init

.gitignore

@@ -0,0 +1,57 @@
+# Python
+__pycache__/
+*.py[cod]
+*.pyo
+*.pyd
+*.egg
+*.egg-info/
+dist/
+build/
+.env
+.venv/
+*.sqlite3
+
+# Jupyter Notebook Checkpoints
+.ipynb_checkpoints/
+
+# VSCode / IDEs
+.vscode/
+.idea/
+
+# OS Files
+.DS_Store
+Thumbs.db
+
+# Log and Cache
+*.log
+*.tmp
+*.bak
+.cache/
+*.swp
+
+# Node.js
+node_modules/
+npm-debug.log*
+
+# Compiled files
+*.o
+*.so
+*.exe
+*.dll
+*.class
+
+# Docker
+.dockerignore
+docker-compose.override.yml
+.docker/
+
+# SATE
+ENNI_test/
+ENNI_test.py
+ENNI_transcript.py
+input/
+performance_eval/
+session_data/
+
+# Model
+CrisperWhisper_local/

CrisperWhisper/.gitignore

@@ -0,0 +1,126 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+docs/_build/doctrees/
+docs/_build/html/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pyright type checker
+.pyright/
+
+# End of https://www.toptal.com/developers/gitignore/api/python
+
+.idea

CrisperWhisper/LICENSE

@@ -0,0 +1,407 @@
+Attribution-NonCommercial 4.0 International
+
+=======================================================================
+
+Creative Commons Corporation ("Creative Commons") is not a law firm and
+does not provide legal services or legal advice. Distribution of
+Creative Commons public licenses does not create a lawyer-client or
+other relationship. Creative Commons makes its licenses and related
+information available on an "as-is" basis. Creative Commons gives no
+warranties regarding its licenses, any material licensed under their
+terms and conditions, or any related information. Creative Commons
+disclaims all liability for damages resulting from their use to the
+fullest extent possible.
+
+Using Creative Commons Public Licenses
+
+Creative Commons public licenses provide a standard set of terms and
+conditions that creators and other rights holders may use to share
+original works of authorship and other material subject to copyright
+and certain other rights specified in the public license below. The
+following considerations are for informational purposes only, are not
+exhaustive, and do not form part of our licenses.
+
+     Considerations for licensors: Our public licenses are
+     intended for use by those authorized to give the public
+     permission to use material in ways otherwise restricted by
+     copyright and certain other rights. Our licenses are
+     irrevocable. Licensors should read and understand the terms
+     and conditions of the license they choose before applying it.
+     Licensors should also secure all rights necessary before
+     applying our licenses so that the public can reuse the
+     material as expected. Licensors should clearly mark any
+     material not subject to the license. This includes other CC-
+     licensed material, or material used under an exception or
+     limitation to copyright. More considerations for licensors:
+    wiki.creativecommons.org/Considerations_for_licensors
+
+     Considerations for the public: By using one of our public
+     licenses, a licensor grants the public permission to use the
+     licensed material under specified terms and conditions. If
+     the licensor's permission is not necessary for any reason--for
+     example, because of any applicable exception or limitation to
+     copyright--then that use is not regulated by the license. Our
+     licenses grant only permissions under copyright and certain
+     other rights that a licensor has authority to grant. Use of
+     the licensed material may still be restricted for other
+     reasons, including because others have copyright or other
+     rights in the material. A licensor may make special requests,
+     such as asking that all changes be marked or described.
+     Although not required by our licenses, you are encouraged to
+     respect those requests where reasonable. More considerations
+     for the public:
+    wiki.creativecommons.org/Considerations_for_licensees
+
+=======================================================================
+
+Creative Commons Attribution-NonCommercial 4.0 International Public
+License
+
+By exercising the Licensed Rights (defined below), You accept and agree
+to be bound by the terms and conditions of this Creative Commons
+Attribution-NonCommercial 4.0 International Public License ("Public
+License"). To the extent this Public License may be interpreted as a
+contract, You are granted the Licensed Rights in consideration of Your
+acceptance of these terms and conditions, and the Licensor grants You
+such rights in consideration of benefits the Licensor receives from
+making the Licensed Material available under these terms and
+conditions.
+
+
+Section 1 -- Definitions.
+
+  a. Adapted Material means material subject to Copyright and Similar
+     Rights that is derived from or based upon the Licensed Material
+     and in which the Licensed Material is translated, altered,
+     arranged, transformed, or otherwise modified in a manner requiring
+     permission under the Copyright and Similar Rights held by the
+     Licensor. For purposes of this Public License, where the Licensed
+     Material is a musical work, performance, or sound recording,
+     Adapted Material is always produced where the Licensed Material is
+     synched in timed relation with a moving image.
+
+  b. Adapter's License means the license You apply to Your Copyright
+     and Similar Rights in Your contributions to Adapted Material in
+     accordance with the terms and conditions of this Public License.
+
+  c. Copyright and Similar Rights means copyright and/or similar rights
+     closely related to copyright including, without limitation,
+     performance, broadcast, sound recording, and Sui Generis Database
+     Rights, without regard to how the rights are labeled or
+     categorized. For purposes of this Public License, the rights
+     specified in Section 2(b)(1)-(2) are not Copyright and Similar
+     Rights.
+  d. Effective Technological Measures means those measures that, in the
+     absence of proper authority, may not be circumvented under laws
+     fulfilling obligations under Article 11 of the WIPO Copyright
+     Treaty adopted on December 20, 1996, and/or similar international
+     agreements.
+
+  e. Exceptions and Limitations means fair use, fair dealing, and/or
+     any other exception or limitation to Copyright and Similar Rights
+     that applies to Your use of the Licensed Material.
+
+  f. Licensed Material means the artistic or literary work, database,
+     or other material to which the Licensor applied this Public
+     License.
+
+  g. Licensed Rights means the rights granted to You subject to the
+     terms and conditions of this Public License, which are limited to
+     all Copyright and Similar Rights that apply to Your use of the
+     Licensed Material and that the Licensor has authority to license.
+
+  h. Licensor means the individual(s) or entity(ies) granting rights
+     under this Public License.
+
+  i. NonCommercial means not primarily intended for or directed towards
+     commercial advantage or monetary compensation. For purposes of
+     this Public License, the exchange of the Licensed Material for
+     other material subject to Copyright and Similar Rights by digital
+     file-sharing or similar means is NonCommercial provided there is
+     no payment of monetary compensation in connection with the
+     exchange.
+
+  j. Share means to provide material to the public by any means or
+     process that requires permission under the Licensed Rights, such
+     as reproduction, public display, public performance, distribution,
+     dissemination, communication, or importation, and to make material
+     available to the public including in ways that members of the
+     public may access the material from a place and at a time
+     individually chosen by them.
+
+  k. Sui Generis Database Rights means rights other than copyright
+     resulting from Directive 96/9/EC of the European Parliament and of
+     the Council of 11 March 1996 on the legal protection of databases,
+     as amended and/or succeeded, as well as other essentially
+     equivalent rights anywhere in the world.
+
+  l. You means the individual or entity exercising the Licensed Rights
+     under this Public License. Your has a corresponding meaning.
+
+
+Section 2 -- Scope.
+
+  a. License grant.
+
+       1. Subject to the terms and conditions of this Public License,
+          the Licensor hereby grants You a worldwide, royalty-free,
+          non-sublicensable, non-exclusive, irrevocable license to
+          exercise the Licensed Rights in the Licensed Material to:
+
+            a. reproduce and Share the Licensed Material, in whole or
+               in part, for NonCommercial purposes only; and
+
+            b. produce, reproduce, and Share Adapted Material for
+               NonCommercial purposes only.
+
+       2. Exceptions and Limitations. For the avoidance of doubt, where
+          Exceptions and Limitations apply to Your use, this Public
+          License does not apply, and You do not need to comply with
+          its terms and conditions.
+
+       3. Term. The term of this Public License is specified in Section
+          6(a).
+
+       4. Media and formats; technical modifications allowed. The
+          Licensor authorizes You to exercise the Licensed Rights in
+          all media and formats whether now known or hereafter created,
+          and to make technical modifications necessary to do so. The
+          Licensor waives and/or agrees not to assert any right or
+          authority to forbid You from making technical modifications
+          necessary to exercise the Licensed Rights, including
+          technical modifications necessary to circumvent Effective
+          Technological Measures. For purposes of this Public License,
+          simply making modifications authorized by this Section 2(a)
+          (4) never produces Adapted Material.
+
+       5. Downstream recipients.
+
+            a. Offer from the Licensor -- Licensed Material. Every
+               recipient of the Licensed Material automatically
+               receives an offer from the Licensor to exercise the
+               Licensed Rights under the terms and conditions of this
+               Public License.
+
+            b. No downstream restrictions. You may not offer or impose
+               any additional or different terms or conditions on, or
+               apply any Effective Technological Measures to, the
+               Licensed Material if doing so restricts exercise of the
+               Licensed Rights by any recipient of the Licensed
+               Material.
+
+       6. No endorsement. Nothing in this Public License constitutes or
+          may be construed as permission to assert or imply that You
+          are, or that Your use of the Licensed Material is, connected
+          with, or sponsored, endorsed, or granted official status by,
+          the Licensor or others designated to receive attribution as
+          provided in Section 3(a)(1)(A)(i).
+
+  b. Other rights.
+
+       1. Moral rights, such as the right of integrity, are not
+          licensed under this Public License, nor are publicity,
+          privacy, and/or other similar personality rights; however, to
+          the extent possible, the Licensor waives and/or agrees not to
+          assert any such rights held by the Licensor to the limited
+          extent necessary to allow You to exercise the Licensed
+          Rights, but not otherwise.
+
+       2. Patent and trademark rights are not licensed under this
+          Public License.
+
+       3. To the extent possible, the Licensor waives any right to
+          collect royalties from You for the exercise of the Licensed
+          Rights, whether directly or through a collecting society
+          under any voluntary or waivable statutory or compulsory
+          licensing scheme. In all other cases the Licensor expressly
+          reserves any right to collect such royalties, including when
+          the Licensed Material is used other than for NonCommercial
+          purposes.
+
+
+Section 3 -- License Conditions.
+
+Your exercise of the Licensed Rights is expressly made subject to the
+following conditions.
+
+  a. Attribution.
+
+       1. If You Share the Licensed Material (including in modified
+          form), You must:
+
+            a. retain the following if it is supplied by the Licensor
+               with the Licensed Material:
+
+                 i. identification of the creator(s) of the Licensed
+                    Material and any others designated to receive
+                    attribution, in any reasonable manner requested by
+                    the Licensor (including by pseudonym if
+                    designated);
+
+                ii. a copyright notice;
+
+               iii. a notice that refers to this Public License;
+
+                iv. a notice that refers to the disclaimer of
+                    warranties;
+
+                 v. a URI or hyperlink to the Licensed Material to the
+                    extent reasonably practicable;
+
+            b. indicate if You modified the Licensed Material and
+               retain an indication of any previous modifications; and
+
+            c. indicate the Licensed Material is licensed under this
+               Public License, and include the text of, or the URI or
+               hyperlink to, this Public License.
+
+       2. You may satisfy the conditions in Section 3(a)(1) in any
+          reasonable manner based on the medium, means, and context in
+          which You Share the Licensed Material. For example, it may be
+          reasonable to satisfy the conditions by providing a URI or
+          hyperlink to a resource that includes the required
+          information.
+
+       3. If requested by the Licensor, You must remove any of the
+          information required by Section 3(a)(1)(A) to the extent
+          reasonably practicable.
+
+       4. If You Share Adapted Material You produce, the Adapter's
+          License You apply must not prevent recipients of the Adapted
+          Material from complying with this Public License.
+
+
+Section 4 -- Sui Generis Database Rights.
+
+Where the Licensed Rights include Sui Generis Database Rights that
+apply to Your use of the Licensed Material:
+
+  a. for the avoidance of doubt, Section 2(a)(1) grants You the right
+     to extract, reuse, reproduce, and Share all or a substantial
+     portion of the contents of the database for NonCommercial purposes
+     only;
+
+  b. if You include all or a substantial portion of the database
+     contents in a database in which You have Sui Generis Database
+     Rights, then the database in which You have Sui Generis Database
+     Rights (but not its individual contents) is Adapted Material; and
+
+  c. You must comply with the conditions in Section 3(a) if You Share
+     all or a substantial portion of the contents of the database.
+
+For the avoidance of doubt, this Section 4 supplements and does not
+replace Your obligations under this Public License where the Licensed
+Rights include other Copyright and Similar Rights.
+
+
+Section 5 -- Disclaimer of Warranties and Limitation of Liability.
+
+  a. UNLESS OTHERWISE SEPARATELY UNDERTAKEN BY THE LICENSOR, TO THE
+     EXTENT POSSIBLE, THE LICENSOR OFFERS THE LICENSED MATERIAL AS-IS
+     AND AS-AVAILABLE, AND MAKES NO REPRESENTATIONS OR WARRANTIES OF
+     ANY KIND CONCERNING THE LICENSED MATERIAL, WHETHER EXPRESS,
+     IMPLIED, STATUTORY, OR OTHER. THIS INCLUDES, WITHOUT LIMITATION,
+     WARRANTIES OF TITLE, MERCHANTABILITY, FITNESS FOR A PARTICULAR
+     PURPOSE, NON-INFRINGEMENT, ABSENCE OF LATENT OR OTHER DEFECTS,
+     ACCURACY, OR THE PRESENCE OR ABSENCE OF ERRORS, WHETHER OR NOT
+     KNOWN OR DISCOVERABLE. WHERE DISCLAIMERS OF WARRANTIES ARE NOT
+     ALLOWED IN FULL OR IN PART, THIS DISCLAIMER MAY NOT APPLY TO YOU.
+
+  b. TO THE EXTENT POSSIBLE, IN NO EVENT WILL THE LICENSOR BE LIABLE
+     TO YOU ON ANY LEGAL THEORY (INCLUDING, WITHOUT LIMITATION,
+     NEGLIGENCE) OR OTHERWISE FOR ANY DIRECT, SPECIAL, INDIRECT,
+     INCIDENTAL, CONSEQUENTIAL, PUNITIVE, EXEMPLARY, OR OTHER LOSSES,
+     COSTS, EXPENSES, OR DAMAGES ARISING OUT OF THIS PUBLIC LICENSE OR
+     USE OF THE LICENSED MATERIAL, EVEN IF THE LICENSOR HAS BEEN
+     ADVISED OF THE POSSIBILITY OF SUCH LOSSES, COSTS, EXPENSES, OR
+     DAMAGES. WHERE A LIMITATION OF LIABILITY IS NOT ALLOWED IN FULL OR
+     IN PART, THIS LIMITATION MAY NOT APPLY TO YOU.
+
+  c. The disclaimer of warranties and limitation of liability provided
+     above shall be interpreted in a manner that, to the extent
+     possible, most closely approximates an absolute disclaimer and
+     waiver of all liability.
+
+
+Section 6 -- Term and Termination.
+
+  a. This Public License applies for the term of the Copyright and
+     Similar Rights licensed here. However, if You fail to comply with
+     this Public License, then Your rights under this Public License
+     terminate automatically.
+
+  b. Where Your right to use the Licensed Material has terminated under
+     Section 6(a), it reinstates:
+
+       1. automatically as of the date the violation is cured, provided
+          it is cured within 30 days of Your discovery of the
+          violation; or
+
+       2. upon express reinstatement by the Licensor.
+
+     For the avoidance of doubt, this Section 6(b) does not affect any
+     right the Licensor may have to seek remedies for Your violations
+     of this Public License.
+
+  c. For the avoidance of doubt, the Licensor may also offer the
+     Licensed Material under separate terms or conditions or stop
+     distributing the Licensed Material at any time; however, doing so
+     will not terminate this Public License.
+
+  d. Sections 1, 5, 6, 7, and 8 survive termination of this Public
+     License.
+
+
+Section 7 -- Other Terms and Conditions.
+
+  a. The Licensor shall not be bound by any additional or different
+     terms or conditions communicated by You unless expressly agreed.
+
+  b. Any arrangements, understandings, or agreements regarding the
+     Licensed Material not stated herein are separate from and
+     independent of the terms and conditions of this Public License.
+
+
+Section 8 -- Interpretation.
+
+  a. For the avoidance of doubt, this Public License does not, and
+     shall not be interpreted to, reduce, limit, restrict, or impose
+     conditions on any use of the Licensed Material that could lawfully
+     be made without permission under this Public License.
+
+  b. To the extent possible, if any provision of this Public License is
+     deemed unenforceable, it shall be automatically reformed to the
+     minimum extent necessary to make it enforceable. If the provision
+     cannot be reformed, it shall be severed from this Public License
+     without affecting the enforceability of the remaining terms and
+     conditions.
+
+  c. No term or condition of this Public License will be waived and no
+     failure to comply consented to unless expressly agreed to by the
+     Licensor.
+
+  d. Nothing in this Public License constitutes or may be interpreted
+     as a limitation upon, or waiver of, any privileges and immunities
+     that apply to the Licensor or You, including from the legal
+     processes of any jurisdiction or authority.
+
+=======================================================================
+
+Creative Commons is not a party to its public
+licenses. Notwithstanding, Creative Commons may elect to apply one of
+its public licenses to material it publishes and in those instances
+will be considered the “Licensor.” The text of the Creative Commons
+public licenses is dedicated to the public domain under the CC0 Public
+Domain Dedication. Except for the limited purpose of indicating that
+material is shared under a Creative Commons public license or as
+otherwise permitted by the Creative Commons policies published at
+creativecommons.org/policies, Creative Commons does not authorize the
+use of the trademark "Creative Commons" or any other trademark or logo
+of Creative Commons without its prior written consent including,
+without limitation, in connection with any unauthorized modifications
+to any of its public licenses or any other arrangements,
+understandings, or agreements concerning use of licensed material. For
+the avoidance of doubt, this paragraph does not form part of the
+public licenses.
+
+Creative Commons may be contacted at creativecommons.org.

CrisperWhisper/README.md

@@ -0,0 +1,325 @@
+# CrisperWhisper
+
+**CrisperWhisper** is an advanced variant of OpenAI's Whisper, designed for fast, precise, and verbatim speech recognition with accurate (**crisp**) word-level timestamps. Unlike the original Whisper, which tends to omit disfluencies and follows more of a intended transcription style, CrisperWhisper aims to transcribe every spoken word exactly as it is, including fillers, pauses, stutters and false starts.
+
+## Key Features
+
+- 🎯 **Accurate Word-Level Timestamps**: Provides precise timestamps, even around disfluencies and pauses, by utilizing an adjusted tokenizer and a custom attention loss during training.
+- 📝 **Verbatim Transcription**: Transcribes every spoken word exactly as it is, including and differentiating fillers like "um" and "uh".
+- 🔍 **Filler Detection**: Detects and accurately transcribes fillers.
+- 🛡️ **Hallucination Mitigation**: Minimizes transcription hallucinations to enhance accuracy.
+
+## Table of Contents
+
+- [Key Features](#key-features)
+- [Highlights](#highlights)
+- [Performance Overview](#1-performance-overview)
+  - [Qualitative Performance Overview](#11-qualitative-performance-overview)
+  - [Quantitative Performance Overview](#12-quantitative-performance-overview)
+    - [Transcription Performance](#transcription-performance)
+    - [Segmentation Performance](#segmentation-performance)
+- [Setup](#2-setup-⚙️)
+  - [Prerequisites](#21-prerequisites)
+  - [Environment Setup](#22-environment-setup)
+- [Usage](#3-usage)
+  - [with transformers](#31-usage-with-🤗-transformers)
+  - [with faster whisper](#32-usage-with-faster-whisper)
+- [Running the Streamlit App](#4-running-the-streamlit-app)
+    - [Prerequisites](#41-prerequisites)
+    - [Steps to Run the Streamlit App](#42-steps-to-run-the-streamlit-app)
+    - [Features of the App](#43-features-of-the-app)
+- [How](#5-how)
+- [License](#license)
+
+
+## Highlights
+
+- 🏆 **1st place** on the [OpenASR Leaderboard](https://huggingface.co/spaces/hf-audio/open_asr_leaderboard) in verbatim datasets (TED, AMI) and overall.
+- 🎓 **Accepted at INTERSPEECH 2024**.
+- 📄 **Paper Drop**: Check out our [paper](https://arxiv.org/abs/2408.16589) for details and reasoning behind our tokenizer adjustment.
+- ✨ **New Feature**: Not mentioned in the paper is a added AttentionLoss to further improve timestamp accuracy. By specifically adding a loss to train the attention scores used for the DTW alignment using timestamped data we significantly boosted the alignment performance.
+
+
+
+## 1. Performance Overview
+
+### 1.1 Qualitative Performance Overview
+
+
+| Audio | Whisper Large V3 | Crisper Whisper |
+|-------|------------------------|------------------------|
+| [Demo de 1](https://github.com/user-attachments/assets/c8608ca8-5e02-4c4a-afd3-8f7c5bff75d5) | Er war kein Genie, aber doch ein fähiger Ingenieur. | Es ist zwar kein. Er ist zwar kein Genie, aber doch ein fähiger Ingenieur.|
+| [Demo de 2](https://github.com/user-attachments/assets/c68414b1-0f84-441c-b39b-29069487edb6) | Leider müssen wir in diesen schweren Zeiten auch unserem Tagesgeschäft nachgehen. Der hier vorgelegte Kulturhaushalt der Ampelregierung strebt an, den Erfolgskurs der Union zumindest fiskalisch fortzuführen. | Leider [UH] müssen wir in diesen [UH] schweren Zeiten auch [UH] unserem [UH] Tagesgeschäft nachgehen. Der hier [UH] vorgelegte [UH] Kulturhaushalt der [UH] Ampelregierung strebt an, den [UH] Erfolgskurs der Union [UH] zumindest [UH] fiskalisch fortzuführen. Es. |
+| [Demo de 3](https://github.com/user-attachments/assets/0c1ed60c-2829-47e4-b7ba-eb584b0a5e9a) | die über alle FRA-Fraktionen hinweg gut im Blick behalten sollten, auch weil sie teilweise sehr teeteuer sind. Aber nicht nur, weil sie teeteuer sind. Wir steigen mit diesem Endentwurf ein in die sogenannten Pandemie-Bereitschaftsverträge.| Die über alle Fr Fraktionen hinweg gut im [UH] Blick behalten sollten, auch weil sie teil teilweise sehr te teuer sind. Aber nicht nur, weil sie te teuer sind. Wir [UH] steigen mit diesem Ent Entwurf ein in die sogenannten Pand Pandemiebereitschaftsverträge. |
+| [Demo en 1](https://github.com/user-attachments/assets/cde5d69c-657f-4ae4-b4ae-b958ea2eacc5) | alternative is you can get like, you have those Dr. Bronner's| Alternative is you can get like [UH] you have those, you know, those doctor Brahmer's. |
+| [Demo en 2](https://github.com/user-attachments/assets/906e307d-5613-4c41-9c61-65f4beede1fd) | influence our natural surrounding? How does it influence our ecosystem? | Influence our [UM] our [UH] our natural surrounding. How does it influence our ecosystem? |
+| [Demo en 3](https://github.com/user-attachments/assets/6c09cd58-a574-4697-9a7e-92e416cf2522) | and always find a place on the street to park and it was easy and you weren't a long distance away from wherever it was that you were trying to go. So I remember that being a lot of fun and easy to do and there were nice places to go and good events to attend. Come downtown and you had the Warner Theater and | And always find a place on the street to park. And and it was it was easy and you weren't a long distance away from wherever it was that you were trying to go. So, I I I remember that being a lot of fun and easy to do and there were nice places to go and, [UM] i good events to attend. Come downtown and you had the Warner Theater and, [UM] |
+| [Demo en 4](https://github.com/user-attachments/assets/7df19486-5e4e-4443-8528-09b07dddf61a) | you know, more masculine, who were rough, and that definitely wasn't me. Then, you know, I was very smart because my father made sure I was smart, you know. So, you know, I hung around those people, you know. And then you had the ones that were just out doing things that they shouldn't have been doing also. So, yeah, I was in the little geek squad. You were in the little geek squad. Yeah. | you know, more masculine, who were rough, and that definitely wasn't me. Then, you know, I was very smart because my father made sure I was smart. You know, so, [UM] you know, I I hung around those people, you know. And then you had the ones that were just just out doing things that they shouldn't have been doing also. So yeah, I was the l I was in the little geek squad. Do you |
+
+### 1.2 Quantitative Performance Overview
+
+#### Transcription Performance
+
+CrisperWhisper significantly outperforms Whisper Large v3, especially on datasets that have a more verbatim transcription style in the ground truth, such as AMI and TED-LIUM.
+
+| Dataset            | CrisperWhisper | Whisper Large v3 | 
+|----------------------|:--------------:|:----------------:|
+| [AMI](https://huggingface.co/datasets/edinburghcstr/ami)                 | **8.72**       | 16.01            |    
+| [Earnings22](https://huggingface.co/datasets/revdotcom/earnings22)           | 12.37          | **11.3**        | 
+| [GigaSpeech](https://huggingface.co/datasets/speechcolab/gigaspeech)         | 10.27          | **10.02**        |     
+| [LibriSpeech clean](https://huggingface.co/datasets/openslr/librispeech_asr)   | **1.74**       | 2.03            |    
+| [LibriSpeech other](https://huggingface.co/datasets/openslr/librispeech_asr)   | 3.97           | **3.91**         |      
+| [SPGISpeech](https://huggingface.co/datasets/kensho/spgispeech)          | **2.71**           | 2.95        |     
+| [TED-LIUM](https://huggingface.co/datasets/LIUM/tedlium)             | **3.35**          | 3.9        |    
+| [VoxPopuli](https://huggingface.co/datasets/facebook/voxpopuli)           | **8.61**           | 9.52         |  
+| [CommonVoice](https://huggingface.co/datasets/mozilla-foundation/common_voice_9_0)       | **8.19**           | 9.67        |      
+| **Average WER**      | **6.66**       | 7.7         |  
+
+#### Segmentation Performance
+
+CrisperWhisper demonstrates superior performance segmentation performance. This performance gap is especially pronounced around disfluencies and pauses.
+The following table uses the metrics as defined in the paper. For this table we used a collar of 50ms. Heads for each Model were selected using the method described in the [How](#5-how) section and the result attaining the highest F1 Score was choosen for each model using varying number of heads.
+
+| Dataset | Metric | CrisperWhisper | Whisper Large v2 | Whisper Large v3 |
+|---------|--------|------------------|------------------|------------------|
+| [AMI IHM](https://groups.inf.ed.ac.uk/ami/corpus/) | F1 Score | **0.79** | 0.63 | 0.66 |
+| | Avg IOU | **0.67** | 0.54 | 0.53 |
+| [Common Voice](https://commonvoice.mozilla.org/en/datasets) | F1 Score | **0.80** | 0.42 | 0.48 |
+| | Avg IOU | **0.70** | 0.32 | 0.43 |
+| [TIMIT](https://catalog.ldc.upenn.edu/LDC93S1) | F1 Score | **0.69** | 0.40 | 0.54 |
+| | Avg IOU | **0.56** | 0.32 | 0.43 |
+
+More plots and ablations can be found in the `run_experiments/plots` folder.
+
+## 2. Setup ⚙️
+
+### 2.1 Prerequisites
+
+- **Python**: 3.10
+- **PyTorch**: 2.0
+- **NVIDIA Libraries**: cuBLAS 11.x and cuDNN 8.x (for GPU execution)
+
+### 2.2 Environment Setup
+
+1. **Clone the Repository**:
+    ```bash
+    git clone https://github.com/nyrahealth/CrisperWhisper.git
+    cd CrisperWhisper
+    ```
+
+2. **Create Python Environment**:
+    ```bash
+    conda create --name crisperWhisper python=3.10
+    conda activate crisperWhisper
+    ```
+
+
+3. **Install Dependencies**:
+    ```bash
+    pip install -r requirements.txt
+    ```
+
+4. **Additional Installations**:
+    Follow OpenAI's instructions to install additional dependencies like `ffmpeg` and `rust`: [Whisper Setup](https://github.com/openai/whisper#setup).
+
+## 3. Usage 
+
+Here's how to use CrisperWhisper in your Python scripts:
+First install our custom transformers fork for the most accurate timestamps:
+```
+pip install git+https://github.com/nyrahealth/transformers.git@crisper_whisper
+```
+
+### 3.1 Usage with 🤗 transformers
+First make sure that you have a huggingface account and accept the licensing of the [model](https://huggingface.co/nyrahealth/CrisperWhisper). Grab you huggingface access token and login so you are certainly able to download the model.
+
+```bash
+huggingface-cli login
+``` 
+
+```python
+import os
+import sys
+import torch
+
+from datasets import load_dataset
+from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
+from utils import adjust_pauses_for_hf_pipeline_output
+
+
+
+device = "cuda:0" if torch.cuda.is_available() else "cpu"
+torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
+
+model_id = "nyrahealth/CrisperWhisper"
+
+model = AutoModelForSpeechSeq2Seq.from_pretrained(
+    model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
+)
+model.to(device)
+
+processor = AutoProcessor.from_pretrained(model_id)
+
+pipe = pipeline(
+    "automatic-speech-recognition",
+    model=model,
+    tokenizer=processor.tokenizer,
+    feature_extractor=processor.feature_extractor,
+    chunk_length_s=30,
+    batch_size=16,
+    return_timestamps='word',
+    torch_dtype=torch_dtype,
+    device=device,
+)
+
+dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation")
+sample = dataset[0]["audio"]
+hf_pipeline_output = pipe(sample)
+crisper_whisper_result = adjust_pauses_for_hf_pipeline_output(hf_pipeline_output)
+print(crisper_whisper_result)
+```
+### 3.2 Usage with faster whisper
+
+We also provide a converted model to be compatible with [faster whisper](https://github.com/SYSTRAN/faster-whisper). However, due to the different implementation of the timestamp calculation in faster whisper or more precisely [CTranslate2](https://github.com/OpenNMT/CTranslate2/) the timestamp accuracy can not be guaranteed. 
+
+First make sure that you have a huggingface account and accept the licensing of the [model](https://huggingface.co/nyrahealth/faster_CrisperWhisper). Grab you huggingface access token and login so you are certainly able to download the model.
+```bash
+huggingface-cli login
+``` 
+
+```python
+from faster_whisper import WhisperModel
+from datasets import load_dataset
+faster_whisper_model = 'nyrahealth/faster_CrisperWhisper'
+
+# Initialize the Whisper model
+
+device = "cuda:0" if torch.cuda.is_available() else "cpu"
+torch_dtype = "float16" if torch.cuda.is_available() else "float32"
+model = WhisperModel(faster_whisper_model, device=device, compute_type="float32")
+dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation")
+sample = dataset[0]["audio"]
+
+segments, info = model.transcribe(sample['array'], beam_size=1, language='en', word_timestamps = True, without_timestamps= True)
+
+for segment in segments:
+    print(segment)
+```
+
+### 3.3 Commandline usage
+
+First make sure that you have a huggingface account and accept the licensing of the model. Grab you huggingface access token and login so you are certainly able to download the model.
+```bash
+    huggingface-cli login
+ ```
+afterwards:
+
+To transcribe an audio file, use the following command:
+
+```bash
+python transcribe.py --f <path_to_audio_file>
+```
+
+## 4. Running the Streamlit App
+
+To use the CrisperWhisper model with a user-friendly interface, you can run the provided Streamlit app. This app allows you to record or upload audio files for transcription and view the results with accurate word-level timestamps.
+
+### 4.1 Prerequisites
+
+Make sure you have followed the [Setup ⚙️](#setup) instructions above and have the `crisperWhisper` environment activated.
+
+### 4.2 Steps to Run the Streamlit App
+
+1. **Activate the Conda Environment**
+
+    Ensure you are in the `crisperWhisper` environment:
+    ```sh
+    conda activate crisperWhisper
+    ```
+
+2. **Navigate to the App Directory**
+
+    Change directory to where the `app.py` script is located:
+
+
+3. **Run the Streamlit App**
+
+    Use the following command to run the app. Make sure to replace `/path/to/your/model` with the actual path to your CrisperWhisper model directory:
+    ```sh
+    streamlit run app.py -- --model_id /path/to/your/model
+    ```
+
+    For example:
+    ```sh
+    streamlit run app.py -- --model_id nyrahealth/CrisperWhisper
+    ```
+
+4. **Access the App**
+
+    After running the command, the Streamlit server will start, and you can access the app in your web browser at:
+    ```
+    http://localhost:8501
+    ```
+
+### 4.3 Features of the App
+
+- **Record Audio**: Record audio directly using your microphone.
+- **Upload Audio**: Upload audio files in formats like WAV, MP3, or OGG.
+- **Transcription**: Get accurate verbatim transcriptions including fillers
+- **Video Generation**: View the transcription with timestamps alongside a video with a black background.
+
+## 5. How?
+
+
+We employ the popular Dynamic Time Warping (DTW) on the Whisper cross-attention scores, as detailed in our [paper](https://arxiv.org/abs/2408.16589) to derive word-level timestamps. By leveraging our retokenization process, this method allows us to consistently detect pauses. Given that the accuracy of the timestamps heavily depends on the DTW cost matrix and, consequently, on the quality of the cross-attentions, we developed a specialized loss function for the selected alignment heads to enhance precision.
+
+Although this loss function was not included in the original [paper](https://arxiv.org/abs/2408.16589) due to time constraints preventing the completion of experiments and training before the submission deadline, it has been used to train our publicly available models.
+Key Features of this loss are as follows:
+
+1. **Data Preparation**
+    - We used datasets with word-level timestamp annotations, such as [AMI IHM](https://groups.inf.ed.ac.uk/ami/corpus/) and [TIMIT](https://catalog.ldc.upenn.edu/LDC93S1)   , but required additional timestamped data.
+    - To address this, we validated the alignment accuracy of several forced alignment tools using a small hand-labeled dataset.
+    - Based on this validation, we chose the [PyTorch CTC aligner](https://pytorch.org/audio/main/tutorials/ctc_forced_alignment_api_tutorial.html) to generate more time-aligned data from the CommonVoice dataset.
+    - Because the [PyTorch CTC aligner](https://pytorch.org/audio/main/tutorials/ctc_forced_alignment_api_tutorial.html) tends to overestimate pause durations, we applied the same pause-splitting method detailed in our [paper](...) to correct these errors. The effectiveness of this correction was confirmed using our hand-labeled dataset.
+
+2. **Token-Word Alignment**
+    - Due to retokenization as detailed in our [paper](https://arxiv.org/abs/2408.16589), each token is either part of a word or a pause/space, but never both
+    - Therefore each token can be cleanly aligned to a word OR a space/pause
+
+3. **Ground Truth Cross-Attention**
+    - We define the cross-attention ground truth for tokens as the L2-normalized vector, where:
+        - A value of 1 indicates that the word is active according to the word-level ground truth timestamp.
+        - A value of 0 indicates that no attention should be paid.
+    - To account for small inaccuracies in the ground truth timestamps, we apply a linear interpolation of 4 steps (8 milliseconds) on both sides of the ground truth vector, transitioning smoothly from 0 to 1.
+
+4. **Loss Calculation**
+- The loss function is defined as `1 - cosine similarity`  between the predicted cross-attention vector (when predicting a token) and the ground truth cross-attention vector.
+- This loss is averaged across all predicted tokens and alignment heads.
+
+5 **Alignment Head selection**
+- To choose the heads for alignment we evaluated the alignment performance of each individual decoder attention head on the timestamped timit dataset.
+- We choose the 15 best performing heads and finetune them using our attention loss.
+
+5. **Training Details**
+- Since most of our samples during training were shorter than 30 seconds we shift the audio sample and corresponding timestamp ground truth around with a 50% probability to mitigate the cross attentions ,,overfitting" to early positions of the encoder output.
+- If we have more than 40ms of silence (before or after shifting) we prepend the ground truth transcript ( and corresponding cross attention ground truth) with a space so the model has to accurately predict the starting time of the first word.
+- We use [WavLM](https://arxiv.org/abs/2110.13900) augmentations during Training adding random speech samples or noise to the audio wave to generally increase robustness of the transcription and stability of the alignment heads.
+- We clip ,,predicted" values in the cross attention vectors 4 seconds before and 4 seconds after the groundtruth word they belong to to 0. This is to decrease the dimensionality of the cross attention vector and therefore emphasize the attention where it counts in the loss and ultimately for the alignment.
+- With a probability of 1% we use samples containing exclusively noise where the model has to return a empty prediction to improve hallucination.
+- The Model is trained in three stages, in the first stage we use around 10000 hours of audio to adjust Whisper to the new tokenizer. In the second stage we exclusively use high quality datasets that are transcribed in a verbatim fashion. Finally we continue training on this verbatim mixture and add the attention loss for another 6000 steps.
+
+
+## License
+
+```markdown
+Shield: [![CC BY-NC 4.0][cc-by-nc-shield]][cc-by-nc]
+
+This work is licensed under a
+[Creative Commons Attribution-NonCommercial 4.0 International License][cc-by-nc].
+
+[![CC BY-NC 4.0][cc-by-nc-image]][cc-by-nc]
+
+[cc-by-nc]: https://creativecommons.org/licenses/by-nc/4.0/
+[cc-by-nc-image]: https://licensebuttons.net/l/by-nc/4.0/88x31.png
+[cc-by-nc-shield]: https://img.shields.io/badge/License-CC%20BY--NC%204.0-lightgrey.svg
+```

CrisperWhisper/app.py

@@ -0,0 +1,173 @@
+import argparse
+import io
+from typing import Any, Dict, List, Tuple, Union
+
+import moviepy.editor as mp
+import numpy as np
+import streamlit as st
+import torch
+import torchaudio
+import torchaudio.transforms as T
+from scipy.io import wavfile
+from streamlit_mic_recorder import mic_recorder
+from transformers import (
+    AutomaticSpeechRecognitionPipeline,
+    AutoModelForSpeechSeq2Seq,
+    AutoProcessor,
+    pipeline,
+)
+
+
+def parse_arguments() -> argparse.Namespace:
+    """Parse command-line arguments."""
+    parser = argparse.ArgumentParser(
+        description="Streamlit app for speech transcription."
+    )
+    parser.add_argument(
+        "--model_id", type=str, required=True, help="Path to the model directory"
+    )
+    return parser.parse_args()
+
+
+# Load model and processor from the specified path
+@st.cache_resource  # type: ignore
+def load_model_and_processor(
+    model_id: str,
+) -> Tuple[AutoModelForSpeechSeq2Seq, AutoProcessor]:
+    model = AutoModelForSpeechSeq2Seq.from_pretrained(
+        model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
+    )
+    model.to(device)
+    model.generation_config.median_filter_width = 3
+    processor = AutoProcessor.from_pretrained(model_id)
+    return model, processor
+
+
+# Setup the pipeline
+@st.cache_resource  # type: ignore
+def setup_pipeline(
+    _model: AutoModelForSpeechSeq2Seq, _processor: AutoProcessor
+) -> AutomaticSpeechRecognitionPipeline:
+    return pipeline(
+        "automatic-speech-recognition",
+        model=_model,
+        tokenizer=_processor.tokenizer,
+        feature_extractor=_processor.feature_extractor,
+        chunk_length_s=30,
+        batch_size=1,
+        return_timestamps=True,
+        torch_dtype=torch_dtype,
+        device=device,
+    )
+
+
+def wav_to_black_mp4(wav_path: str, output_path: str, fps: int = 25) -> None:
+    """Convert WAV file to a black-screen MP4 with the same audio."""
+    waveform, sample_rate = torchaudio.load(wav_path)
+    duration: float = waveform.shape[1] / sample_rate
+    audio = mp.AudioFileClip(wav_path)
+    black_clip = mp.ColorClip((256, 250), color=(0, 0, 0), duration=duration)
+    final_clip = black_clip.set_audio(audio)
+    final_clip.write_videofile(output_path, fps=fps)
+
+
+def timestamps_to_vtt(timestamps: List[Dict[str, Union[str, Any]]]) -> str:
+    """Convert timestamps to VTT format."""
+    vtt_content: str = "WEBVTT\n\n"
+    for word in timestamps:
+        start_time, end_time = word["timestamp"]
+        start_time_str = f"{int(start_time // 3600)}:{int(start_time // 60 % 60):02d}:{start_time % 60:06.3f}"
+        end_time_str = f"{int(end_time // 3600)}:{int(end_time // 60 % 60):02d}:{end_time % 60:06.3f}"
+        vtt_content += f"{start_time_str} --> {end_time_str}\n{word['text']}\n\n"
+    return vtt_content
+
+
+def process_audio_bytes(audio_bytes: bytes) -> torch.Tensor:
+    """Process audio bytes to the required format."""
+    audio_stream = io.BytesIO(audio_bytes)
+    sr, y = wavfile.read(audio_stream)
+    y = y.astype(np.float32)
+    y_mean = np.mean(y)
+    y_std = np.std(y)
+    y_normalized = (y - y_mean) / y_std
+    transform = T.Resample(sr, 16000)
+    waveform = transform(torch.unsqueeze(torch.tensor(y_normalized / 8), 0))
+    torchaudio.save("sample.wav", waveform, sample_rate=16000)
+    return waveform
+
+
+def transcribe(audio_bytes: bytes) -> Dict[str, Any]:
+    """Transcribe the given audio bytes."""
+    waveform = process_audio_bytes(audio_bytes)
+    transcription = pipe(waveform[0, :].numpy(), return_timestamps="word")
+    return transcription
+
+
+args = parse_arguments()
+model_id = args.model_id
+
+# Set up device and data type for processing
+device: str = "cuda:0" if torch.cuda.is_available() else "cpu"
+torch_dtype: torch.dtype = torch.float16 if torch.cuda.is_available() else torch.float32
+model, processor = load_model_and_processor(model_id)
+pipe = setup_pipeline(model, processor)
+
+# Streamlit app interface
+st.title("CrisperWhisper++ 🦻")
+st.subheader("Caution when using. Make sure you can handle the crispness. ⚠️")
+st.write("🎙️ Record an audio to transcribe or 📁 upload an audio file.")
+
+# Audio recorder component
+audio = mic_recorder(
+    start_prompt="Start recording",
+    stop_prompt="Stop recording",
+    just_once=False,
+    use_container_width=False,
+    format="wav",
+    callback=None,
+    args=(),
+    kwargs={},
+    key=None,
+)
+
+audio_bytes: Union[bytes, None] = audio["bytes"] if audio else None
+
+# Audio file upload handling
+audio_file = st.file_uploader("Or upload an audio file", type=["wav", "mp3", "ogg"])
+
+if audio_file is not None:
+    audio_bytes = audio_file.getvalue()
+
+if audio_bytes:
+    try:
+        transcription = transcribe(audio_bytes)
+        vtt = timestamps_to_vtt(transcription["chunks"])
+
+        with open("subtitles.vtt", "w") as file:
+            file.write(vtt)
+
+        wav_to_black_mp4("sample.wav", "video.mp4")
+
+        st.video("video.mp4", subtitles="subtitles.vtt")
+        st.subheader("Transcription:")
+        st.markdown(
+            f"""
+            <div style="background-color: #f0f0f0; padding: 10px; border-radius: 5px;">
+                <p style="font-size: 16px; color: #333;">{transcription['text']}</p>
+            </div>
+            """,
+            unsafe_allow_html=True,
+        )
+    except Exception as e:
+        st.error(f"An error occurred during transcription: {e}")
+
+# Footer
+st.markdown(
+    """
+    <hr>
+    <footer>
+        <p style="text-align: center;">© 2024 nyra health GmbH</p>
+    </footer>
+    """,
+    unsafe_allow_html=True,
+)

CrisperWhisper/requirements.txt

@@ -0,0 +1,9 @@
+torch
+torchaudio
+git+https://github.com/nyrahealth/transformers.git@crisper_whisper
+accelerate
+scipy
+streamlit
+moviepy
+streamlit_mic_recorder
+librosa

CrisperWhisper/transcribe.py

@@ -0,0 +1,56 @@
+import argparse
+import os
+import sys
+import torch
+from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
+
+
+def transcribe_audio(file_path):
+    device = "cuda:0" if torch.cuda.is_available() else "cpu"
+    torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
+
+    model_id = "nyrahealth/CrisperWhisper"  # You can change this to a different model if needed
+
+    model = AutoModelForSpeechSeq2Seq.from_pretrained(
+        model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
+    )
+    model.to(device)
+
+    processor = AutoProcessor.from_pretrained(model_id)
+
+    pipe = pipeline(
+        "automatic-speech-recognition",
+        model=model,
+        tokenizer=processor.tokenizer,
+        feature_extractor=processor.feature_extractor,
+        chunk_length_s=30,
+        batch_size=16,
+        return_timestamps="word",
+        torch_dtype=torch_dtype,
+        device=device,
+    )
+
+    result = pipe(file_path)
+    return result
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Transcribe an audio file.")
+    parser.add_argument("--f", type=str, required=True, help="Path to the audio file")
+    args = parser.parse_args()
+
+    if not os.path.exists(args.f):
+        print(f"Error: The file '{args.f}' does not exist.")
+        sys.exit(1)
+
+    try:
+        transcription = transcribe_audio(args.f)
+        print("Transcription:")
+        print(transcription["text"])
+    except Exception as e:
+        print(f"An error occurred while transcribing the audio: {str(e)}")
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()

CrisperWhisper/utils.py

@@ -0,0 +1,29 @@
+def adjust_pauses_for_hf_pipeline_output(pipeline_output, split_threshold=0.12):
+    """
+    Adjust pause timings by distributing pauses up to the threshold evenly between adjacent words.
+    """
+
+    adjusted_chunks = pipeline_output["chunks"].copy()
+
+    for i in range(len(adjusted_chunks) - 1):
+        current_chunk = adjusted_chunks[i]
+        next_chunk = adjusted_chunks[i + 1]
+
+        current_start, current_end = current_chunk["timestamp"]
+        next_start, next_end = next_chunk["timestamp"]
+        pause_duration = next_start - current_end
+
+        if pause_duration > 0:
+            if pause_duration > split_threshold:
+                distribute = split_threshold / 2
+            else:
+                distribute = pause_duration / 2
+
+            # Adjust current chunk end time
+            adjusted_chunks[i]["timestamp"] = (current_start, current_end + distribute)
+
+            # Adjust next chunk start time
+            adjusted_chunks[i + 1]["timestamp"] = (next_start - distribute, next_end)
+    pipeline_output["chunks"] = adjusted_chunks
+
+    return pipeline_output

Dockerfile

@@ -0,0 +1,13 @@
+FROM continuumio/miniconda3:24.11.1-0
+WORKDIR /sate
+
+COPY . .
+
+RUN conda env create -f environment_sate_0.11.yml
+
+SHELL ["conda", "run", "-n", "SATE", "/bin/bash", "-c"]
+
+EXPOSE 5000
+
+CMD ["conda", "run", "--no-capture-output", "-n", "SATE", "python", "main_socket.py"]
+

annotation.py

@@ -0,0 +1,78 @@
+import os
+import json
+
+def annotate_transcript(session_id, base_dir="session_data"):
+
+    session_dir = os.path.join(base_dir, session_id)
+    json_file = os.path.join(session_dir, f"{session_id}_transcriptionCW.json")
+    output_file = os.path.join(session_dir, "annotation_result.txt")
+    
+    if not os.path.exists(json_file):
+        print(f"Error: could not find {json_file}")
+        return
+    
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+    
+    segments = data.get("segments", [])
+    annotated_lines = []
+    
+    for segment in segments:
+        speaker = segment.get("speaker", "UNKNOWN")
+        words = segment.get("words", [])
+        n = len(words)
+        
+        pause_map = {}
+        for pause in segment.get("pauses", []):
+            pause_start = pause.get("start")
+            duration = pause.get("duration")
+            for idx, token in enumerate(words):
+                if abs(token.get("end", 0) - pause_start) < 0.01:
+                    pause_map.setdefault(idx, []).append(f"({duration})")
+                    break
+        
+        rep_map = {}
+        for rep in segment.get("repetitions", []):
+            indices = rep.get("words", [])
+            if indices:
+                start_idx = indices[0]
+                end_idx = indices[-1]
+                rep_content = rep.get("content", "")
+                rep_map[start_idx] = (end_idx, rep_content)
+        
+        annotated_tokens = []
+        i = 0
+        while i < n:
+            if i in rep_map:
+                rep_end, rep_content = rep_map[i]
+                rep_str = f"<{rep_content}> [/]"
+                annotated_tokens.append(rep_str)
+                if rep_end in pause_map:
+                    for pause_marker in pause_map[rep_end]:
+                        annotated_tokens.append(pause_marker)
+                i = rep_end + 1
+            else:
+                token_word = words[i].get("word", "")
+                annotated_tokens.append(token_word)
+                if i in pause_map:
+                    for pause_marker in pause_map[i]:
+                        annotated_tokens.append(pause_marker)
+                i += 1
+        
+        # join all transcript
+        transcript = " ".join(annotated_tokens)
+        # gen *SPEAKER:\ttranscript
+        line = f"*{speaker}\t{transcript}"
+        annotated_lines.append(line)
+    
+    # write annotation_result.txt
+    with open(output_file, "w", encoding="utf-8") as f:
+        for line in annotated_lines:
+            f.write(line + "\n")
+    
+    print(f"Annotation done in {output_file}")
+    return output_file
+
+if __name__ == "__main__":
+
+    annotate_transcript("000030")

environment.yml

@@ -0,0 +1,166 @@
+name: SATE
+channels:
+  - conda-forge
+  - nvidia
+  - defaults
+  - https://repo.anaconda.com/pkgs/main
+  - https://repo.anaconda.com/pkgs/r
+dependencies:
+  - _libgcc_mutex=0.1
+  - _openmp_mutex=5.1
+  - bzip2=1.0.8
+  - ca-certificates=2025.2.25
+  - cudatoolkit=11.1.74
+  - cudnn=8.0.4
+  - ffmpeg=4.3.2
+  - freetype=2.10.4
+  - gmp=6.2.1
+  - gnutls=3.6.13
+  - lame=3.100
+  - ld_impl_linux-64=2.40
+  - libffi=3.4.4
+  - libgcc-ng=11.2.0
+  - libgomp=11.2.0
+  - libpng=1.6.37
+  - libstdcxx-ng=11.2.0
+  - libuuid=1.41.5
+  - ncurses=6.4
+  - nettle=3.6
+  - openh264=2.1.1
+  - openssl=3.0.16
+  - pip=25.0
+  - python=3.10.16
+  - readline=8.2
+  - setuptools=75.8.0
+  - sqlite=3.45.3
+  - tk=8.6.14
+  - wheel=0.45.1
+  - x264=1!161.3030
+  - xz=5.6.4
+  - zlib=1.2.13
+  - pip:
+      - accelerate==1.2.1
+      - aiohappyeyeballs==2.6.1
+      - aiohttp==3.11.14
+      - aiosignal==1.3.2
+      - alembic==1.15.1
+      - antlr4-python3-runtime==4.9.3
+      - asteroid-filterbanks==0.4.0
+      - async-timeout==5.0.1
+      - attrs==25.3.0
+      - av==14.2.0
+      - blinker==1.9.0
+      - certifi==2025.1.31
+      - cffi==1.17.1
+      - charset-normalizer==3.4.1
+      - click==8.1.8
+      - coloredlogs==15.0.1
+      - colorlog==6.9.0
+      - contourpy==1.3.1
+      - ctranslate2==4.4.0
+      - cycler==0.12.1
+      - docopt==0.6.2
+      - einops==0.8.1
+      - faster-whisper==1.1.0
+      - filelock==3.18.0
+      - flask==3.1.0
+      - flatbuffers==25.2.10
+      - fonttools==4.56.0
+      - frozenlist==1.5.0
+      - fsspec==2025.3.0
+      - greenlet==3.1.1
+      - huggingface-hub==0.29.3
+      - humanfriendly==10.0
+      - hyperpyyaml==1.2.2
+      - idna==3.10
+      - itsdangerous==2.2.0
+      - jinja2==3.1.6
+      - joblib==1.4.2
+      - julius==0.2.7
+      - kiwisolver==1.4.8
+      - lightning==2.5.1
+      - lightning-utilities==0.14.2
+      - mako==1.3.9
+      - markdown-it-py==3.0.0
+      - markupsafe==3.0.2
+      - matplotlib==3.10.1
+      - mdurl==0.1.2
+      - mpmath==1.3.0
+      - multidict==6.2.0
+      - networkx==3.4.2
+      - nltk==3.9.1
+      - numpy==2.0.2
+      - nvidia-cublas-cu12==12.4.5.8
+      - nvidia-cuda-cupti-cu12==12.4.127
+      - nvidia-cuda-nvrtc-cu12==12.4.127
+      - nvidia-cuda-runtime-cu12==12.4.127
+      - nvidia-cudnn-cu12==9.1.0.70
+      - nvidia-cufft-cu12==11.2.1.3
+      - nvidia-curand-cu12==10.3.5.147
+      - nvidia-cusolver-cu12==11.6.1.9
+      - nvidia-cusparse-cu12==12.3.1.170
+      - nvidia-cusparselt-cu12==0.6.2
+      - nvidia-nccl-cu12==2.21.5
+      - nvidia-nvjitlink-cu12==12.4.127
+      - nvidia-nvtx-cu12==12.4.127
+      - omegaconf==2.3.0
+      - onnxruntime==1.21.0
+      - optuna==4.2.1
+      - packaging==24.2
+      - pandas==2.2.3
+      - pillow==11.1.0
+      - primepy==1.3
+      - propcache==0.3.0
+      - protobuf==6.30.1
+      - psutil==7.0.0
+      - pyannote-audio==3.3.2
+      - pyannote-core==5.0.0
+      - pyannote-database==5.1.3
+      - pyannote-metrics==3.2.1
+      - pyannote-pipeline==3.0.1
+      - pycparser==2.22
+      - pydub==0.25.1
+      - pygments==2.19.1
+      - pyparsing==3.2.3
+      - python-dateutil==2.9.0.post0
+      - pytorch-lightning==2.5.1
+      - pytorch-metric-learning==2.8.1
+      - pytz==2025.2
+      - pyyaml==6.0.2
+      - regex==2024.11.6
+      - requests==2.32.3
+      - rich==13.9.4
+      - ruamel-yaml==0.18.10
+      - ruamel-yaml-clib==0.2.12
+      - safetensors==0.5.3
+      - scikit-learn==1.6.1
+      - scipy==1.15.2
+      - semver==3.0.4
+      - sentencepiece==0.2.0
+      - shellingham==1.5.4
+      - six==1.17.0
+      - sortedcontainers==2.4.0
+      - soundfile==0.12.1
+      - speechbrain==1.0.2
+      - sqlalchemy==2.0.39
+      - sympy==1.13.1
+      - tabulate==0.9.0
+      - tensorboardx==2.6.2.2
+      - threadpoolctl==3.6.0
+      - tokenizers==0.15.2
+      - torch==2.5.1
+      - torch-audiomentations==0.12.0
+      - torch-pitch-shift==1.2.5
+      - torchaudio==2.5.1
+      - torchmetrics==1.7.0
+      - tqdm==4.67.1
+      - transformers==4.37.2
+      - triton==3.1.0
+      - typer==0.15.2
+      - typing-extensions==4.12.2
+      - tzdata==2025.2
+      - urllib3==2.3.0
+      - werkzeug==3.1.3
+      - whisperx==3.3.1
+      - yarl==1.18.3
+prefix: /home/easgrad/shuweiho/workspace/miniconda3/envs/SATE

environment_sate_0.11.yml

@@ -0,0 +1,169 @@
+name: SATE
+channels:
+  - conda-forge
+  - nvidia
+  - defaults
+  - https://repo.anaconda.com/pkgs/main
+  - https://repo.anaconda.com/pkgs/r
+dependencies:
+  - _libgcc_mutex=0.1=main
+  - _openmp_mutex=5.1=1_gnu
+  - bzip2=1.0.8=h5eee18b_6
+  - ca-certificates=2025.2.25=h06a4308_0
+  - cudatoolkit=11.1.74=h6bb024c_0
+  - cudnn=8.0.4=cuda11.1_0
+  - ffmpeg=4.3.2=hca11adc_0
+  - freetype=2.10.4=h0708190_1
+  - gmp=6.2.1=h58526e2_0
+  - gnutls=3.6.13=h85f3911_1
+  - lame=3.100=h7f98852_1001
+  - ld_impl_linux-64=2.40=h12ee557_0
+  - libffi=3.4.4=h6a678d5_1
+  - libgcc-ng=11.2.0=h1234567_1
+  - libgomp=11.2.0=h1234567_1
+  - libpng=1.6.37=h21135ba_2
+  - libstdcxx-ng=11.2.0=h1234567_1
+  - libuuid=1.41.5=h5eee18b_0
+  - ncurses=6.4=h6a678d5_0
+  - nettle=3.6=he412f7d_0
+  - openh264=2.1.1=h780b84a_0
+  - openssl=3.0.16=h5eee18b_0
+  - pip=25.0=py310h06a4308_0
+  - python=3.10.16=he870216_1
+  - readline=8.2=h5eee18b_0
+  - setuptools=75.8.0=py310h06a4308_0
+  - sqlite=3.45.3=h5eee18b_0
+  - tk=8.6.14=h39e8969_0
+  - wheel=0.45.1=py310h06a4308_0
+  - x264=1!161.3030=h7f98852_1
+  - xz=5.6.4=h5eee18b_1
+  - zlib=1.2.13=h5eee18b_1
+  - pip:
+      - accelerate==1.2.1
+      - aiohappyeyeballs==2.6.1
+      - aiohttp==3.11.14
+      - aiosignal==1.3.2
+      - alembic==1.15.1
+      - antlr4-python3-runtime==4.9.3
+      - asteroid-filterbanks==0.4.0
+      - async-timeout==5.0.1
+      - attrs==25.3.0
+      - av==14.2.0
+      - blinker==1.9.0
+      - certifi==2025.1.31
+      - cffi==1.17.1
+      - charset-normalizer==3.4.1
+      - click==8.1.8
+      - coloredlogs==15.0.1
+      - colorlog==6.9.0
+      - contourpy==1.3.1
+      - ctranslate2==4.4.0
+      - cycler==0.12.1
+      - docopt==0.6.2
+      - einops==0.8.1
+      - faster-whisper==1.1.0
+      - filelock==3.18.0
+      - flask==3.1.0
+      - flatbuffers==25.2.10
+      - fonttools==4.56.0
+      - frozenlist==1.5.0
+      - fsspec==2025.3.0
+      - greenlet==3.1.1
+      - huggingface-hub==0.29.3
+      - humanfriendly==10.0
+      - hyperpyyaml==1.2.2
+      - idna==3.10
+      - imbalanced-learn==0.13.0
+      - itsdangerous==2.2.0
+      - jinja2==3.1.6
+      - joblib==1.4.2
+      - julius==0.2.7
+      - kiwisolver==1.4.8
+      - lightning==2.5.1
+      - lightning-utilities==0.14.2
+      - mako==1.3.9
+      - markdown-it-py==3.0.0
+      - markupsafe==3.0.2
+      - matplotlib==3.10.1
+      - mdurl==0.1.2
+      - mpmath==1.3.0
+      - multidict==6.2.0
+      - networkx==3.4.2
+      - nltk==3.9.1
+      - numpy==2.0.2
+      - nvidia-cublas-cu12==12.4.5.8
+      - nvidia-cuda-cupti-cu12==12.4.127
+      - nvidia-cuda-nvrtc-cu12==12.4.127
+      - nvidia-cuda-runtime-cu12==12.4.127
+      - nvidia-cudnn-cu12==9.1.0.70
+      - nvidia-cufft-cu12==11.2.1.3
+      - nvidia-curand-cu12==10.3.5.147
+      - nvidia-cusolver-cu12==11.6.1.9
+      - nvidia-cusparse-cu12==12.3.1.170
+      - nvidia-cusparselt-cu12==0.6.2
+      - nvidia-nccl-cu12==2.21.5
+      - nvidia-nvjitlink-cu12==12.4.127
+      - nvidia-nvtx-cu12==12.4.127
+      - omegaconf==2.3.0
+      - onnxruntime==1.21.0
+      - optuna==4.2.1
+      - packaging==24.2
+      - pandas==2.2.3
+      - pillow==11.1.0
+      - primepy==1.3
+      - propcache==0.3.0
+      - protobuf==6.30.1
+      - psutil==7.0.0
+      - pyannote-audio==3.3.2
+      - pyannote-core==5.0.0
+      - pyannote-database==5.1.3
+      - pyannote-metrics==3.2.1
+      - pyannote-pipeline==3.0.1
+      - pycparser==2.22
+      - pydub==0.25.1
+      - pygments==2.19.1
+      - pyparsing==3.2.3
+      - python-dateutil==2.9.0.post0
+      - pytorch-lightning==2.5.1
+      - pytorch-metric-learning==2.8.1
+      - pytz==2025.2
+      - pyyaml==6.0.2
+      - regex==2024.11.6
+      - requests==2.32.3
+      - rich==13.9.4
+      - ruamel-yaml==0.18.10
+      - ruamel-yaml-clib==0.2.12
+      - safetensors==0.5.3
+      - scikit-learn==1.6.1
+      - scipy==1.15.2
+      - semver==3.0.4
+      - sentencepiece==0.2.0
+      - shellingham==1.5.4
+      - six==1.17.0
+      - sklearn-compat==0.1.3
+      - sortedcontainers==2.4.0
+      - soundfile==0.12.1
+      - speechbrain==1.0.2
+      - sqlalchemy==2.0.39
+      - sympy==1.13.1
+      - tabulate==0.9.0
+      - tensorboardx==2.6.2.2
+      - threadpoolctl==3.6.0
+      - tokenizers==0.15.2
+      - torch==2.5.1
+      - torch-audiomentations==0.12.0
+      - torch-pitch-shift==1.2.5
+      - torchaudio==2.5.1
+      - torchmetrics==1.7.0
+      - tqdm==4.67.1
+      - transformers==4.37.2
+      - triton==3.1.0
+      - typer==0.15.2
+      - typing-extensions==4.12.2
+      - tzdata==2025.2
+      - urllib3==2.3.0
+      - werkzeug==3.1.3
+      - whisperx==3.3.1
+      - xgboost==3.0.0
+      - yarl==1.18.3
+prefix: /home/easgrad/shuweiho/workspace/miniconda3/envs/SATE

feature_extraction.py

@@ -0,0 +1,393 @@
+import os
+import json
+import re
+from collections import Counter
+
+def target_speaker(segments):
+    speakers = [seg.get("speaker", "UNKNOWN") for seg in segments]
+    most_common = Counter(speakers).most_common(1)
+    return most_common[0][0] if most_common else "UNKNOWN"
+
+def pause_feature(session_id, base_dir="session_data"):
+
+    json_file = os.path.join(base_dir, session_id, f"{session_id}_transcriptionCW.json")
+    feature_file = os.path.join(base_dir, session_id, f"{session_id}_feature.json")
+
+    if not os.path.exists(json_file):
+        print(f"[Error] transcriptionCW file not found: {json_file}")
+        return
+
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+
+    segments = data.get("segments", [])
+    tgt_speaker = target_speaker(segments)
+    tgt_segments = [seg for seg in segments if seg.get("speaker") == tgt_speaker]
+
+    total_words = 0
+    total_duration = 0.0
+    all_pauses = []
+
+    for seg in tgt_segments:
+        words = seg.get("words", [])
+        total_words += len(words)
+        duration = seg.get("end", 0) - seg.get("start", 0)
+        total_duration += duration
+        pauses = seg.get("pauses", [])
+        all_pauses.extend(pauses)
+
+    total_pauses = len(all_pauses)
+    pause_durations = [p["duration"] for p in all_pauses]
+    pause_total_duration = sum(pause_durations)
+    avg_pause_duration = sum(pause_durations) / total_pauses if total_pauses > 0 else 0
+    longest_pause = max(pause_durations) if pause_durations else 0
+
+    pause_density_1 = (total_pauses / total_words * 100) if total_words > 0 else 0
+    pause_density_2 = (total_pauses / total_duration * 60) if total_duration > 0 else 0
+    pause_proportion = (pause_total_duration / total_duration) * 100 if total_duration > 0 else 0
+
+    pause_feat = {
+        "Pause Density I": round(pause_density_1, 3),
+        "Pause Density II": round(pause_density_2, 3),
+        "Pause Proportion": round(pause_proportion, 3),
+        "Average Pause Duration": round(avg_pause_duration, 3),
+        "Longest Pause Duration": round(longest_pause, 3)
+    }
+
+    if os.path.exists(feature_file):
+        with open(feature_file, "r", encoding="utf-8") as f:
+            feature_data = json.load(f)
+    else:
+        feature_data = {}
+
+    feature_data.setdefault("pause", []).append(pause_feat)
+
+    with open(feature_file, "w", encoding="utf-8") as f:
+        json.dump(feature_data, f, indent=4, ensure_ascii=False)
+
+    print(f"[Done] Pause features written to {feature_file}")
+
+
+
+def get_syllable_weight(cv_pattern):
+    weight1_patterns = {'CV', 'CVC', 'VC', 'V'}
+    weight2_patterns = {'VCC', 'CVCC', 'CCV', 'CCVC'}
+    if cv_pattern in weight1_patterns:
+        return 1
+    elif cv_pattern in weight2_patterns:
+        return 10
+    else:
+        return 100
+        
+
+def syllable_feature(session_id, base_dir="session_data"):
+    json_file = os.path.join(base_dir, session_id, f"{session_id}_transcriptionCW.json")
+    feature_file = os.path.join(base_dir, session_id, f"{session_id}_feature.json")
+
+    if not os.path.exists(json_file):
+        print(f"[Error] transcriptionCW file not found: {json_file}")
+        return
+
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+
+    segments = data.get("segments", [])
+    tgt_speaker = target_speaker(segments)
+    tgt_segments = [seg for seg in segments if seg.get("speaker") == tgt_speaker]
+
+    total_duration = sum(seg["end"] - seg["start"] for seg in tgt_segments)
+    syllable_types = set()
+    total_syllables = 0
+    total_weight = 0
+    max_syllable_len = 0
+    word_syllable_count = {}
+    word_counter = 0 
+
+    for seg in tgt_segments:
+        syllables = seg.get("syllables", [])
+        for syl in syllables:
+            cv = syl.get("CV_pattern", "")
+            phonemes = syl.get("phonemes", [])
+            syllable_types.add(cv)
+            total_syllables += 1
+            total_weight += get_syllable_weight(cv)
+            max_syllable_len = max(max_syllable_len, len(phonemes))
+            
+            local_idx = syl.get("word_index")
+            global_idx = word_counter + local_idx
+            word_syllable_count[global_idx] = word_syllable_count.get(global_idx, 0) + 1
+        word_counter += len(seg.get("words", []))
+
+    all_words = []
+    for seg in tgt_segments:
+        all_words.extend(seg.get("words", []))
+    total_words = len(all_words)
+
+    multi_syll_2 = sum(1 for c in word_syllable_count.values() if c >= 2)
+    multi_syll_3 = sum(1 for c in word_syllable_count.values() if c >= 3)
+
+    feat = {
+        "Number of Syllable Types": len(syllable_types),
+        "Syllable Complexity Index": round(total_weight / total_syllables, 3) if total_syllables > 0 else 0,
+        "Average Syllable Rate": round(total_syllables / total_duration, 3) if total_duration > 0 else 0,
+        "Longest Syllable Length": max_syllable_len,
+        "Proportion of Multisyllabic Words I": round((multi_syll_2 / total_words) * 100, 3) if total_words > 0 else 0,
+        "Proportion of Multisyllabic Words II": round((multi_syll_3 / total_words) * 100, 3) if total_words > 0 else 0
+    }
+
+    if os.path.exists(feature_file):
+        with open(feature_file, "r", encoding="utf-8") as f:
+            feature_data = json.load(f)
+    else:
+        feature_data = {}
+
+    feature_data.setdefault("syllable", []).append(feat)
+
+    with open(feature_file, "w", encoding="utf-8") as f:
+        json.dump(feature_data, f, indent=4, ensure_ascii=False)
+
+    print(f"[Done] Syllable features written to {feature_file}")
+
+
+
+def get_rep_weight(length):
+    if length == 1:
+        return 1
+    elif length == 2:
+        return 5
+    elif length == 3:
+        return 10
+    elif length == 4:
+        return 20
+    else:
+        return 40
+
+
+def repetition_feature(session_id, base_dir="session_data"):
+    json_file = os.path.join(base_dir, session_id, f"{session_id}_transcriptionCW.json")
+    feature_file = os.path.join(base_dir, session_id, f"{session_id}_feature.json")
+
+    if not os.path.exists(json_file):
+        print(f"[Error] transcriptionCW file not found: {json_file}")
+        return
+
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+
+    segments = data.get("segments", [])
+    tgt_speaker = target_speaker(segments)
+    tgt_segments = [seg for seg in segments if seg.get("speaker") == tgt_speaker]
+
+    total_duration = sum(seg["end"] - seg["start"] for seg in tgt_segments)
+    total_words = sum(len(seg.get("words", [])) for seg in tgt_segments)
+
+    rep_count = 0
+    rep_weights = 0
+    rep_lengths = []
+    rep_durations = []
+
+    for seg in tgt_segments:
+        words = seg.get("words", [])
+        repetitions = seg.get("repetitions", [])
+        for rep in repetitions:
+            indices = rep.get("words", [])
+            if not indices:
+                continue
+            length = len(indices)
+            rep_lengths.append(length)
+            rep_weights += get_rep_weight(length)
+            rep_count += 1
+
+            # compute duration based on word timestamps
+            start_idx = indices[0]
+            end_idx = indices[-1]
+            if 0 <= start_idx < len(words) and 0 <= end_idx < len(words):
+                start_time = words[start_idx]["start"]
+                end_time = words[end_idx]["end"]
+                rep_durations.append(end_time - start_time)
+
+    total_rep_time = sum(rep_durations)
+    longest_rep = max(rep_lengths) if rep_lengths else 0
+    avg_rep_len = sum(rep_lengths) / len(rep_lengths) if rep_lengths else 0
+
+    rep_feat = {
+        "Word Repetition Index": round(rep_weights / rep_count, 3) if rep_count > 0 else 0,
+        "Repetition Density I": round(rep_count / total_words * 100, 3) if total_words > 0 else 0,
+        "Repetition Density II": round(rep_count / total_duration * 60, 3) if total_duration > 0 else 0,
+        "Repetition Proportion": round((total_rep_time / total_duration) * 100, 3) if total_duration > 0 else 0,
+        "Longest Repetition Length": longest_rep,
+        "Average Repetition Length": round(avg_rep_len, 3)
+    }
+
+    if os.path.exists(feature_file):
+        with open(feature_file, "r", encoding="utf-8") as f:
+            feature_data = json.load(f)
+    else:
+        feature_data = {}
+
+    feature_data.setdefault("repetition", []).append(rep_feat)
+
+    with open(feature_file, "w", encoding="utf-8") as f:
+        json.dump(feature_data, f, indent=4, ensure_ascii=False)
+
+    print(f"[Done] Repetition features written to {feature_file}")
+
+
+
+def fillerword_feature(session_id, base_dir="session_data"):
+    json_file = os.path.join(base_dir, session_id, f"{session_id}_transcriptionCW.json")
+    feature_file = os.path.join(base_dir, session_id, f"{session_id}_feature.json")
+
+    if not os.path.exists(json_file):
+        print(f"[Error] transcriptionCW file not found: {json_file}")
+        return
+
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+
+    segments = data.get("segments", [])
+    tgt_speaker = target_speaker(segments)
+    tgt_segments = [seg for seg in segments if seg.get("speaker") == tgt_speaker]
+
+    total_duration = sum(seg["end"] - seg["start"] for seg in tgt_segments)
+    total_words = sum(len(seg.get("words", [])) for seg in tgt_segments)
+
+    filler_total = 0
+    filler_durations = []
+    filler_intervals = []
+
+    for seg in tgt_segments:
+        fillers = seg.get("fillerwords", [])
+        words = seg.get("words", [])
+
+        filler_total += len(fillers)
+        filler_durations += [fw["duration"] for fw in fillers]
+
+        if len(fillers) >= 2:
+            indices = []
+            for fw in fillers:
+                for i, w in enumerate(words):
+                    if abs(w.get("start", 0) - fw["start"]) < 0.01:
+                        indices.append(i)
+                        break
+            indices.sort()
+            intervals = [indices[i+1] - indices[i] - 1 for i in range(len(indices)-1)]
+            if intervals:
+                filler_intervals.append(sum(intervals) / len(intervals))
+
+    avg_duration = sum(filler_durations) / len(filler_durations) if filler_durations else 0
+    longest_duration = max(filler_durations) if filler_durations else 0
+    avg_interval = sum(filler_intervals) / len(filler_intervals) if filler_intervals else 0
+
+    feat = {
+        "Filler Word Density I": round(filler_total / total_words * 100, 3) if total_words > 0 else 0,
+        "Filler Word Density II": round(filler_total / total_duration * 60, 3) if total_duration > 0 else 0,
+        "Filler Word Proportion": round((sum(filler_durations) / total_duration) * 100, 3) if total_duration > 0 else 0,
+        "Average Filler Word Duration": round(avg_duration, 3),
+        "Longest Filler Word Duration": round(longest_duration, 3),
+        "Average Filler Word Interval": round(avg_interval, 3)
+    }
+
+    if os.path.exists(feature_file):
+        with open(feature_file, "r", encoding="utf-8") as f:
+            feature_data = json.load(f)
+    else:
+        feature_data = {}
+
+    feature_data.setdefault("fillerword", []).append(feat)
+
+    with open(feature_file, "w", encoding="utf-8") as f:
+        json.dump(feature_data, f, indent=4, ensure_ascii=False)
+
+    print(f"[Done] Filler word features written to {feature_file}")
+
+
+def plm_feature(session_id, base_dir="session_data"):
+
+
+    json_file = os.path.join(base_dir, session_id, f"{session_id}_transcriptionCW.json")
+    feature_file = os.path.join(base_dir, session_id, f"{session_id}_feature.json")
+
+    if not os.path.exists(json_file):
+        print(f"[Error] transcriptionCW file not found: {json_file}")
+        return
+
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+
+    segments = data.get("segments", [])
+    tgt_speaker = target_speaker(segments)
+    tgt_segments = [seg for seg in segments if seg.get("speaker") == tgt_speaker]
+
+    sequence = ''.join(seg.get("mispronunciation", "") for seg in tgt_segments)
+    sequence = re.sub(r"[^CE]", "", sequence.upper())
+
+    n = len(sequence)
+    if n == 0:
+        feat = {
+            "Mispronunciation Density": 0,
+            "Normalized Transition Count": 0,
+            "Average Common Correct": 0,
+            "Average Common Error": 0,
+            "Longest Common Correct": 0,
+            "Longest Common Error": 0
+        }
+    else:
+        transitions = sum(1 for i in range(1, n) if sequence[i] != sequence[i - 1])
+        MPD = sum(1 for ch in sequence if ch == 'E') / n
+        NTC = transitions / n
+
+        def run_lengths(s, ch):
+            return [len(g) for g in re.findall(f"{ch}+", s)]
+
+        c_runs = run_lengths(sequence, 'C')
+        e_runs = run_lengths(sequence, 'E')
+
+        ACC = sum(c_runs) / len(c_runs) if c_runs else 0
+        ACE = sum(e_runs) / len(e_runs) if e_runs else 0
+        LCC = max(c_runs) if c_runs else 0
+        LCE = max(e_runs) if e_runs else 0
+
+        feat = {
+            "Mispronunciation Density": round(MPD, 3),
+            "Normalized Transition Count": round(NTC, 3),
+            "Average Common Correct": round(ACC, 3),
+            "Average Common Error": round(ACE, 3),
+            "Longest Common Correct": LCC,
+            "Longest Common Error": LCE
+        }
+
+    if os.path.exists(feature_file):
+        with open(feature_file, "r", encoding="utf-8") as f:
+            feature_data = json.load(f)
+    else:
+        feature_data = {}
+
+    feature_data.setdefault("plm", []).append(feat)
+
+    with open(feature_file, "w", encoding="utf-8") as f:
+        json.dump(feature_data, f, indent=4, ensure_ascii=False)
+
+    print(f"[Done] PLM features written to {feature_file}")
+
+
+
+def feature_extraction(session_id, base_dir="session_data"):
+
+    feature_file = os.path.join(base_dir, session_id, f"{session_id}_feature.json")
+
+    if not os.path.exists(feature_file):
+        with open(feature_file, "w", encoding="utf-8") as f:
+            json.dump({}, f, indent=4)
+
+    pause_feature(session_id, base_dir=base_dir)
+    syllable_feature(session_id, base_dir=base_dir)
+    repetition_feature(session_id, base_dir=base_dir)
+    fillerword_feature(session_id, base_dir=base_dir)
+    plm_feature(session_id, base_dir=base_dir)
+
+    print(f"[All Analysis Done] Feature extraction complete for session {session_id}")
+
+
+if __name__ == "__main__":
+    feature_extraction("000030")

fillerword.py

@@ -0,0 +1,37 @@
+import os
+import json
+import re
+
+def annotate_fillerwords(session_id, base_dir="session_data"):
+    session_dir = os.path.join(base_dir, session_id)
+    json_file = os.path.join(session_dir, f"{session_id}_transcriptionCW.json")
+
+    if not os.path.exists(json_file):
+        print(f"[Error] File not found: {json_file}")
+        return
+
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+
+    for segment in data.get("segments", []):
+        words = segment.get("words", [])
+        fillerwords = []
+        for w in words:
+            word_content = w.get("word", "").strip()
+            if re.fullmatch(r"\[.*?\]", word_content):
+                fillerwords.append({
+                    "start": w.get("start"),
+                    "end": w.get("end"),
+                    "content": word_content,
+                    "duration": round(w.get("end", 0) - w.get("start", 0), 3)
+                })
+        segment["fillerwords"] = fillerwords
+
+    with open(json_file, "w", encoding="utf-8") as f:
+        json.dump(data, f, ensure_ascii=False, indent=4)
+
+    print(f"Session {session_id} fillerword annotation done: {json_file}")
+    return data
+
+if __name__ == "__main__":
+    annotate_fillerwords("000003")

main_socket.py

@@ -0,0 +1,91 @@
+import os
+import tempfile
+import json
+from flask import Flask, request, jsonify
+from werkzeug.utils import secure_filename
+
+from preprocess import process_audio_file
+from pause import annotate_pauses
+from repetition import annotate_repetitions
+from syllable import annotate_syllables
+from fillerword import annotate_fillerwords
+
+from annotation import annotate_transcript
+
+
+
+app = Flask(__name__)
+
+@app.route('/process_old', methods=['POST'])
+def process_audio_old():
+    data = request.get_json()
+    if not data or 'input_audio_file' not in data:
+        return jsonify({'error': 'Missing input_audio_file parameter'}), 400
+
+    input_audio_file = data['input_audio_file']
+    device = data.get('device', 'cuda')
+    pause_threshold = data.get('pause_threshold', 0.5)
+    num_speakers = data.get('num_speakers', 2)
+
+    app.logger.info(f"Processing audio file: {input_audio_file}")
+
+    session_id = process_audio_file(input_audio_file, num_speakers=num_speakers, device=device)
+    annotate_pauses(session_id, pause_threshold)
+    annotate_repetitions(session_id)
+    annotate_syllables(session_id)
+    annotate_fillerwords(session_id)
+
+    output_annotation = annotate_transcript(session_id)
+
+    result = {
+        'session_id': session_id,
+        'annotation_result': output_annotation
+    }
+    return jsonify(result), 200
+
+
+
+@app.route('/process', methods=['POST'])
+def process_audio():
+    if 'audio_file' not in request.files:
+        return jsonify({'error': 'Missing audio file '}), 400
+    audio_file = request.files['audio_file']
+    filename = secure_filename(audio_file.filename)
+    
+    suffix = os.path.splitext(filename)[1] or '.wav'
+    with tempfile.NamedTemporaryFile(delete=False, suffix=suffix) as tmp:
+        audio_path = tmp.name
+        audio_file.save(audio_path)
+
+    device = request.form.get('device', 'cuda')
+    pause_threshold = float(request.form.get('pause_threshold', 0.5))
+    num_speakers = int(request.form.get('num_speakers', 2))
+
+    app.logger.info(f"Processing uploaded audio: {audio_path}")
+
+    session_id = process_audio_file(audio_path, num_speakers=num_speakers, device=device)
+    annotate_pauses(session_id, pause_threshold)
+    annotate_repetitions(session_id)
+    annotate_syllables(session_id)
+    annotate_fillerwords(session_id)
+    # annotate_transcript(session_id)
+
+
+    json_path = f"/sate/session_data/{session_id}/{session_id}_transcriptionCW.json"
+    if not os.path.isfile(json_path):
+        return jsonify({'error': f"Annotation file {json_path} not found"}), 500
+
+    with open(json_path, 'r', encoding='utf-8') as f:
+        transcription = json.load(f)
+
+
+    try:
+        os.remove(audio_path)
+    except OSError:
+        pass
+
+    return jsonify(transcription), 200
+
+
+if __name__ == "__main__":
+    app.run(host="0.0.0.0", port=5000, debug=True)

mispronunciation.py

@@ -0,0 +1,67 @@
+import os
+import json
+import base64
+import requests
+
+def wav_to_base64_url(wav_path):
+    with open(wav_path, "rb") as f:
+        b64_audio = base64.b64encode(f.read()).decode()
+    return f"data:audio/wav;base64,{b64_audio}"
+
+def is_api_available(api_url):
+    try:
+        response = requests.get(api_url)
+        return response.status_code < 500
+    except Exception as e:
+        print("[Error] Cannot reach API:", e)
+        return False
+
+def annotate_mispronunciation(session_id, api_url="http://localhost:8080", base_dir="session_data"):
+    json_path = os.path.join(base_dir, session_id, f"{session_id}_transcriptionCW.json")
+    if not os.path.exists(json_path):
+        print(f"[Error] File not found: {json_path}")
+        return
+
+    if not is_api_available(api_url):
+        print(f"[Error] API not available at {api_url}")
+        return
+
+    with open(json_path, "r", encoding="utf-8") as f:
+        data = json.load(f)
+
+    segments = data.get("segments", [])
+    for segment in segments:
+        start = segment["start"]
+        end = segment["end"]
+        speaker = segment["speaker"]
+        filename = f"{session_id}-{start:.2f}-{end:.2f}-{speaker}.wav"
+        filepath = os.path.join(base_dir, session_id, filename)
+        if not os.path.exists(filepath):
+            print(f"[Warning] Audio file missing: {filename}")
+            continue
+
+        audio_url = wav_to_base64_url(filepath)
+        try:
+            resp = requests.post(
+                f"{api_url}/vocallens/api/analyze",
+                headers={"Content-Type": "application/json"},
+                data=json.dumps([audio_url])
+            )
+            if resp.status_code == 200:
+                result = resp.json()
+                segment["mispronunciation"] = result.get("ce", "")
+            else:
+                print(f"[Warning] API error {resp.status_code} for {filename}")
+                segment["mispronunciation"] = ""
+        except Exception as e:
+            print(f"[Error] Exception during API call for {filename}:", e)
+            segment["mispronunciation"] = ""
+
+    with open(json_path, "w", encoding="utf-8") as f:
+        json.dump(data, f, indent=4, ensure_ascii=False)
+    
+    print(f"Session {session_id} mispronunciation annotation done: {json_path}")
+    return data
+
+if __name__ == "__main__":
+    annotate_mispronunciation("000030")

pause.py

@@ -0,0 +1,44 @@
+import os
+import json
+
+def annotate_pauses(session_id, threshold, base_dir="session_data"):
+
+    session_dir = os.path.join(base_dir, session_id)
+    json_file = os.path.join(session_dir, f"{session_id}_transcriptionCW.json")
+    
+    if not os.path.exists(json_file):
+        print(f"Error: could not finf {json_file}")
+        return
+    
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+    
+    segments = data.get("segments", [])
+    for segment in segments:
+        words = segment.get("words", [])
+        if "pauses" in segment:
+            del segment["pauses"]
+            
+        pauses = []
+        if words and len(words) > 1:
+            for i in range(1, len(words)):
+                prev_word = words[i - 1]
+                current_word = words[i]
+                gap = current_word["start"] - prev_word["end"]
+                if gap > threshold:
+                    pause_info = {
+                        "start": round(prev_word["end"], 3),
+                        "end": round(current_word["start"], 3),
+                        "duration": round(gap, 3)
+                    }
+                    pauses.append(pause_info)
+        segment["pauses"] = pauses
+    
+    with open(json_file, "w", encoding="utf-8") as f:
+        json.dump(data, f, ensure_ascii=False, indent=4)
+    
+    print(f"Session {session_id} pause annotation done: {json_file}")
+    return data
+
+if __name__ == "__main__":
+    annotated_data = annotate_pauses("000030", 0.1)

preprocess.py

@@ -0,0 +1,241 @@
+import os
+import json
+from pathlib import Path
+import whisperx
+import soundfile as sf
+import numpy as np
+import re
+import torch
+from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
+import sys
+
+from dotenv import load_dotenv
+load_dotenv()
+token = os.getenv("HF_TOKEN")
+
+print("Start Preprocessing ... ...")
+
+sys.path.append('./CrisperWhisper/')
+from utils import adjust_pauses_for_hf_pipeline_output
+
+def generate_session_id():
+    session_root = "session_data"
+    if not os.path.exists(session_root):
+        os.makedirs(session_root)
+        return "000001"
+    
+    existing_ids = [d for d in os.listdir(session_root)
+                    if os.path.isdir(os.path.join(session_root, d)) and d.isdigit()]
+    if existing_ids:
+        new_id = max(int(x) for x in existing_ids) + 1
+    else:
+        new_id = 1
+    return f"{new_id:06d}"
+
+def assign_speakers(segments, diarization_segments):
+    speaker_map = {}
+    for segment in segments:
+        segment_start = segment["start"]
+        segment_end = segment["end"]
+        max_overlap = 0
+        assigned_speaker = "Unknown"
+        
+        for _, diar in diarization_segments.iterrows():
+            speaker = diar["speaker"]
+            diar_start = diar["start"]
+            diar_end = diar["end"]
+            overlap_start = max(segment_start, diar_start)
+            overlap_end = min(segment_end, diar_end)
+            overlap_duration = max(0, overlap_end - overlap_start)
+            if overlap_duration > max_overlap:
+                max_overlap = overlap_duration
+                assigned_speaker = speaker
+        
+        speaker_map[segment_start] = assigned_speaker
+    return speaker_map
+
+def load_audio_for_split(input_audio_file):
+
+    if input_audio_file.lower().endswith('.mp3'):
+        from pydub import AudioSegment
+        audio_seg = AudioSegment.from_file(input_audio_file)
+        sr = audio_seg.frame_rate
+        samples = np.array(audio_seg.get_array_of_samples()).astype(np.float32)
+        samples = samples / 32768.0
+        if audio_seg.channels > 1:
+            samples = samples.reshape((-1, audio_seg.channels))
+        return samples, sr
+    else:
+        return sf.read(input_audio_file)
+
+def process_audio_file(input_audio_file, num_speakers, device="cuda"):
+    
+    print("Loading WhisperX model (English)...")
+    model = whisperx.load_model("medium", device, language="en")
+    
+    audio = whisperx.load_audio(input_audio_file)
+    
+    print("Transcribing audio with WhisperX...")
+    result = model.transcribe(audio)
+    
+    print("Performing forced alignment with WhisperX...")
+    alignment_model, metadata = whisperx.load_align_model(language_code="en", device=device)
+    result_aligned = whisperx.align(result["segments"], alignment_model, metadata, audio, device, return_char_alignments=True)
+    
+    print("Detecting speakers with WhisperX...")
+    diarization_model = whisperx.DiarizationPipeline(use_auth_token=token,
+                                                     device=device)
+    diarization_segments = diarization_model(audio)
+    
+    speaker_map = assign_speakers(result_aligned["segments"], diarization_segments)
+    for segment in result_aligned["segments"]:
+        segment["speaker"] = speaker_map.get(segment["start"], "Unknown")
+        segment.pop("chars", None)
+
+    session_id = generate_session_id()
+    session_dir = os.path.join("session_data", session_id)
+    os.makedirs(session_dir, exist_ok=True)
+    
+    data, sr = load_audio_for_split(input_audio_file)
+    
+    for segment in result_aligned["segments"]:
+        start_time = segment["start"]
+        end_time = segment["end"]
+        speaker = segment["speaker"]
+        start_sample = int(start_time * sr)
+        end_sample = int(end_time * sr)
+        segment_audio = data[start_sample:end_sample]
+        
+        segment_filename = f"{session_id}-{start_time:.2f}-{end_time:.2f}-{speaker}.wav"
+        segment_filepath = os.path.join(session_dir, segment_filename)
+        sf.write(segment_filepath, segment_audio, sr)
+        print(f"Saved segment: {segment_filepath}")
+    
+
+    transcript_path = os.path.join(session_dir, f"{session_id}_transcription.txt")
+    with open(transcript_path, "w", encoding="utf-8") as f:
+        for segment in result_aligned["segments"]:
+            f.write(f"[{segment['start']} - {segment['end']}] (Speaker {segment['speaker']}): {segment['text']}\n")
+    
+
+    del model
+    torch.cuda.empty_cache()
+    
+
+    print("Loading CrisperWhisper model...")
+    device_str = "cuda:0" if torch.cuda.is_available() else "cpu"
+    torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
+    local_model_dir = "./CrisperWhisper_local"
+
+    cw_model = AutoModelForSpeechSeq2Seq.from_pretrained(
+        local_model_dir,
+        torch_dtype=torch_dtype,
+        low_cpu_mem_usage=True,
+        use_safetensors=True
+    )
+    cw_model.to(device_str)
+    
+    processor = AutoProcessor.from_pretrained(local_model_dir)
+    
+    asr_pipeline = pipeline(
+        "automatic-speech-recognition",
+        model=cw_model,
+        tokenizer=processor.tokenizer,
+        feature_extractor=processor.feature_extractor,
+        chunk_length_s=30,         
+        batch_size=4,
+        return_timestamps='word',  
+        torch_dtype=torch_dtype,
+        device=0 if torch.cuda.is_available() else -1,
+        generate_kwargs={"language": "en"}
+    )
+    
+    segments_cw = []
+    skipped_segments = []
+    segment_files = [f for f in os.listdir(session_dir) if f.endswith('.wav')]
+    for seg_file in sorted(segment_files):
+
+        match = re.match(r'^(\d+)-(\d+\.\d+)-(\d+\.\d+)-(.+)\.wav$', seg_file)
+        if not match:
+            continue
+
+        seg_session_id = match.group(1)
+        start_time = float(match.group(2))
+        end_time = float(match.group(3))
+        speaker = match.group(4)
+        seg_path = os.path.join(session_dir, seg_file)
+        
+        print(f"Processing segment with CrisperWhisper: {seg_path}")
+        try:
+            cw_output = asr_pipeline(seg_path)
+            cw_result = adjust_pauses_for_hf_pipeline_output(cw_output)
+        except Exception as e:
+            print(f"[Warning] CrisperWhisper error, skiped this segment: {seg_path}\nError Message: {e}")
+            skipped_segments.append(seg_path)
+            continue
+        
+        text = cw_result.get('text', '').strip()
+        if not text:
+            print(f"********** No text returned, skiped this segment: {seg_path} **********")
+            skipped_segments.append(seg_path)
+            continue
+        
+        chunks = cw_result.get('chunks', [])
+        words_info = []
+        for i, chunk in enumerate(chunks):
+            word_text = chunk['text'].strip()
+            if not word_text:
+                continue
+
+            chunk_start, chunk_end = chunk['timestamp']
+            
+            if chunk_start is None:
+                if i == 0:
+                    chunk_start = 0.0
+                else:
+                    chunk_start = words_info[-1]['end'] - start_time
+
+            if chunk_end is None:
+                if i < len(chunks) - 1:
+                    next_chunk_start, _ = chunks[i+1]['timestamp']
+                    if next_chunk_start is None:
+                        next_chunk_start = chunk_start
+                    chunk_end = next_chunk_start
+                else:
+                    chunk_end = end_time - start_time 
+
+            word_start = round(start_time + chunk_start, 3)
+            word_end = round(start_time + chunk_end, 3)
+            words_info.append({
+                "word": word_text,
+                "start": word_start,
+                "end": word_end
+            })
+        
+        segment_entry = {
+            "start": round(start_time, 3),
+            "end": round(end_time, 3),
+            "speaker": speaker,
+            "text": text,
+            "words": words_info
+        }
+        segments_cw.append(segment_entry)
+    
+    segments_cw = sorted(segments_cw, key=lambda x: x["start"])
+    cw_json_path = os.path.join(session_dir, f"{session_id}_transcriptionCW.json")
+    with open(cw_json_path, "w", encoding="utf-8") as f:
+        json.dump({"segments": segments_cw}, f, ensure_ascii=False, indent=4)
+    print(f"CrisperWhisper transcription saved to: {cw_json_path}")
+    
+    if skipped_segments:
+        skipped_file = os.path.join(session_dir, "skipped_segments.txt")
+        with open(skipped_file, "w", encoding="utf-8") as f:
+            for s in sorted(skipped_segments):
+                f.write(s + "\n")
+        print(f"Skipped segments recorded in: {skipped_file}")
+    
+    return session_id
+
+if __name__ == "__main__":
+    session = process_audio_file("/home/easgrad/shuweiho/workspace/volen/SATE_docker_test/input/454.mp3", num_speakers=2, device="cuda")
+    print("Processing complete. Session ID:", session)

readme.md

@@ -0,0 +1,53 @@
+This is the SATE MVP, integrate the all the pipelines into one framework.
+Contain the main entrance, build for docker.
+
+main.py:
+Input: Entire audio file
+Output: Transcription with annotation
+
+Preprocess:
+Segmentation + speaker diarization -> crisper whisper transcriptions for each segmentation
+
+
+P.S. Should keep transcript consist in each pipelines.
+
+
+
+IMAGE CREATION:
+
+docker build -t sate_0.11 .
+
+
+(New) HOW TO USE after image created:
+
+docker run --gpus all -it --rm \
+  -v /home/easgrad/shuweiho/workspace/volen/SATE_docker_test/input:/sate/input \
+  -v /home/easgrad/shuweiho/workspace/volen/SATE_docker_test/session_data:/sate/session_data \
+  -p 5000:5000 \
+  sate_0.11
+
+
+curl -X POST http://localhost:5000/process \
+  -F "audio_file=@/home/easgrad/shuweiho/workspace/volen/SATE_docker_test/input/454.mp3" \
+  -F "device=cuda" \
+  -F "pause_threshold=0.25"
+
+
+
+
+(Old - don't follow it) HOW TO USE after image created:
+
+docker run --gpus all -it --rm \
+  -v /home/easgrad/shuweiho/workspace/volen/SATE_docker_test/input:/sate/input \
+  -v /home/easgrad/shuweiho/workspace/volen/SATE_docker_test/session_data:/sate/session_data \
+  -p 5000:5000 \
+  sate_0.10
+
+
+curl -X POST http://localhost:5000/process \
+     -H "Content-Type: application/json" \
+     -d '{
+           "input_audio_file": "/sate/input/454.mp3",
+           "device": "cuda",
+           "pause_threshold": 0.5
+         }'

repetition.py

@@ -0,0 +1,57 @@
+import os
+import json
+
+def annotate_repetitions(session_id, base_dir="session_data"):
+
+    session_dir = os.path.join(base_dir, session_id)
+    json_file = os.path.join(session_dir, f"{session_id}_transcriptionCW.json")
+    
+    if not os.path.exists(json_file):
+        print(f"Error: could not find {json_file}")
+        return
+    
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+    
+    segments = data.get("segments", [])
+    for segment in segments:
+        if "repetitions" in segment:
+            del segment["repetitions"]
+            
+        words_list = segment.get("words", [])
+        tokens = [w.get("word", "") for w in words_list]
+        reps = []
+        i = 0
+        n = len(tokens)
+        while i < n:
+            found = False
+            maxL = (n - i) // 2
+            for L in range(maxL, 0, -1):
+                if tokens[i:i+L] == tokens[i+L:i+2*L]:
+                    count = 2
+                    while i + count * L <= n and tokens[i:i+L] == tokens[i+(count-1)*L:i+count*L]:
+                        count += 1
+                    count -= 1 
+                    
+                    rep_count = count - 1
+                    rep_obj = {
+                        "content": " ".join(tokens[i:i+L] * rep_count),
+                        "words": list(range(i, i + rep_count * L)),
+                        "mark_location": i + rep_count * L - 1
+                    }
+                    reps.append(rep_obj)
+                    i += count * L 
+                    found = True
+                    break
+            if not found:
+                i += 1
+        segment["repetitions"] = reps
+    
+    with open(json_file, "w", encoding="utf-8") as f:
+        json.dump(data, f, ensure_ascii=False, indent=4)
+    
+    print(f"Session {session_id} repetition annotation done: {json_file}")
+    return data
+
+if __name__ == "__main__":
+    annotate_repetitions("000030")

syllable.py

@@ -0,0 +1,98 @@
+import os
+import json
+import re
+import string
+
+# Load Dict
+def load_custom_dict(dict_path):
+    with open(dict_path, 'r', encoding='utf-8') as f:
+        return json.load(f)
+
+custom_dict_path = "./syllable_dict_ENNI_refine.json"
+custom_dict = load_custom_dict(custom_dict_path)
+
+vowels_phonemes = [
+    "iː", "uː", "ɜː", "ɔː", "ɑː",
+    "ɪ", "ʊ", "e", "ə", "æ", "ʌ", "ɛ", "ɒ",
+    "eɪ", "aɪ", "ɔɪ", "aʊ", "əʊ", "ɪə", "eə", "ʊə"
+]
+
+def phoneme_type(phoneme):
+    return 'V' if phoneme in vowels_phonemes else 'C'
+
+def get_pronunciation_from_dict(word):
+    clean_word = word.strip(string.punctuation).lower()
+    return custom_dict.get(clean_word, "")
+
+def split_ipa_into_syllables(ipa_str):
+    ipa_str = ipa_str.replace("ˈ", ".").replace("ˌ", ".")
+    return [s for s in ipa_str.split('.') if s.strip()]
+
+def split_syllable_into_phonemes(syllable):
+    vowels_sorted = sorted(vowels_phonemes, key=len, reverse=True)
+    phonemes = []
+    i = 0
+    while i < len(syllable):
+        matched = None
+        for v in vowels_sorted:
+            if syllable[i:i+len(v)] == v:
+                matched = v
+                break
+        if matched:
+            phonemes.append(matched)
+            i += len(matched)
+        else:
+            phonemes.append(syllable[i])
+            i += 1
+    return phonemes
+
+def analyze_word_syllables(word):
+    ipa_str = get_pronunciation_from_dict(word)
+    if not ipa_str:
+        return []
+    syllables_ipa = split_ipa_into_syllables(ipa_str)
+    syllable_data = []
+    for syl in syllables_ipa:
+        phs = split_syllable_into_phonemes(syl)
+        CV = ''.join(phoneme_type(p) for p in phs)
+        syllable_data.append({
+            "syllable": ''.join(phs),
+            "phonemes": phs,
+            "CV_pattern": CV
+        })
+    return syllable_data
+
+def annotate_syllables(session_id, base_dir="session_data"):
+    json_file = os.path.join(base_dir, session_id, f"{session_id}_transcriptionCW.json")
+    if not os.path.exists(json_file):
+        print(f"[Error] Cannot find file: {json_file}")
+        return
+
+    with open(json_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+
+    for segment in data.get("segments", []):
+        text = segment.get("text", "")
+        words_info = segment.get("words", [])
+        syllables = []
+
+        for idx, word_obj in enumerate(words_info):
+            word = word_obj.get("word", "")
+            if re.fullmatch(r"\[.*?\]", word):  # 跳过 filler
+                continue
+            word_syllables = analyze_word_syllables(word)
+            for syl in word_syllables:
+                syl["word_index"] = idx
+            syllables.extend(word_syllables)
+
+        segment["syllables"] = syllables
+
+    with open(json_file, "w", encoding="utf-8") as f:
+        json.dump(data, f, indent=4, ensure_ascii=False)
+
+    print(f"Session {session_id} syllable annotation done: {json_file}")
+    return data
+
+
+if __name__ == "__main__":
+    annotate_syllables("000030")

syllable_dict_ENNI_refine.json

@@ -0,0 +1,3793 @@
+{
+    "A": "eɪ",
+    "ABC": "ˌeɪ.biːˈsiː",
+    "ARGH": "",
+    "Aaron": "",
+    "About": "əˈbaʊt",
+    "Accident": "ˈæk.sɪ.dənt",
+    "Accidentally": "ˌæk.sɪˈden.təl.i",
+    "Actually": "ˈæk.tʃu.ə.li",
+    "Adam": "ˈæd.əm",
+    "Adamor": "",
+    "Addison": "ˈæd.ɪ.sənz dɪˌziːz",
+    "Adia": "",
+    "Admiral": "ˈæd.mər.əl",
+    "African": "ˈæf.rɪ.kən",
+    "After": "ˈɑːf.tər",
+    "Again": "əˈɡen",
+    "Ah": "ɑː",
+    "Aha": "ɑːˈhɑː",
+    "Airpin": "",
+    "Airplane": "ˈeə.pleɪn",
+    "Alana": "",
+    "Albert": "",
+    "Alex": "",
+    "Alfie": "",
+    "Alfred": "",
+    "Ali": "",
+    "Alison": "",
+    "All": "ɔːl",
+    "Alley": "ˈæl.i",
+    "Allison": "",
+    "Almost": "ˈɔːl.məʊst",
+    "Alon": "",
+    "Already": "ɔːlˈred.i",
+    "Alright": "ɔːlˈraɪt",
+    "Also": "ˈɔːl.səʊ",
+    "Although": "ɔːlˈðəʊ",
+    "Alvin": "",
+    "Alway": "ˈɔːl.weɪ",
+    "Amanda": "",
+    "Amber": "ˈæm.bər",
+    "An": "æn",
+    "And": "ænd",
+    "Andra": "",
+    "Andrea": "",
+    "Andrew": "",
+    "Angelina": "",
+    "Animal": "ˈæn.ɪ.məl",
+    "Ann": "",
+    "Anna": "",
+    "Annette": "",
+    "Annie": "",
+    "Another": "əˈnʌð.ər",
+    "Anthony": "",
+    "Anton": "",
+    "Anybody": "ˈen.iˌbɒd.i",
+    "Anyone": "ˈen.i.wʌn",
+    "Anything": "ˈen.i.θɪŋ",
+    "Anyway": "ˈen.i.weɪ",
+    "Apple": "ˈæp.əl",
+    "April": "ˈeɪ.prəl",
+    "Are": "ɑːr",
+    "Arlene": "",
+    "Ashley": "",
+    "Ask": "ɑːsk",
+    "Asked": "ɑːsk",
+    "Askin": "",
+    "At": "æt",
+    "Aunt": "ɑːnt",
+    "Avalon": "",
+    "Average": "ˈæv.ər.ɪdʒ",
+    "Aw": "ɔː",
+    "Awesome": "ˈɔː.səm",
+    "Aww": "ɔː",
+    "BOOM": "buːm",
+    "Bab": "bæb",
+    "Back": "bæk",
+    "Bake": "beɪk",
+    "Ball": "bɔːl",
+    "Balloon": "bəˈluːn",
+    "Ballooner": "",
+    "Band": "bænd",
+    "Band-Aid": "ˈbænd.eɪd",
+    "Bang": "bæŋ",
+    "Barbera": "",
+    "Basket": "ˈbɑː.skɪt",
+    "Be": "biː",
+    "Bea": "",
+    "Bec": "",
+    "Because": "bɪˈkəz",
+    "Before": "bɪˈfɔːr",
+    "Bell": "bel",
+    "Belvedere": "ˈbel.və.dɪər",
+    "Ben": "ben",
+    "Benny": "ˈben.i",
+    "Beth": "",
+    "Bianca": "",
+    "Big": "bɪɡ",
+    "Bill": "bɪl",
+    "Billy": "ˈbɪl.i",
+    "Bit": "bɪt",
+    "Blake": "",
+    "Ble": "",
+    "Bloom": "bluːm",
+    "Bloomsailor": "",
+    "Blue": "bluː",
+    "Bob": "bɒb",
+    "Body": "ˈbɒd.i",
+    "Boing": "",
+    "Bonnie": "",
+    "Book": "bʊk",
+    "Boom": "buːm",
+    "Boon": "buːn",
+    "Boop": "buːp",
+    "Both": "bəʊθ",
+    "Bounce": "baʊns",
+    "Bow": "baʊ",
+    "Boy": "bɔɪ",
+    "Brace": "breɪs",
+    "Braden": "",
+    "Brady": "",
+    "Braeden": "",
+    "Bray": "breɪ",
+    "Break": "breɪk",
+    "Brent": "",
+    "Brett": "",
+    "Brianna": "",
+    "Bringing": "brɪŋ",
+    "Brittany": "",
+    "Broke": "brəʊk",
+    "Brought": "brɔːt",
+    "Bryce": "ˌbraɪs ˌkæn.jən ˌnæʃ.ən.əl ˈpɑːk",
+    "Buddy": "ˈbʌd.i",
+    "Bull": "bʊl",
+    "Bunch": "bʌntʃ",
+    "Bunny": "ˈbʌn.i",
+    "Bushabai": "",
+    "But": "bʌt",
+    "Butt": "bʌt",
+    "Buying": "baɪ",
+    "By": "baɪ",
+    "Bye": "baɪ",
+    "Bye-bye": "baɪ",
+    "C": "siː",
+    "CC": "ˌsiːˈsiː",
+    "Caitlin": "",
+    "Caitlyn": "",
+    "Cal": "kæl",
+    "Caleb": "",
+    "Call": "kɔːl",
+    "Calm": "kɑːm",
+    "Calvin": "",
+    "Came": "keɪm",
+    "Cameron": "",
+    "Can": "kæn",
+    "Can't": "kɑːnt",
+    "Captain": "ˈkæp.tɪn",
+    "Careful": "ˈkeə.fəl",
+    "Carl": "",
+    "Carrick": "",
+    "Carrot": "ˈkær.ət",
+    "Cash": "kæʃ",
+    "Cast": "kɑːst",
+    "Castle": "ˈkɑː.səl",
+    "Casual": "ˈkæʒ.ju.əl",
+    "Cat": "kæt",
+    "Catchy": "ˈkætʃ.i",
+    "Catherine": "ˈkæθ.rɪn ˌwiːl",
+    "Cause": "kɔːz",
+    "Causing": "kɔːz",
+    "Chadiah": "",
+    "Challenge": "ˈtʃæl.ɪndʒ",
+    "Check": "tʃek",
+    "Checked": "tʃekt",
+    "Cheer": "tʃɪər",
+    "Chela": "",
+    "Chelsea": "ˌtʃel.si ˈbʌn",
+    "Cheyla": "",
+    "Chose": "tʃəʊz",
+    "Chri": "",
+    "Christina": "",
+    "Christma": "",
+    "Christopher": "",
+    "Chrysantha": "",
+    "Chug": "tʃʌɡ",
+    "Clean": "kliːn",
+    "Coach": "kəʊtʃ",
+    "Cody": "",
+    "Colby": "ˈkɒl.bi",
+    "Come": "kʌm",
+    "Cone": "kəʊn",
+    "Connie": "",
+    "Connor": "",
+    "Cook": "kʊk",
+    "Cookie": "ˈkʊk.i",
+    "Cooking": "ˈkʊk.ɪŋ",
+    "Cool": "kuːl",
+    "Copper": "ˈkɒp.ər",
+    "Corbin": "",
+    "Could": "kʊd",
+    "Couldn't": "ˈkʊd.ənt",
+    "Courtney": "",
+    "Cow": "kaʊ",
+    "Crabby": "ˈkræb.i",
+    "Crisatha": "",
+    "Cron": "",
+    "Crying": "ˈkraɪ.ɪŋ",
+    "Crystal": "ˈkrɪs.təl",
+    "Cure": "kjʊər",
+    "Curti": "",
+    "Dad": "dæd",
+    "Daddy": "ˈdæd.i",
+    "Daffron": "",
+    "Damien": "",
+    "Dan": "dæn",
+    "Dana": "",
+    "Danciba": "",
+    "Danciber": "",
+    "Dang": "dæŋ",
+    "Daniel": "",
+    "Danielle": "",
+    "Darlene": "",
+    "Darlin": "",
+    "Darling": "ˈdɑː.lɪŋ",
+    "Darryl": "",
+    "Darwin": "ˈdɑː.wɪn",
+    "Daryl": "",
+    "Dave": "",
+    "Davern": "",
+    "David": "ˌstɑː.r əv ˈdeɪ.vɪd",
+    "Davin": "",
+    "Davrin": "",
+    "Dawn": "dɔːn",
+    "Debbie": "",
+    "Demi": "ˈdem.i-",
+    "Devin": "",
+    "Devon": "ˈdev.ən",
+    "Diba": "",
+    "Did": "dɪd",
+    "Didn't": "ˈdɪd.ənt",
+    "Different": "ˈdɪf.ər.ənt",
+    "Digging": "ˈdɪɡ.ɪŋ",
+    "Dina": "",
+    "Diving": "ˈdaɪ.vɪŋ",
+    "Dizzy": "ˈdɪz.i",
+    "Do": "də",
+    "Doc": "dɒk",
+    "Doctor": "ˈdɒk.tər",
+    "Dodger": "ˈdɒdʒ.ər",
+    "Doe": "dəʊ",
+    "Doesn't": "ˈdʌz.ənt",
+    "Dog": "dɒɡ",
+    "Doggie": "ˈdɔ·ɡi",
+    "Doggy": "ˈdɒɡ.i",
+    "Doing": "ˈduː.ɪŋ",
+    "Dolly": "ˈdɒl.i",
+    "Don": "dɒn",
+    "Don't": "dəʊnt",
+    "Done": "dʌn",
+    "Donkey": "ˈdɒŋ.ki",
+    "Doorman": "ˈdɔː.mən",
+    "Dora": "",
+    "Doran": "",
+    "Dot": "dɒt",
+    "Doug": "",
+    "Down": "daʊn",
+    "Dr": "ˈdɑk·tər",
+    "Draft": "drɑːft",
+    "Dragon": "ˈdræɡ.ən",
+    "Drava": "",
+    "Dried": "draɪd",
+    "Drive": "draɪv",
+    "Drop": "drɒp",
+    "Dropped": "",
+    "Drossage": "",
+    "Duckie": "",
+    "Dump": "dʌmp",
+    "Dumped": "dʌmp",
+    "Dun": "dʌn",
+    "Duncan": "",
+    "Dusty": "dʌst",
+    "Dylan": "",
+    "E.N": "ˌɒn ˈblɒk",
+    "END": "end",
+    "Each": "iːtʃ",
+    "Easter": "ˈiː.stər",
+    "Easy": "ˈiː.zi",
+    "Eat": "iːt",
+    "Eating": "iːt",
+    "Edmonton": "ˈed.mən.tən",
+    "Egg": "eɡ",
+    "Elephant": "ˈel.ɪ.fənt",
+    "Elf": "elf",
+    "Elfin": "ˈel.fɪn",
+    "Elisa": "",
+    "Eliza": "",
+    "Ella": "",
+    "Ellie": "",
+    "Elvin": "",
+    "Emily": "",
+    "Emma": "",
+    "End": "end",
+    "Enjoy": "ɪnˈdʒɔɪ",
+    "Eric": "",
+    "Erin": "",
+    "Ernie": "",
+    "Even": "ˈiː.vən",
+    "Everlet": "",
+    "Every": "ˈev.ri",
+    "Everybody": "ˈev.riˌbɒd.i",
+    "Everyone": "ˈev.ri.wʌn",
+    "Everything": "ˈev.ri.θɪŋ",
+    "Exactly": "ɪɡˈzækt.li",
+    "Excellent": "ˈek.səl.ənt",
+    "Except": "ɪkˈsept",
+    "Excuse": "ɪkˈskjuːz",
+    "Eyebrow": "ˈaɪ.braʊ",
+    "Faked": "feɪk",
+    "Fall": "fɔːl",
+    "Fantastic": "fænˈtæs.tɪk",
+    "Far": "fɑːr",
+    "Fell": "fel",
+    "Ferguson": "",
+    "Finally": "ˈfaɪ.nəl.i",
+    "Finish": "ˈfɪn.ɪʃ",
+    "First": "ˈfɜːst",
+    "Five": "faɪv",
+    "Fixed": "fɪkst",
+    "Fixicane": "",
+    "Fluffball": "ˈflʌf ˌbɔːl",
+    "Fluffy": "ˈflʌf.i",
+    "Fly": "flaɪ",
+    "Fog": "fɒɡ",
+    "Follow": "ˈfɒl.əʊ",
+    "For": "fɔːr",
+    "Forget": "fəˈɡet",
+    "From": "frɒm",
+    "Full": "fʊl",
+    "Fulton": "",
+    "Gave": "ɡeɪv",
+    "George": "",
+    "Get": "ɡet",
+    "Getting": "",
+    "Giddy": "ˈɡɪd.i",
+    "Giraffe": "dʒɪˈrɑːf",
+    "Girl": "ɡɜːl",
+    "Give": "ɡɪv",
+    "Glen": "ɡlen",
+    "Go": "ɡəʊ",
+    "God": "ɡɒd",
+    "Goe": "",
+    "Goed": "",
+    "Going": "ɡəʊ",
+    "Good": "ɡʊd",
+    "Good-bye": "",
+    "Goodbye": "ɡʊdˈbaɪ",
+    "Gossed": "",
+    "Got": "ɡɒt",
+    "Grabbed": "",
+    "Grace": "ɡreɪs",
+    "Graf": "",
+    "Graham": "ˈɡreɪ.əm ˌkræk.ər",
+    "Grandma": "ˈɡræn.mɑː",
+    "Graydon": "",
+    "Great": "ɡreɪt",
+    "Gro": "",
+    "Grrr": "",
+    "Ha": "hɑː",
+    "Had": "hæd",
+    "Hailey": "",
+    "Haley": "",
+    "Half": "hɑːf",
+    "Hang": "hæŋ",
+    "Hannah": "",
+    "Happy": "ˈhæp.i",
+    "Hare": "heər",
+    "Harold": "",
+    "Harry": "ˈhær.i",
+    "Have": "hæv",
+    "Having": "",
+    "Hayley": "",
+    "He": "hiː",
+    "He'd": "hiːd",
+    "Hear": "hɪər",
+    "Heather": "ˈheð.ər",
+    "Held": "held",
+    "Hello": "heˈləʊ",
+    "Help": "help",
+    "Helping": "ˈhel.pɪŋ",
+    "Her": "hɜːr",
+    "Here": "hɪər",
+    "Hey": "heɪ",
+    "Hi": "haɪ",
+    "High": "haɪ",
+    "Hillary": "",
+    "Hip": "hɪp",
+    "Hippo": "ˌhɪp.əˈpɒt.ə.məs",
+    "Hmm": "həm",
+    "Hold": "həʊld",
+    "Homesteader": "ˈhəʊmˌsted.ər",
+    "Hook": "hʊk",
+    "Hope": "həʊp",
+    "Horse": "hɔːs",
+    "How": "haʊ",
+    "Hugged": "",
+    "Hugh": "",
+    "Huh": "hə",
+    "Hungry": "ˈhʌŋ.ɡri",
+    "Hurry": "ˈhʌr.i",
+    "Hurt": "hɜːt",
+    "I": "aɪ",
+    "I'd": "aɪd",
+    "I'll": "aɪl",
+    "I'm": "aɪm",
+    "I've": "aɪv",
+    "Ian": "æn",
+    "If": "ɪf",
+    "In": "ɪn",
+    "Inside": "ɪnˈsaɪd",
+    "Instead": "ɪnˈsted",
+    "Into": "ˈɪn.tuː",
+    "Involve": "ɪnˈvɒlv",
+    "It": "ɪt",
+    "It'll": "ˈɪt.əl",
+    "JSCOT": "",
+    "Jabba": "",
+    "Jack": "dʒæk",
+    "Jaff": "",
+    "Jaffer": "",
+    "Jake": "",
+    "Jame": "",
+    "Jamie": "",
+    "Jap": "",
+    "Jared": "",
+    "Jason": "",
+    "Jayden": "",
+    "Je": "ˌʒə nə seɪ ˈkwɑː",
+    "Jeff": "ef",
+    "Jelinda": "",
+    "Jelly": "ˈdʒel.i",
+    "Jennifer": "",
+    "Jenny": "ˌkriː.pɪŋ ˈdʒen.i",
+    "Jermone": "",
+    "Jerome": "",
+    "Jessalyn": "",
+    "Jessica": "",
+    "Jillian": "",
+    "Jim": "ˌdʒɪm ˈkrəʊ",
+    "Jiro": "",
+    "Joanna": "",
+    "Jocko": "",
+    "Joe": "dʒəʊ",
+    "Joff": "",
+    "John": "ˈdʒɒn",
+    "Johnny": "ˈdʒɒn.i",
+    "Jonathan": "ˈdʒɒn.ə.θən",
+    "Joph": "",
+    "Jordan": "ˈdʒɔː.dən",
+    "Jory": "",
+    "Joseph": "",
+    "Joy": "dʒɔɪ",
+    "Juice": "dʒuːs",
+    "Julia": "",
+    "Juliana": "",
+    "Julie": "",
+    "Juliet": "ˌdʒuː.li.ət ˈbæl.kə.ni",
+    "July": "dʒuˈlaɪ",
+    "Jump": "dʒʌmp",
+    "June": "dʒuːn",
+    "Jung": "",
+    "Junk": "dʒʌŋk",
+    "Just": "dʒʌst",
+    "Justin": "",
+    "Juvi": "",
+    "K": "keɪ",
+    "Kaia": "",
+    "Kangaroo": "ˌkæŋ.ɡərˈuː",
+    "Kate": "",
+    "Katie": "",
+    "Kay": "",
+    "Kayla": "",
+    "Kaylee": "",
+    "Keegan": "",
+    "Keenan": "",
+    "Keep": "kiːp",
+    "Keisha": "",
+    "Kevin": "",
+    "Ki": "",
+    "Kicked": "kɪk",
+    "Kimberly": "",
+    "Kind": "kaɪnd",
+    "Kleenex": "ˈkliː.neks",
+    "Kovacevic": "",
+    "Kristen": "",
+    "Kyle": "",
+    "Kylo": "",
+    "LA": "lɑː",
+    "Lady": "ˈleɪ.di",
+    "Lala": "",
+    "Lane": "leɪn",
+    "Larissa": "",
+    "Last": "lɑːst",
+    "Later": "ˈleɪ.tər",
+    "Lauderdale": "",
+    "Laughing": "lɑːf",
+    "Lauren": "",
+    "Laying": "leɪ",
+    "Leah": "",
+    "Leanne": "",
+    "Lee": "liː",
+    "Lefker": "",
+    "Leo": "ˈliː.əʊ",
+    "Leslie": "",
+    "Let": "let",
+    "Level": "ˈlev.əl",
+    "Lifeguard": "ˈlaɪf.ɡɑːd",
+    "Like": "laɪk",
+    "Lily": "ˈlɪl.i",
+    "Lincoln": "ˌlɪŋ.kən ˈred",
+    "Little": "ˈlɪt.əl",
+    "Long": "lɒŋ",
+    "Long-Neck": "ˈlɒŋ.nek",
+    "Long-Nose": "",
+    "Longnose": "",
+    "Look": "lʊk",
+    "Look!\"": "lʊk",
+    "Looked": "lʊk",
+    "Looking": "ˈlʊk.ɪŋ",
+    "Lot": "lɒt",
+    "Louie": "",
+    "Love": "lʌv",
+    "Loved": "lʌv",
+    "Luca": "",
+    "Luke": "",
+    "M": "em",
+    "Machine": "məˈʃiːn",
+    "Mackenzie": "",
+    "Mad": "mæd",
+    "Maddie": "",
+    "Make": "meɪk",
+    "Making": "ˈmeɪ.kɪŋ",
+    "Mama": "məˈmɑː",
+    "Man": "mæn",
+    "Manager": "ˈmæn.ɪ.dʒər",
+    "Marie": "ˌbæn.məˈriː",
+    "Marnie": "",
+    "Martha": "",
+    "Mary": "ˌheɪl ˈmeə.ri",
+    "Matt": "mæt",
+    "Matthew": "",
+    "Maybe": "ˈmeɪ.bi",
+    "McKenzie": "",
+    "McKinsey": "",
+    "Me": "miː",
+    "Megan": "",
+    "Melanie": "",
+    "Merry": "ˈmer.i",
+    "Mesa": "ˈmeɪ.sə",
+    "Mi": "miː",
+    "Michael": "",
+    "Michelle": "",
+    "Mike": "maɪk",
+    "Mikey": "",
+    "Milk": "mɪlk",
+    "Miller": "ˈmɪl.ər",
+    "Mitchell": "",
+    "Mm": "",
+    "Mm-hmm": "",
+    "Mom": "mɒm",
+    "Mommy": "ˈmɒm.i",
+    "Money": "ˈmʌn.i",
+    "Monica": "",
+    "Monster": "ˈmɒn.stər",
+    "Moon": "muːn",
+    "Moore": "",
+    "Moose": "muːs",
+    "More": "mɔːr",
+    "Morgan": "",
+    "Mose": "",
+    "Most": "məʊst",
+    "Mother": "ˈmʌð.ər",
+    "Mouse": "maʊs",
+    "Mr": "ˈmɪs.tər",
+    "Munching": "mʌntʃ",
+    "My": "maɪ",
+    "Nathan": "",
+    "Neck": "nek",
+    "New": "njuː",
+    "Next": "nekst",
+    "Nice": "naɪs",
+    "Nichola": "",
+    "Nick": "nɪk",
+    "Nicole": "",
+    "No": "nəʊ",
+    "Nobody": "ˈnəʊ.bə.di",
+    "Nooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo": "",
+    "Nope": "nəʊp",
+    "Norse": "nɔːs",
+    "North": "nɔːθ",
+    "Nose": "nəʊz",
+    "Not": "nɒt",
+    "Nothing": "ˈnʌθ.ɪŋ",
+    "November": "nəʊˈvem.bər",
+    "Now": "naʊ",
+    "Number": "ˈnʌm.bər",
+    "Nurse": "nɜːs",
+    "O": "əʊ",
+    "OK": "ˌəʊˈkeɪ",
+    "Of": "əv",
+    "Off": "ɒf",
+    "Office": "ˈɒf.ɪs",
+    "Oh": "əʊ",
+    "Ohad": "",
+    "Okay": "ˌəʊˈkeɪ",
+    "Old": "əʊld",
+    "On": "ɒn",
+    "Once": "wʌns",
+    "One": "wʌn",
+    "Only": "ˈəʊn.li",
+    "Onto": "ˈɒn.tu",
+    "Ooh": "uː",
+    "Oop": "ˌæl.iˈuːp",
+    "Oopsie": "",
+    "Open": "ˈəʊ.pən",
+    "Or": "ɔːr",
+    "Orange": "ˈɒr.ɪndʒ",
+    "Oscar": "ˈɒs.kər",
+    "Other": "ˈʌð.ər",
+    "Ouch": "aʊtʃ",
+    "Our": "aʊər",
+    "Out": "aʊt",
+    "Outward": "ˈaʊt.wəd",
+    "Over": "ˈəʊ.vər",
+    "Ow": "aʊ",
+    "PJ": "",
+    "Papa": "pəˈpɑː",
+    "Pardon": "ˈpɑː.dən",
+    "Part": "pɑːt",
+    "Paulina": "",
+    "Pauline": "",
+    "Pax": "pæks",
+    "Peanut": "ˈpiː.nʌt",
+    "Percy": "",
+    "Peter": "ˈpiː.tə",
+    "Pick": "pɪk",
+    "Picking": "ˈpɪkɪŋz",
+    "Picnic": "ˈpɪk.nɪk",
+    "Pig": "pɪɡ",
+    "Piggy": "ˈpɪɡ.i",
+    "Pimperray": "",
+    "Pina": "ˌpiːnə kəˈlɑː.də",
+    "Pink": "pɪŋk",
+    "Pitcher": "ˈpɪtʃ.ər",
+    "Place": "pleɪs",
+    "Play": "pleɪ",
+    "Playgirl": "",
+    "Playing": "kɑːd",
+    "Please": "pliːz",
+    "Plyfar": "",
+    "Po": "ˌpiːˈəʊ",
+    "Pointing": "ˈpɔɪn.tɪŋ",
+    "Polka": "ˈpɒl.kə",
+    "Polkadot": "ˈpɒl.kə ˌdɒt",
+    "Pool": "puːl",
+    "Pop": "pɒp",
+    "Poppy": "ˈpɒp.i",
+    "Pour": "pɔːr",
+    "Pre": "priː-",
+    "Pretty": "ˈprɪt.i",
+    "Price": "praɪs",
+    "Probably": "ˈprɒb.ə.bəl",
+    "Proof": "pruːf",
+    "Proud": "praʊd",
+    "Puddle": "ˈpʌd.əl",
+    "Pudge": "",
+    "Pull": "pʊl",
+    "Pulling": "pʊl",
+    "Puppy": "ˈpʌp.i",
+    "Push": "pʊʃ",
+    "Pushed": "pʊʃt",
+    "Pushing": "ˈpʊʃ.ɪŋ",
+    "Put": "pʊt",
+    "Putting": "",
+    "Puzzle": "ˈpʌz.əl",
+    "Quarter": "ˈkwɔː.tər",
+    "Quick": "kwɪk",
+    "Quickly": "kwɪk",
+    "R": "ɑːr",
+    "RJ": "",
+    "Rabbi": "ˈræb.aɪ",
+    "Rabbit": "ˈræb.ɪt",
+    "Rabid": "ˈræb.ɪd",
+    "Rachel": "",
+    "Rachelle": "",
+    "Rafi": "",
+    "Randy": "ˈræn.di",
+    "Raph": "",
+    "Rattle": "ˈræt.əl",
+    "Ready": "ˈred.i",
+    "Reagan": "",
+    "Real": "rɪəl",
+    "Really": "ˈrɪə.li",
+    "Rebe": "",
+    "Red": "red",
+    "Refresh": "rɪˈfreʃ",
+    "Reggie": "",
+    "Remember": "rɪˈmem.bər",
+    "Revée": "",
+    "Rex": "reks",
+    "Rhian": "",
+    "Ricky": "",
+    "Right": "raɪt",
+    "Rihanna": "",
+    "Riley": "",
+    "Ringing": "ˈrɪŋ.ɪŋ",
+    "River": "ˈrɪv.ər",
+    "Rivet": "ˈrɪv.ɪt",
+    "Robert": "ˌhɜːb ˈrɒb.ət",
+    "Robin": "ˈrɒb.ɪn",
+    "Rochelle": "",
+    "Rocking": "rɒk",
+    "Roger": "ˈrɒdʒ.ər",
+    "Ronald": "",
+    "Rose": "rəʊz",
+    "Rosie": "ˌrəʊ.zi ˈliː",
+    "Rudy": "",
+    "Run": "rʌn",
+    "Running": "ˈrʌn.ɪŋ",
+    "Ryan": "",
+    "S": "es",
+    "Sad": "sæd",
+    "Said": "sed",
+    "Sally": "ˈsæl.i",
+    "Sam": "sæm",
+    "Samantha": "",
+    "Same": "seɪm",
+    "Sand": "sænd",
+    "Sandash": "",
+    "Sandcastle": "ˈsændˌkɑː.səl",
+    "Sandy": "ˈsæn.dɪ",
+    "Santa": "ˈsæn.tə ˌklɔːz",
+    "Sarah": "",
+    "Saria": "",
+    "Save": "seɪv",
+    "Say": "seɪ",
+    "Saying": "ˈseɪ.ɪŋ",
+    "Scared": "skeəd",
+    "School": "skuːl",
+    "Scott": "",
+    "Scratch": "skrætʃ",
+    "Second": "ˈsek.ənd",
+    "See": "siː",
+    "Seeing": "ˈsiː.ɪŋ",
+    "Seven": "ˈsev.ən",
+    "Shaking": "",
+    "Shall": "ʃæl",
+    "Shalooza": "",
+    "Shane": "",
+    "Shani": "",
+    "Sharon": "ˈʃær.ən ˌfruːt",
+    "She": "ʃiː",
+    "Sheila": "ˈʃiː.lə",
+    "Shelby": "",
+    "Shell": "ʃel",
+    "Sherlock": "",
+    "Shh": "ʃ",
+    "Shine": "ʃaɪn",
+    "Shooting": "ˈʃuː.tɪŋ",
+    "Shopping": "ˈʃɒp.ɪŋ",
+    "Short": "ʃɔːt",
+    "Should": "ʃʊd",
+    "Shouldn't": "ˈʃʊd.ənt",
+    "Show": "ʃəʊ",
+    "Shut": "ʃʌt",
+    "Shy": "ʃaɪ",
+    "Signing": "ˈsaɪ.nɪŋ",
+    "Since": "sɪns",
+    "Sink": "sɪŋk",
+    "Sip": "sɪp",
+    "Sit": "sɪt",
+    "Sitting": "ˈsɪt.ɪŋ",
+    "Skye": "ˌskaɪ ˈter.i.ər",
+    "Slipped": "",
+    "Slippery": "ˈslɪp.ər.i",
+    "Slopper": "",
+    "Smaller": "smɔːl",
+    "Snack": "snæk",
+    "Snap": "snæp",
+    "Snout": "snaʊt",
+    "Snow": "snəʊ",
+    "So": "es",
+    "Solly": "",
+    "Some": "sʌm",
+    "Somebody": "ˈsʌm.bə.di",
+    "Someone": "ˈsʌm.wʌn",
+    "Something": "ˈsʌm.θɪŋ",
+    "Sometime": "ˈsʌm.taɪm",
+    "Soon": "suːn",
+    "Sorensen": "",
+    "Sorry": "ˈsɒr.i",
+    "Sound": "saʊnd",
+    "Sounded": "saʊnd",
+    "Spanked": "spæŋk",
+    "Speak": "spiːk",
+    "Spilling": "spɪl",
+    "Spilt": "",
+    "Splash": "splæʃ",
+    "Spooz": "",
+    "St": "",
+    "Stained": "steɪn",
+    "Standing": "ˈstæn.dɪŋ",
+    "Star": "stɑːr",
+    "Start": "stɑːt",
+    "Started": "stɑːt",
+    "Starting": "ˈstɑː.tɪŋ",
+    "Stay": "steɪ",
+    "Stefan": "",
+    "Steinha": "",
+    "Stephane": "",
+    "Stephanie": "",
+    "Stephon": "",
+    "Steve": "",
+    "Stewart": "",
+    "Stick": "stɪk",
+    "Sticked": "ˈhaɪˌstɪk",
+    "Still": "stɪl",
+    "Stole": "stəʊl",
+    "Stop": "stɒp",
+    "Store": "stɔːr",
+    "Story": "ˈstɔː.ri",
+    "Suck": "sʌk",
+    "Suddenly": "ˈsʌd.ən.li",
+    "Sumpra": "",
+    "Super": "ˈsuː.pər",
+    "Supervisor": "ˈsuː.pə.vaɪz",
+    "Sure": "ʃɔːr",
+    "Susan": "ˌleɪ.zi ˈsuː.zən",
+    "Suzanne": "",
+    "Sway": "sweɪ",
+    "Swim": "swɪm",
+    "Swimmed": "",
+    "Swimming": "swɪm",
+    "T": "tiː",
+    "T-Rex": "ˈtiːˌreks",
+    "T-Star": "",
+    "T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-T-": "tiː",
+    "T.O.S": "ˌtiː.əʊˈes",
+    "TJ": "",
+    "Ta-da": "təˈdɑː",
+    "Take": "teɪk",
+    "Taking": "",
+    "Tali": "",
+    "Talk": "tɔːk",
+    "Talked": "tɔːk",
+    "Talking": "tɔːk",
+    "Tammy": "",
+    "Taryn": "",
+    "Tassel": "ˈtæs.əl",
+    "Tata": "",
+    "Tate": "",
+    "Tatum": "",
+    "Tayden": "",
+    "Taylor": "ˈteɪ.lə rɪˌpɔːt",
+    "Tea": "tiː",
+    "Tear": "teər",
+    "Ted": "ˈted.i ˌbɔɪ",
+    "Teddy": "ˈted.i",
+    "Teden": "",
+    "Teeth": "tiːθ",
+    "Tell": "tel",
+    "Teppany": "",
+    "Terry": "ˈter.i",
+    "Tevi": "",
+    "Tevin": "",
+    "Thank": "θæŋk",
+    "That": "ðæt",
+    "The": "ðiː",
+    "Their": "ðeər",
+    "Then": "ðen",
+    "There": "ðeər",
+    "These": "ðiːz",
+    "They": "ðeɪ",
+    "They'll": "ðeɪl",
+    "They're": "ðeər",
+    "Thi": "",
+    "Think": "θɪŋk",
+    "Third": "θɜːd",
+    "Thoma": "",
+    "Those": "ðəʊz",
+    "Three": "θriː",
+    "Through": "θruː",
+    "Thumper": "ˈbaɪ.bəlˌθʌmp.ɚ",
+    "Tia": "ˌtiː.aɪˈeɪ",
+    "Tidy": "ˈtaɪ.di",
+    "Tiger": "ˈtaɪ.ɡər",
+    "Till": "tɪl",
+    "Tim": "",
+    "Time": "taɪm",
+    "Timmy": "",
+    "Tini": "",
+    "To": "tuː",
+    "Toby": "ˈtəʊ.bi ˌdʒʌɡ",
+    "Today": "təˈdeɪ",
+    "Tolly": "",
+    "Tommy": "ˈtɒm.i ɡʌn",
+    "Tomorrow": "təˈmɒr.əʊ",
+    "Tony": "ˈtəʊn.i",
+    "Too": "tuː",
+    "Took": "tʊk",
+    "Tornadoe": "",
+    "Touched": "tʌtʃt",
+    "Tour": "tʊər",
+    "Toward": "tɔːrd",
+    "Toy": "tɔɪ",
+    "Tree": "triː",
+    "Trev": "",
+    "Tried": "traɪd",
+    "Trinity": "ˈtrɪn.ə.ti",
+    "Try": "traɪ",
+    "Trying": "ˈtraɪ.ɪŋ",
+    "Tully": "",
+    "Tumble": "ˈtʌm.bəl",
+    "Turn": "tɜːn",
+    "Turned": "tɜːn",
+    "Two": "tuː",
+    "Ty": "",
+    "Tyler": "",
+    "U": "juː",
+    "Uh": "ɜː",
+    "Uh-huh": "ʌˈhʌ",
+    "Uh-oh": "ˌʌˈəʊ",
+    "Uh-uh": "ɜː",
+    "Uh-uh-uh": "ɜː",
+    "Um": "əm",
+    "Untied": "ʌnˈtaɪ",
+    "Until": "ənˈtɪl",
+    "Up": "ʌp",
+    "Upon": "əˈpɒn",
+    "Val": "",
+    "Vanessa": "",
+    "Velociraptor": "vɪˌlɒs.ɪˈræp.tər",
+    "Verola": "",
+    "Veronica": "vəˈrɒn.ɪ.kə",
+    "Very": "ˈver.i",
+    "Viva": "ˈvaɪ.və",
+    "Wa": "",
+    "Wade": "weɪd",
+    "Wagon": "ˈwæɡ.ən",
+    "Wait": "weɪt",
+    "Wake": "weɪk",
+    "Walk": "wɔːk",
+    "Walked": "wɔːk",
+    "Walking": "ˈwɔː.kɪŋ",
+    "Wanna": "ˈwɒn.ə",
+    "Want": "wɒnt",
+    "Wanted": "ˈwɒn.tɪd",
+    "Wasn't": "ˈwɒz.ənt",
+    "Watch": "wɒtʃ",
+    "Wave": "weɪv",
+    "Waving": "weɪv",
+    "Way": "weɪ",
+    "We": "wiː",
+    "We'll": "wiːl",
+    "We're": "wɪər",
+    "Weird": "wɪəd",
+    "Well": "wel",
+    "Went": "went",
+    "Were": "wɜːr",
+    "Wham": "wæm",
+    "What": "wɒt",
+    "Whatever": "wɒtˈev.ər",
+    "Whee": "wiː",
+    "When": "wen",
+    "Where": "weər",
+    "Which": "wɪtʃ",
+    "While": "waɪl",
+    "White": "waɪt",
+    "Whitney": "",
+    "Who": "huː",
+    "Whoa": "wəʊ",
+    "Whoop": "wuːp",
+    "Whoops-a-daisy": "",
+    "Whoopsie": "",
+    "Why": "waɪ",
+    "Why'd": "",
+    "Wife": "waɪf",
+    "Wilderill": "",
+    "With": "wɪð",
+    "Won": "wʌn",
+    "Wonderful": "ˈwʌn.də.fəl",
+    "Would": "wʊd",
+    "Wow": "waʊ",
+    "Wrappy": "",
+    "X": "eks",
+    "Ya-haaa": "",
+    "Yahoo": "ˈjɑː.huː",
+    "Yay": "jeɪ",
+    "Ye": "jiː",
+    "Yeah": "jeə",
+    "Yeek": "",
+    "Yep": "jep",
+    "Yippee": "jɪˈpiː",
+    "You": "juː",
+    "You'll": "juːl",
+    "You're": "jɔːr",
+    "You've": "juːv",
+    "Your": "jɔːr",
+    "Zach": "",
+    "Zachary": "",
+    "Zebra": "ˈzeb.rə",
+    "Zenyak": "",
+    "a": "eɪ",
+    "abandoned": "əˈbæn.dənd",
+    "able": "ˈeɪ.bəl",
+    "about": "əˈbaʊt",
+    "above": "əˈbʌv",
+    "accident": "ˈæk.sɪ.dənt",
+    "accidentally": "ˌæk.sɪˈden.təl.i",
+    "accidently": "",
+    "according": "əˈkɔːd",
+    "ache": "eɪk",
+    "acro": "",
+    "acting": "ˈæk.tɪŋ",
+    "actually": "ˈæk.tʃu.ə.li",
+    "ad": "æd",
+    "add": "æd",
+    "added": "ˈæd.ɪd",
+    "adder": "ˈæd.ər",
+    "adding": "æd",
+    "adept": "əˈdept",
+    "adly": "",
+    "admire": "ədˈmaɪər",
+    "admiring": "ədˈmaɪə.rɪŋ",
+    "adult": "ˈæd.ʌlt",
+    "aeroplane": "ˈeə.rə.pleɪn",
+    "afe": "",
+    "afety": "",
+    "afford": "əˈfɔːd",
+    "afraid": "əˈfreɪd",
+    "after": "ˈɑːf.tər",
+    "afternoon": "ˌɑːf.təˈnuːn",
+    "again": "əˈɡen",
+    "against": "əˈɡenst",
+    "age": "eɪdʒ",
+    "ago": "əˈɡəʊ",
+    "agree": "əˈɡriː",
+    "ah": "ɑː",
+    "ahead": "əˈhed",
+    "aid": "eɪd",
+    "ail": "eɪl",
+    "ailed": "eɪl",
+    "ailing": "ˈeɪ.lɪŋ",
+    "ailman": "",
+    "ailor": "",
+    "ain't": "eɪnt",
+    "air": "eər",
+    "airplane": "ˈeə.pleɪn",
+    "airport": "ˈeə.pɔːt",
+    "aisle": "aɪl",
+    "alad": "",
+    "ale": "eɪl",
+    "alesman": "",
+    "alesmen": "",
+    "alesperson": "",
+    "alive": "əˈlaɪv",
+    "all": "ɔːl",
+    "alley": "ˈæl.i",
+    "alligator": "ˈæl.ɪ.ɡeɪ.tər",
+    "allowance": "əˈlaʊ.əns",
+    "allowed": "əˈlaʊ",
+    "allying": "əˈlaɪ",
+    "almon": "",
+    "almost": "ˈɔːl.məʊst",
+    "alone": "əˈləʊn",
+    "along": "əˈlɒŋ",
+    "already": "ɔːlˈred.i",
+    "alright": "ɔːlˈraɪt",
+    "also": "ˈɔːl.səʊ",
+    "alt": "ɒlt",
+    "altered": "ˈɒl.təd",
+    "although": "ɔːlˈðəʊ",
+    "alway": "ˈɔːl.weɪ",
+    "am": "æm",
+    "amazed": "əˈmeɪzd",
+    "amazement": "əˈmeɪz.mənt",
+    "amazing": "əˈmeɪ.zɪŋ",
+    "ame": "ˈeɪˌemˈiː",
+    "ample": "ˈæm.pəl",
+    "amusing": "əˈmjuː.zɪŋ",
+    "an": "æn",
+    "and": "ænd",
+    "andal": "",
+    "andbar": "",
+    "andbox": "",
+    "andcastle": "",
+    "andcloth": "",
+    "andpack": "",
+    "andpit": "",
+    "andwich": "",
+    "andwiche": "",
+    "angel": "ˈeɪn.dʒəl",
+    "anger": "ˈæŋ.ɡər",
+    "angry": "ˈæŋ.ɡri",
+    "animal": "ˈæn.ɪ.məl",
+    "ank": "",
+    "ankle": "ˈæŋ.kəl",
+    "annoyed": "əˈnɔɪd",
+    "another": "əˈnʌð.ər",
+    "answered": "ˈɑːn.sər",
+    "answering": "ˈɑːn.sər",
+    "ant": "ænt",
+    "antler": "ˈænt.lər",
+    "any": "ˈen.i",
+    "anybody": "ˈen.iˌbɒd.i",
+    "anymore": "ˌen.iˈmɔːr",
+    "anyone": "ˈen.i.wʌn",
+    "anything": "ˈen.i.θɪŋ",
+    "anytime": "ˈen.i.taɪm",
+    "anyway": "ˈen.i.weɪ",
+    "anywhere": "ˈen.i.weər",
+    "apart": "əˈpɑːt",
+    "apartment": "əˈpɑːt.mənt",
+    "ape": "eɪp",
+    "apologize": "əˈpɒl.ə.dʒaɪz",
+    "apologizing": "əˈpɒl.ə.dʒaɪz",
+    "apparel": "əˈpær.əl",
+    "appear": "əˈpɪər",
+    "apple": "ˈæp.əl",
+    "appointment": "əˈpɔɪnt.mənt",
+    "appreciate": "əˈpriː.ʃi.eɪt",
+    "apprised": "əˈpraɪz",
+    "arc": "ɑːk",
+    "are": "ɑːr",
+    "aren't": "ɑːnt",
+    "argh": "",
+    "argument": "ˈɑːɡ.jə.mənt",
+    "arm": "ɑːm",
+    "around": "əˈraʊnd",
+    "arrived": "əˈraɪv",
+    "arrow": "ˈær.əʊ",
+    "ashamed": "əˈʃeɪmd",
+    "ask": "ɑːsk",
+    "asked": "ɑːsk",
+    "asking": "ɑːsk",
+    "asleep": "əˈsliːp",
+    "assle": "",
+    "at": "æt",
+    "ate": "et",
+    "attached": "əˈtætʃt",
+    "attack": "əˈtæk",
+    "attendance": "əˈten.dəns",
+    "attention": "əˈten.ʃən",
+    "aunt": "ɑːnt",
+    "avalanche": "ˈæv.əl.ɑːntʃ",
+    "ave": "æv",
+    "aved": "",
+    "average": "ˈæv.ər.ɪdʒ",
+    "avocadion": "",
+    "aw": "ɔː",
+    "awake": "əˈweɪk",
+    "awakened": "əˈweɪ.kən",
+    "away": "əˈweɪ",
+    "awed": "ɔːd",
+    "awesome": "ˈɔː.səm",
+    "awful": "ˈɔː.fəl",
+    "ay": "aɪ",
+    "aying": "",
+    "azote": "",
+    "ba-ba": "ˌbiːˈeɪ",
+    "babie": "",
+    "baboon": "bəˈbuːn",
+    "baby": "ˈbeɪ.bi",
+    "back": "bæk",
+    "backed": "-bækt",
+    "backpack": "ˈbæk.pæk",
+    "backward": "ˈbæk.wəd",
+    "backyard": "ˌbækˈjɑːd",
+    "bad": "bæd",
+    "badly": "ˈbæd.li",
+    "bag": "bæɡ",
+    "baker": "ˈbeɪ.kər",
+    "bakery": "ˈbeɪ.kər.i",
+    "balance": "ˈbæl.əns",
+    "ball": "bɔːl",
+    "ball-": "bɔːl",
+    "balloon": "bəˈluːn",
+    "ballooned": "bəˈluːn",
+    "banana": "bəˈnɑː.nə",
+    "band": "bænd",
+    "band-aid": "ˈbænd.eɪd",
+    "bandage": "ˈbæn.dɪdʒ",
+    "bandaged": "ˈbæn.dɪdʒ",
+    "bandaid": "",
+    "banded": "bænd",
+    "banding": "bænd",
+    "bandy": "ˈbæn.di",
+    "bang": "bæŋ",
+    "banged": "bæŋ",
+    "banging": "ˈbæŋ.ɪŋ",
+    "bank": "bæŋk",
+    "banker": "ˈbæŋ.kər",
+    "banner": "ˈbæn.ər",
+    "bar": "bɑːr",
+    "bare": "beər",
+    "barely": "ˈbeə.li",
+    "bark": "bɑːk",
+    "barking": "ˈbɑː.kɪŋ",
+    "barrel": "ˈbær.əl",
+    "baseball": "ˈbeɪs.bɔːl",
+    "baseballing": "",
+    "basket": "ˈbɑː.skɪt",
+    "basketball": "ˈbɑː.skɪt.bɔːl",
+    "bat": "bæt",
+    "bath": "bɑːθ",
+    "bathing": "ˈbeɪ.ðɪŋ",
+    "bathroom": "ˈbɑːθ.ruːm",
+    "batter": "ˈbæt.ər",
+    "battling": "ˈbæt.əl",
+    "be": "biː",
+    "beach": "biːtʃ",
+    "beam": "biːm",
+    "bear": "beər",
+    "beat": "biːt",
+    "beautiful": "ˈbjuː.tɪ.fəl",
+    "became": "bɪˈkeɪm",
+    "because": "bɪˈkəz",
+    "bed": "bed",
+    "bee": "biː",
+    "beehole": "",
+    "been": "biːn",
+    "before": "bɪˈfɔːr",
+    "began": "bɪˈɡæn",
+    "begged": "",
+    "begging": "",
+    "begin": "bɪˈɡɪn",
+    "beginning": "bɪˈɡɪn.ɪŋ",
+    "behind": "bɪˈhaɪnd",
+    "being": "ˈbiː.ɪŋ",
+    "believe": "bɪˈliːv",
+    "bell": "bel",
+    "belly": "ˈbel.i",
+    "bellyache": "ˈbel.i.eɪk",
+    "belong": "bɪˈlɒŋ",
+    "belonged": "bɪˈlɒŋ",
+    "below": "bɪˈləʊ",
+    "belting": "belt",
+    "bench": "bentʃ",
+    "benche": "",
+    "bend": "bend",
+    "bendy": "ˈben.di",
+    "bent": "bent",
+    "berrie": "",
+    "beside": "bɪˈsaɪd",
+    "best": "best",
+    "bestest": "ˈbest.ɪst",
+    "better": "ˈbet.ər",
+    "between": "bɪˈtwiːn",
+    "big": "bɪɡ",
+    "bigger": "",
+    "biggest": "",
+    "bill": "bɪl",
+    "billion": "ˈbɪl.jən",
+    "bin": "bɪn",
+    "binder": "ˈbaɪn.dər",
+    "bird": "bɜːd",
+    "birthday": "ˈbɜːθ.deɪ",
+    "biscuit": "ˈbɪs.kɪt",
+    "bit": "bɪt",
+    "bite": "baɪt",
+    "bitten": "ˈbɪt.ən",
+    "bitter": "ˈbɪt.ər",
+    "black": "blæk",
+    "blah": "blɑː",
+    "blaming": "bleɪm",
+    "blanket": "ˈblæŋ.kɪt",
+    "blast": "blɑːst",
+    "bleeding": "ˈbliː.dɪŋ",
+    "blend": "blend",
+    "blew": "bluː",
+    "bloat": "bləʊt",
+    "block": "blɒk",
+    "blonde": "blɒnd",
+    "blood": "blʌd",
+    "bloom": "bluːm",
+    "blow": "bləʊ",
+    "blowed": "",
+    "blower": "ˈbləʊ.ər",
+    "blowing": "bləʊ",
+    "blue": "bluː",
+    "blunt": "blʌnt",
+    "blush": "blʌʃ",
+    "blushed": "blʌʃ",
+    "blushing": "blʌʃ",
+    "bo": "ˌbiːˈəʊ",
+    "board": "bɔːd",
+    "boat": "bəʊt",
+    "body": "ˈbɒd.i",
+    "bogeyman": "ˈbəʊ.ɡi.mæn",
+    "boil": "bɔɪl",
+    "bone": "bəʊn",
+    "bonked": "bɒŋk",
+    "boo": "buː",
+    "boo-boo": "ˈbuː.buː",
+    "boob": "buːb",
+    "boobie": "",
+    "booboo": "",
+    "booing": "buː",
+    "book": "bʊk",
+    "boom": "buːm",
+    "boomerang": "ˈbuː.mə.ræŋ",
+    "boon": "buːn",
+    "boot": "buːt",
+    "bored": "bɔːd",
+    "borrow": "ˈbɒr.əʊ",
+    "borrowed": "ˈbɒr.əʊ",
+    "both": "bəʊθ",
+    "bottle": "ˈbɒt.əl",
+    "bottom": "ˈbɒt.əm",
+    "bought": "bɔːt",
+    "bounce": "baʊns",
+    "bounced": "baʊns",
+    "bouncing": "ˈbaʊn.sɪŋ",
+    "bouncy": "ˈbaʊn.si",
+    "bound": "baʊnd",
+    "bounty": "ˈbaʊn.ti",
+    "bow": "baʊ",
+    "bowl": "bəʊl",
+    "bowling": "ˈbəʊ.lɪŋ",
+    "box": "bɒks",
+    "boxe": "",
+    "boy": "bɔɪ",
+    "boyfriend": "ˈbɔɪ.frend",
+    "brain": "breɪn",
+    "braked": "breɪk",
+    "branch": "brɑːntʃ",
+    "brand": "brænd",
+    "brave": "breɪv",
+    "bread": "bred",
+    "break": "breɪk",
+    "breaking": "ˈbreɪ.kɪŋ",
+    "breath": "breθ",
+    "breathe": "briːð",
+    "brick": "brɪk",
+    "bridge": "brɪdʒ",
+    "bright": "braɪt",
+    "bring": "brɪŋ",
+    "bringing": "brɪŋ",
+    "brink": "brɪŋk",
+    "broke": "brəʊk",
+    "broken": "ˈbrəʊ.kən",
+    "broom": "bruːm",
+    "brother": "ˈbrʌð.ər",
+    "brought": "brɔːt",
+    "bruise": "bruːz",
+    "bruised": "bruːzd",
+    "brush": "brʌʃ",
+    "brushe": "",
+    "bubble": "ˈbʌb.əl",
+    "buck": "bʌk",
+    "bucket": "ˈbʌk.ɪt",
+    "buddie": "",
+    "budding": "ˈbʌd.ɪŋ",
+    "buddy": "ˈbʌd.i",
+    "budget": "ˈbʌdʒ.ɪt",
+    "bug": "bʌɡ",
+    "bugging": "",
+    "build": "bɪld",
+    "build-up": "bɪld",
+    "builded": "",
+    "building": "ˈbɪl.dɪŋ",
+    "built": "bɪlt",
+    "bull": "bʊl",
+    "bump": "bʌmp",
+    "bumped": "bʌmp",
+    "bun": "bʌn",
+    "bunch": "bʌntʃ",
+    "bunion": "ˈbʌn.jən",
+    "bunnie": "",
+    "bunny": "ˈbʌn.i",
+    "burie": "",
+    "buried": "ˈber.i",
+    "burn": "bɜːn",
+    "burned": "bɜːn",
+    "burning": "ˈbɜː.nɪŋ",
+    "burnt": "bɜːnt",
+    "burp": "bɜːp",
+    "burped": "bɜːp",
+    "burping": "bɜːp",
+    "burst": "bɜːst",
+    "bury": "ˈber.i",
+    "buse": "",
+    "busine": "",
+    "busse": "",
+    "busy": "ˈbɪz.i",
+    "but": "bʌt",
+    "butter": "ˈbʌt.ər",
+    "butterflie": "",
+    "butterfly": "ˈbʌt.ə.flaɪ",
+    "button": "ˈbʌt.ən",
+    "buy": "baɪ",
+    "buying": "baɪ",
+    "buzze": "",
+    "by": "baɪ",
+    "bye": "baɪ",
+    "cab": "kæb",
+    "cabbed": "",
+    "cake": "keɪk",
+    "call": "kɔːl",
+    "called": "kɔːl",
+    "calling": "ˈkɔː.lɪŋ",
+    "calm": "kɑːm",
+    "came": "keɪm",
+    "camel": "ˈkæm.əl",
+    "camera": "ˈkæm.rə",
+    "camp": "kæmp",
+    "can": "kæn",
+    "can't": "kɑːnt",
+    "cancel": "ˈkæn.səl",
+    "candy": "ˈkæn.di",
+    "cane": "keɪn",
+    "canned": "kænd",
+    "cannonball": "ˈkæn.ən.bɔːl",
+    "cannot": "ˈkæn.ɒt",
+    "cant": "kænt",
+    "cape": "keɪp",
+    "car": "kɑːr",
+    "card": "kɑːd",
+    "care": "keər",
+    "cared": "keər",
+    "caredy": "",
+    "careful": "ˈkeə.fəl",
+    "carefully": "ˈkeə.fəl.i",
+    "carpet": "ˈkɑː.pɪt",
+    "carriage": "��kær.ɪdʒ",
+    "carrie": "",
+    "carried": "ˈkær.i",
+    "carrier": "ˈkær.i.ər",
+    "carrot": "ˈkær.ət",
+    "carry": "ˈkær.i",
+    "carrying": "ˈkær.i",
+    "cart": "kɑːt",
+    "carton": "ˈkɑː.tən",
+    "cary": "",
+    "case": "keɪs",
+    "cash": "kæʃ",
+    "cashier": "kæʃˈɪər",
+    "cast": "kɑːst",
+    "castle": "ˈkɑː.səl",
+    "casual": "ˈkæʒ.ju.əl",
+    "cat": "kæt",
+    "catch": "kætʃ",
+    "catche": "",
+    "catched": "",
+    "catcher": "ˈkætʃ.ər",
+    "catching": "ˈkætʃ.ɪŋ",
+    "caught": "kɔːt",
+    "cause": "kɔːz",
+    "ceased": "siːs",
+    "cell": "sel",
+    "cellar": "ˈsel.ər",
+    "cement": "sɪˈment",
+    "cene": "",
+    "cent": "sent",
+    "center": "ˈsen.tər",
+    "cereal": "ˈsɪə.ri.əl",
+    "chair": "tʃeər",
+    "change": "tʃeɪndʒ",
+    "changed": "tʃeɪndʒd",
+    "chapter": "ˈtʃæp.tər",
+    "character": "ˈkær.ək.tər",
+    "charge": "tʃɑːdʒ",
+    "charged": "tʃɑːdʒd",
+    "charming": "ˈtʃɑː.mɪŋ",
+    "chase": "tʃeɪs",
+    "chased": "tʃeɪs",
+    "chasing": "tʃeɪs",
+    "chat": "tʃæt",
+    "chatted": "tʃæt",
+    "check": "tʃek",
+    "checked": "tʃekt",
+    "checking": "tʃek",
+    "checkup": "ˈtʃek.ʌp",
+    "cheek": "tʃiːk",
+    "cheer": "tʃɪər",
+    "cheering": "ˈtʃɪə.rɪŋ",
+    "cheese": "tʃiːz",
+    "cheetah": "ˈtʃiː.tə",
+    "chick": "tʃɪk",
+    "chief": "tʃiːf",
+    "child": "tʃaɪld",
+    "children": "ˈtʃɪl.drən",
+    "chilly": "ˈtʃɪl.i",
+    "chip": "tʃɪp",
+    "chocolate": "ˈtʃɒk.lət",
+    "choked": "tʃəʊkt",
+    "choo-choo": "ˈtʃuː.tʃuː",
+    "chool": "",
+    "choose": "tʃuːz",
+    "chopped": "",
+    "chopper": "ˈtʃɒp.ər",
+    "chopping": "",
+    "chose": "tʃəʊz",
+    "chosen": "ˈtʃəʊ.zən",
+    "chowing": "",
+    "chubby": "ˈtʃʌb.i",
+    "chucked": "tʃʌk",
+    "chug": "tʃʌɡ",
+    "circle": "ˈsɜː.kəl",
+    "circly": "",
+    "circu": "",
+    "city": "ˈsɪt.i",
+    "cla": "",
+    "clapped": "",
+    "classe": "",
+    "clean": "kliːn",
+    "cleaned": "kliːn",
+    "cleaner": "ˈkliː.nər",
+    "cleaning": "ˈkliː.nɪŋ",
+    "clear": "klɪər",
+    "clearing": "ˈklɪə.rɪŋ",
+    "clerk": "klɑːk",
+    "clever": "ˈklev.ər",
+    "climb": "klaɪm",
+    "climbed": "klaɪm",
+    "climbing": "ˈklaɪ.mɪŋ",
+    "clip": "klɪp",
+    "close": "kləʊz",
+    "closed": "kləʊzd",
+    "closer": "ˈkləʊ.zər",
+    "closing": "ˈkləʊ.zɪŋ",
+    "cloth": "klɒθ",
+    "clothe": "kləʊð",
+    "cloud": "klaʊd",
+    "clown": "klaʊn",
+    "club": "klʌb",
+    "clump": "klʌmp",
+    "coach": "kəʊtʃ",
+    "coache": "",
+    "coast": "kəʊst",
+    "coat": "kəʊt",
+    "coated": "ˈkəʊ.tɪd",
+    "coconut": "ˈkəʊ.kə.nʌt",
+    "coin": "kɔɪn",
+    "cold": "kəʊld",
+    "collapse": "kəˈlæps",
+    "collapsed": "kəˈlæpst",
+    "collecting": "kəˈlekt",
+    "color": "ˈkʌl.ər",
+    "colossal": "kəˈlɒs.əl",
+    "comb": "kəʊm",
+    "come": "kʌm",
+    "comfort": "ˈkʌm.fət",
+    "coming": "ˈkʌm.ɪŋ",
+    "comment": "ˈkɒm.ent",
+    "commonly": "ˈkɒm.ən.li",
+    "complain": "kəmˈpleɪn",
+    "complaining": "kəmˈpleɪ.nɪŋ",
+    "complimented": "ˈkɒm.plɪ.mənt",
+    "condition": "kənˈdɪʃ.ən",
+    "cone": "kəʊn",
+    "confused": "kənˈfjuːzd",
+    "consciou": "",
+    "constantly": "ˈkɒn.stənt.li",
+    "continue": "kənˈtɪn.juː",
+    "continued": "kənˈtɪn.juːd",
+    "control": "kənˈtrəʊl",
+    "cook": "kʊk",
+    "cookie": "ˈkʊk.i",
+    "cool": "kuːl",
+    "coop": "kuːp",
+    "cooped": "ˌkuːpt ˈʌp",
+    "cooper": "ˈkuː.pər",
+    "cooping": "kuːp",
+    "cooter": "",
+    "cord": "kɔːd",
+    "core": "kɔːr",
+    "corner": "ˈkɔː.nər",
+    "cost": "kɒst",
+    "costed": "kɒst",
+    "costume": "ˈkɒs.tʃuːm",
+    "could": "kʊd",
+    "couldn't": "ˈkʊd.ənt",
+    "council": "ˈkaʊn.səl",
+    "count": "kaʊnt",
+    "country": "ˈkʌn.tri",
+    "couple": "ˈkʌp.əl",
+    "courage": "ˈkʌr.ɪdʒ",
+    "course": "kɔːs",
+    "court": "kɔːt",
+    "covered": "ˈkʌv.əd",
+    "covering": "ˈkʌv.ər.ɪŋ",
+    "cow": "kaʊ",
+    "cow-cow-cow": "kaʊ",
+    "cow-like": "",
+    "cowardly": "ˈkaʊ.əd.li",
+    "crabby": "ˈkræb.i",
+    "crack": "kræk",
+    "craft": "krɑːft",
+    "cranky": "ˈkræŋ.ki",
+    "crap": "kræp",
+    "crape": "kreɪp",
+    "craped": "",
+    "crash": "kræʃ",
+    "crashe": "",
+    "crashed": "kræʃ",
+    "crashing": "ˈkræʃ.ɪŋ",
+    "cratch": "",
+    "cratched": "",
+    "crawled": "krɔːl",
+    "crazy": "ˈkreɪ.zi",
+    "cream": "kriːm",
+    "creamed": "kriːm",
+    "creaming": "kriːm",
+    "crewing": "kruː",
+    "crie": "",
+    "cried": "kraɪd",
+    "cro": "ˌkrəʊˈmæɡ.nən",
+    "crossed": "krɒst",
+    "crow": "krəʊ",
+    "crowd": "kraʊd",
+    "crowded": "ˈkraʊ.dɪd",
+    "crown": "kraʊn",
+    "crumb": "krʌm",
+    "crumpled": "ˈkrʌm.pəld",
+    "crushe": "",
+    "crushed": "krʌʃ",
+    "cry": "kraɪ",
+    "crying": "ˈkraɪ.ɪŋ",
+    "culture": "ˈkʌl.tʃər",
+    "cup": "kʌp",
+    "curb": "kɜːb",
+    "cure": "kjʊər",
+    "curiou": "",
+    "curse": "kɜːs",
+    "customer": "ˈkʌs.tə.mər",
+    "cut": "kʌt",
+    "cute": "kjuːt",
+    "cutting": "ˈkʌt.ɪŋ",
+    "cuz": "kəz",
+    "dad": "dæd",
+    "dada": "ˈdɑː.dɑː",
+    "daddy": "ˈdæd.i",
+    "dainty": "ˈdeɪn.ti",
+    "daisy": "ˈdeɪ.zi",
+    "daisying": "",
+    "damaging": "ˈdæm.ɪ.dʒɪŋ",
+    "damien": "",
+    "damper": "ˈdæm.pər",
+    "dance": "dɑːns",
+    "dancing": "ˈdɑːn.sɪŋ",
+    "dang": "dæŋ",
+    "danger": "ˈdeɪn.dʒər",
+    "dark": "dɑːk",
+    "darker": "dɑːk",
+    "darlin": "",
+    "darling": "ˈdɑː.lɪŋ",
+    "date": "deɪt",
+    "daughter": "ˈdɔː.tər",
+    "day": "deɪ",
+    "daycare": "ˈdeɪ ˌkeər",
+    "dazzled": "ˈdæz.əl",
+    "dead": "ded",
+    "deaf": "def",
+    "dear": "dɪər",
+    "decide": "dɪˈsaɪd",
+    "decided": "dɪˈsaɪ.dɪd",
+    "deck": "dek",
+    "decoration": "ˌdek.əˈreɪ.ʃən",
+    "deep": "diːp",
+    "deeper": "diːp",
+    "definitely": "ˈdef.ɪ.nət.li",
+    "deliciou": "",
+    "delightful": "dɪˈlaɪt.fəl",
+    "densely": "ˈdens.li",
+    "dentist": "ˈden.tɪst",
+    "dessert": "dɪˈzɜːt",
+    "destroy": "dɪˈstrɔɪ",
+    "destroyed": "dɪˈstrɔɪ",
+    "devil": "ˈdev.əl",
+    "did": "dɪd",
+    "did-": "dɪd",
+    "didn't": "ˈdɪd.ənt",
+    "die": "daɪ",
+    "died": "daɪ",
+    "diet": "ˈdaɪ.ət",
+    "different": "ˈdɪf.ər.ənt",
+    "dig": "dɪɡ",
+    "digged": "",
+    "digger": "ˈdɪɡ.ər",
+    "digging": "ˈdɪɡ.ɪŋ",
+    "dinner": "ˈdɪn.ər",
+    "dinosaur": "ˈdaɪ.nə.sɔːr",
+    "dip": "dɪp",
+    "dipped": "",
+    "dippity": "",
+    "dirt": "dɜːt",
+    "dirty": "ˈdɜː.ti",
+    "disappear": "ˌdɪs.əˈpɪər",
+    "disappeared": "ˌdɪs.əˈpɪəd",
+    "disappointed": "ˌdɪs.əˈpɔɪn.tɪd",
+    "disc": "dɪsk",
+    "disciplining": "ˈdɪs.ə.plɪn",
+    "discouraged": "dɪˈskʌr.ɪdʒd",
+    "dish": "dɪʃ",
+    "diskette": "dɪsˈket",
+    "display": "dɪˈspleɪ",
+    "distance": "ˈdɪs.təns",
+    "distract": "dɪˈstrækt",
+    "ditty": "ˈdɪt.i",
+    "dive": "daɪv",
+    "dived": "daɪv",
+    "diver": "ˈdaɪ.vər",
+    "diving": "ˈdaɪ.vɪŋ",
+    "dizzier": "ˈdɪz.i",
+    "dizzy": "ˈdɪz.i",
+    "do": "də",
+    "doc": "dɒk",
+    "doctor": "ˈdɒk.tər",
+    "dodge": "dɒdʒ",
+    "doe": "dəʊ",
+    "doesn't": "ˈdʌz.ənt",
+    "dog": "dɒɡ",
+    "doggie": "ˈdɔ·ɡi",
+    "doggy": "ˈdɒɡ.i",
+    "doing": "ˈduː.ɪŋ",
+    "doll": "dɒl",
+    "dollar": "ˈdɒl.ər",
+    "dolly": "ˈdɒl.i",
+    "dolphin": "ˈdɒl.fɪn",
+    "dome": "dəʊm",
+    "don't": "dəʊnt",
+    "done": "dʌn",
+    "donkey": "ˈdɒŋ.ki",
+    "door": "dɔːr",
+    "doorkeeper": "ˈdɔːˌkiː.pər",
+    "doorman": "ˈdɔː.mən",
+    "dot": "dɒt",
+    "double": "ˈdʌb.əl",
+    "doubt": "daʊt",
+    "dough": "dəʊ",
+    "dove": "dʌv",
+    "down": "daʊn",
+    "down-bait": "",
+    "draft": "drɑːft",
+    "drafted": "drɑːft",
+    "drag": "dræɡ",
+    "dragged": "",
+    "dragging": "",
+    "dragon": "ˈdræɡ.ən",
+    "drank": "dræŋk",
+    "drastery": "",
+    "draw": "drɔː",
+    "drawing": "ˈdrɔː.ɪŋ",
+    "dre": "",
+    "dressed": "drest",
+    "dribbling": "ˈdrɪb.əl.ɪŋ",
+    "dried": "draɪd",
+    "drift": "drɪft",
+    "drifted": "drɪft",
+    "drill": "drɪl",
+    "drink": "drɪŋk",
+    "drinking": "ˈdrɪŋ.kɪŋ",
+    "dripping": "ˈdrɪp.ɪŋ",
+    "drippy": "ˈdrɪp.i",
+    "drive": "draɪv",
+    "drived": "ˈtest ˌdraɪv",
+    "driver": "ˈdraɪ.vər",
+    "driving": "ˈdraɪ.vɪŋ",
+    "drooling": "druːl",
+    "drop": "drɒp",
+    "dropped": "",
+    "dropping": "",
+    "drove": "drəʊv",
+    "drown": "draʊn",
+    "drowned": "draʊn",
+    "drowning": "ˈdraʊn.ɪŋ",
+    "drowse": "draʊz",
+    "drowsy": "ˈdraʊ.zi",
+    "drug": "drʌɡ",
+    "drunk": "drʌŋk",
+    "dry": "draɪ",
+    "duck": "dʌk",
+    "duckle": "",
+    "dude": "duːd",
+    "dug": "dʌɡ",
+    "dumb": "dʌm",
+    "dumbstruck": "ˌdʌmˈfaʊn.dɪd",
+    "dummy": "ˈdʌm.i",
+    "dump": "dʌmp",
+    "dumped": "dʌmp",
+    "dumping": "ˈdʌm.pɪŋ",
+    "dun": "dʌn",
+    "dungeon": "ˈdʌn.dʒən",
+    "duper": "ˌsuː.pəˈduː.pər",
+    "dusty": "ˈdʌs.ti",
+    "ea": "ˌiːˈeɪ",
+    "each": "iːtʃ",
+    "eagle": "ˈiː.ɡəl",
+    "ear": "ɪər",
+    "earched": "",
+    "early": "ˈɜː.li",
+    "earned": "ɜːn",
+    "earphone": "ˈiə.fəʊn",
+    "earring": "ˈɪə.rɪŋ",
+    "earth": "ɜːθ",
+    "easick": "",
+    "easier": "ˈiː.zi",
+    "easily": "ˈiː.zəl.i",
+    "easy": "ˈiː.zi",
+    "eat": "iːt",
+    "eaten": "",
+    "eater": "ˈiː.tər",
+    "eating": "iːt",
+    "ec": "ˌiːˈsiː",
+    "econd": "",
+    "ecret": "",
+    "ection": "",
+    "ecurity": "",
+    "edge": "edʒ",
+    "ee": "-iː",
+    "eed": "",
+    "eeded": "",
+    "eeing": "",
+    "eem": "",
+    "eemed": "",
+    "een": "",
+    "effect": "ɪˈfekt",
+    "effort": "ˈef.ət",
+    "egg": "eɡ",
+    "egg-stricken": "",
+    "eh": "eɪ",
+    "eight": "eɪt",
+    "eight-year-old": "",
+    "either": "ˈaɪ.ðər",
+    "elephant": "ˈel.ɪ.fənt",
+    "elephant-sized": "",
+    "elf": "elf",
+    "ell": "",
+    "eller": "",
+    "elling": "",
+    "else": "els",
+    "elve": "",
+    "elven": "",
+    "embarrassed": "ɪmˈbær.əst",
+    "emergency": "ɪˈmɜː.dʒən.si",
+    "empty": "ˈemp.ti",
+    "encouraged": "ɪnˈkʌr.ɪdʒd",
+    "end": "end",
+    "ended": "",
+    "ending": "",
+    "enemy": "ˈen.ə.mi",
+    "enforcer": "ɪnˈfɔː.sər",
+    "engaged": "ɪnˈɡeɪdʒd",
+    "enjoy": "ɪnˈdʒɔɪ",
+    "enjoyed": "ɪnˈdʒɔɪ",
+    "enjoying": "ɪnˈdʒɔɪ",
+    "enormou": "",
+    "enough": "ɪˈnʌf",
+    "ense": "",
+    "ent": "ˌiː.enˈtiː",
+    "entence": "",
+    "entire": "ɪnˈtaɪər",
+    "entrance": "ˈen.trəns",
+    "equal": "ˈiː.kwəl",
+    "er": "ɜːr",
+    "errand": "ˈer.ənd",
+    "et": "et ˈæl",
+    "ettle": "",
+    "even": "ˈiː.vən",
+    "evening": "ˈiːv.nɪŋ",
+    "event": "ɪˈvent",
+    "eventh": "",
+    "ever": "ˈev.ər",
+    "every": "ˈev.ri",
+    "everybody": "ˈev.riˌbɒd.i",
+    "everyone": "ˈev.ri.wʌn",
+    "everything": "ˈev.ri.θɪŋ",
+    "everywhere": "ˈev.ri.weər",
+    "ew": "ˈiːuː",
+    "ewing": "ˌjuː.ɪŋz sɑːˈkəʊ.mə",
+    "ewn": "",
+    "exact": "ɪɡˈzækt",
+    "exactly": "ɪɡˈzækt.li",
+    "examine": "ɪɡˈzæm.ɪn",
+    "examining": "ɪɡˈzæm.ɪn",
+    "excellent": "ˈek.səl.ənt",
+    "except": "ɪkˈsept",
+    "excited": "ɪkˈsaɪ.tɪd",
+    "excitement": "ɪkˈsaɪt.mənt",
+    "excuse": "ɪkˈskjuːz",
+    "explain": "ɪkˈspleɪn",
+    "explained": "ɪkˈspleɪn",
+    "explaining": "ɪkˈspleɪ.nɪŋ",
+    "exploded": "ɪkˈspləʊd",
+    "expression": "ɪkˈspreʃ.ən",
+    "extra": "ˈek.strə",
+    "eye": "aɪ",
+    "eyeball": "ˈaɪ.bɔːl",
+    "eyebrow": "ˈaɪ.braʊ",
+    "face": "feɪs",
+    "fact": "fækt",
+    "factory": "ˈfæk.tər.i",
+    "faddled": "",
+    "fade": "feɪd",
+    "faded": "ˈfeɪ.dɪd",
+    "failed": "feɪld",
+    "faint": "feɪnt",
+    "fainted": "feɪnt",
+    "fainting": "feɪnt",
+    "fair": "feər",
+    "fairy": "ˈfeə.ri",
+    "faith": "feɪθ",
+    "fake": "feɪk",
+    "faked": "feɪk",
+    "faking": "feɪk",
+    "fall": "fɔːl",
+    "falled": "",
+    "fallen": "ˈfɔː.lən",
+    "falling": "ˈfɔː.lɪŋ",
+    "family": "ˈfæm.əl.i",
+    "fanciest": "ˈfæn.si",
+    "fantastic": "fænˈtæs.tɪk",
+    "far": "fɑːr",
+    "farther": "ˈfɑː.ðər",
+    "fast": "fɑːst",
+    "faster": "fɑːst",
+    "fastest": "fɑːst",
+    "fat": "fæt",
+    "father": "ˈfɑː.ðər",
+    "fatter": "",
+    "fault": "fɒlt",
+    "favorite": "ˈfeɪ.vər.ɪt",
+    "fear": "fɪər",
+    "fed": "fed",
+    "feel": "fiːl",
+    "feeling": "ˈfiː.lɪŋ",
+    "feet": "fiːt",
+    "fell": "fel",
+    "felled": "fel",
+    "felt": "felt",
+    "fetch": "fetʃ",
+    "fetcher": "",
+    "fetching": "ˈfetʃ.ɪŋ",
+    "fever": "ˈfiː.vər",
+    "few": "fjuː",
+    "field": "fiːld",
+    "fifty": "ˈfɪf.ti",
+    "fight": "faɪt",
+    "fighting": "ˈfaɪ.tɪŋ",
+    "figure": "ˈfɪɡ.ər",
+    "fill": "fɪl",
+    "filled": "-fɪld",
+    "filling": "ˈfɪl.ɪŋ",
+    "film": "fɪlm",
+    "final": "ˈfaɪ.nəl",
+    "finally": "ˈfaɪ.nəl.i",
+    "find": "faɪnd",
+    "finding": "ˈfaɪn.dɪŋ",
+    "fine": "faɪn",
+    "finger": "ˈfɪŋ.ɡər",
+    "finish": "ˈfɪn.ɪʃ",
+    "finished": "ˈfɪn.ɪʃt",
+    "finishing": "ˈfɪn.ɪʃ",
+    "fire": "faɪər",
+    "fired": "-faɪəd",
+    "firmly": "ˈfɜːm.li",
+    "first": "ˈfɜːst",
+    "fish": "fɪʃ",
+    "fishe": "",
+    "fished": "fɪʃ",
+    "fishing": "ˈfɪʃ.ɪŋ",
+    "fishnet": "ˈfɪʃ.net",
+    "fist": "fɪst",
+    "fit": "fɪt",
+    "five": "faɪv",
+    "fix": "fɪks",
+    "fixe": "ˌiː.deɪ ˈfiːks",
+    "fixed": "fɪkst",
+    "fixing": "ˈfɪk.sɪŋ",
+    "flag": "flæɡ",
+    "flame": "fleɪm",
+    "flapping": "",
+    "flat": "flæt",
+    "flattening": "ˈflæt.ən",
+    "flatting": "",
+    "flaw": "flɔː",
+    "flew": "fluː",
+    "flie": "",
+    "flied": "flaɪ",
+    "flighted": "ˈflaɪ.tɪd",
+    "fling": "flɪŋ",
+    "flinging": "flɪŋ",
+    "flip": "flɪp",
+    "flipped": "",
+    "float": "fləʊt",
+    "floated": "fləʊt",
+    "floating": "ˈfləʊ.tɪŋ",
+    "floor": "flɔːr",
+    "flow": "fləʊ",
+    "flowed": "fləʊ",
+    "flower": "flaʊər",
+    "flowing": "ˈfləʊ.ɪŋ",
+    "fluid": "ˈfluː.ɪd",
+    "fluttering": "ˈflʌt.ər",
+    "fly": "flaɪ",
+    "flyed": "",
+    "flying": "ˈflaɪ.ɪŋ",
+    "foal": "fəʊl",
+    "fola": "",
+    "fold": "fəʊld",
+    "folding": "ˈfəʊl.dɪŋ",
+    "follow": "ˈfɒl.əʊ",
+    "followed": "ˈfɒl.əʊ",
+    "following": "ˈfɒl.əʊ.ɪŋ",
+    "food": "fuːd",
+    "fool": "fuːl",
+    "fooled": "fuːl",
+    "foot": "fʊt",
+    "for": "fɔːr",
+    "forehead": "ˈfɒr.ɪd",
+    "forest": "ˈfɒr.ɪst",
+    "forever": "fəˈre.vər",
+    "forget": "fəˈɡet",
+    "forgetting": "fəˈɡet",
+    "forgot": "fəˈɡet",
+    "form": "fɔːm",
+    "fought": "fɔːt",
+    "found": "faʊnd",
+    "four": "fɔːr",
+    "fox": "fɒks",
+    "frame": "freɪm",
+    "fraud": "frɔːd",
+    "freak": "friːk",
+    "freaked": "friːk",
+    "freaking": "ˈfriː.kɪŋ",
+    "free": "friː",
+    "fricked": "",
+    "friend": "frend",
+    "fright": "fraɪt",
+    "frightened": "ˈfraɪ.tənd",
+    "from": "frɒm",
+    "front": "frʌnt",
+    "frown": "fraʊn",
+    "frowning": "fraʊn",
+    "froze": "frəʊz",
+    "fruit": "fruːt",
+    "fu": "ˌkʌŋ ˈfuː",
+    "fucking": "ˈfʌk.ɪŋ",
+    "full": "fʊl",
+    "fuller": "fʊl",
+    "fun": "fʌn",
+    "funner": "",
+    "funny": "ˈfʌn.i",
+    "furiou": "",
+    "further": "ˈfɜː.ðər",
+    "ga": "",
+    "gag": "ɡæɡ",
+    "game": "ɡeɪm",
+    "garbage": "ˈɡɑː.bɪdʒ",
+    "garden": "ˈɡɑː.dən",
+    "gather": "ˈɡæð.ər",
+    "gave": "ɡeɪv",
+    "gay": "ɡeɪ",
+    "gear": "ɡɪər",
+    "geniu": "",
+    "get": "ɡet",
+    "getting": "",
+    "ghetto": "ˈɡet.əʊ",
+    "ghost": "ɡəʊst",
+    "ghosty": "",
+    "giddy": "ˈɡɪd.i",
+    "gifted": "ˈɡɪf.tɪd",
+    "gigantic": "ˌdʒaɪˈɡæn.tɪk",
+    "giggled": "ˈɡɪɡ.əl",
+    "giraffe": "dʒɪˈrɑːf",
+    "giraffe-like": "",
+    "girl": "ɡɜːl",
+    "girlfriend": "ˈɡɜːl.frend",
+    "give": "ɡɪv",
+    "giving": "ˈɡɪvɪŋ",
+    "gla": "",
+    "glad": "ɡlæd",
+    "glasse": "",
+    "glide": "ɡlaɪd",
+    "glided": "ɡlaɪd",
+    "glowed": "ɡləʊ",
+    "go": "ɡəʊ",
+    "goal": "ɡəʊl",
+    "goat": "ɡəʊt",
+    "gobble": "ˈɡɒb.əl",
+    "gobbling": "ˈɡɒb.əl",
+    "god": "ɡɒd",
+    "godson": "ˈɡɒd.sʌn",
+    "goe": "",
+    "going": "ˈɡəʊ.ɪŋ",
+    "gold": "ɡəʊld",
+    "gone": "ɡɒn",
+    "gonna": "ˈɡə.nə",
+    "goo-lop": "",
+    "good": "ɡʊd",
+    "goodbye": "ɡʊdˈbaɪ",
+    "goodie": "",
+    "goodne": "",
+    "goofy": "ˈɡuː.fi",
+    "gorilla": "ɡəˈrɪl.ə",
+    "gosh": "ɡɒʃ",
+    "gossanet": "",
+    "gosset": "",
+    "got": "ɡɒt",
+    "gotta": "ˈɡɒt.ə",
+    "gotten": "ˈɡɒt.ən",
+    "gra": "",
+    "grab": "ɡræb",
+    "grabbed": "",
+    "grabbing": "-ɡræb.ɪŋ",
+    "grade": "ɡreɪd",
+    "grandma": "ˈɡræn.mɑː",
+    "grandmother": "ˈɡræn.mʌð.ər",
+    "grandpa": "ˈɡræn.pɑː",
+    "grandson": "ˈɡræn.sʌn",
+    "granny": "ˈɡræn.i",
+    "graph": "ɡrɑːf",
+    "gray": "ɡreɪ",
+    "great": "ɡreɪt",
+    "greedy": "ˈɡriː.di",
+    "green": "ɡriːn",
+    "greet": "ɡriːt",
+    "grew": "ɡruː",
+    "grin": "ɡrɪn",
+    "grinched": "",
+    "grinning": "ˈɡrɪn.ɪŋ",
+    "grip": "ɡrɪp",
+    "gro": "",
+    "grocer": "ˈɡrəʊ.sər",
+    "grocerie": "",
+    "grocery": "ˈɡrəʊ.sər.i",
+    "ground": "ɡraʊnd",
+    "grounded": "ˈɡraʊn.dɪd",
+    "grow": "ɡrəʊ",
+    "growl": "ɡraʊl",
+    "growled": "ɡraʊl",
+    "growling": "ˈɡraʊ.lɪŋ",
+    "grown": "ɡrəʊn",
+    "grown-up": "ˌɡrəʊn ˈʌp",
+    "grueling": "ˈɡruː.ə.lɪŋ",
+    "grumpy": "ˈɡrʌm.pi",
+    "guard": "ɡɑːd",
+    "guarded": "ˈɡɑː.dɪd",
+    "guarding": "ɡɑːd",
+    "gue": "",
+    "guilty": "ˈɡɪl.ti",
+    "gurgling": "ˈɡɜː.ɡəl",
+    "gushed": "ɡʌʃ",
+    "guy": "ɡaɪ",
+    "gym": "dʒɪm",
+    "ha": "hɑː",
+    "habedapa": "",
+    "habit": "ˈhæb.ɪt",
+    "had": "hæd",
+    "hade": "",
+    "hadn't": "ˈhæd.ənt",
+    "hair": "heər",
+    "hake": "heɪk",
+    "haking": "",
+    "half": "hɑːf",
+    "halfway": "ˌhɑːfˈweɪ",
+    "hall": "hɔːl",
+    "hallway": "ˈhɔːl.weɪ",
+    "hammer": "ˈhæm.ər",
+    "hand": "hænd",
+    "handed": "-hæn.dɪd",
+    "handle": "ˈhæn.dəl",
+    "handy-dandy": "",
+    "hang": "hæŋ",
+    "hangar": "ˈhæŋ.ər",
+    "hanging": "ˈhæŋ.ɪŋ",
+    "hape": "",
+    "haped": "",
+    "haping": "",
+    "happen": "ˈhæp.ən",
+    "happened": "ˈhæp.ən",
+    "happening": "ˈhæp.ən.ɪŋ",
+    "happier": "ˈhæp.i",
+    "happily": "ˈhæp.əl.i",
+    "happy": "ˈhæp.i",
+    "hard": "hɑːd",
+    "harder": "hɑːd",
+    "hardly": "ˈhɑːd.li",
+    "hare": "heər",
+    "hared": "heər",
+    "hark": "hɑːk",
+    "hasn't": "ˈhæz.ənt",
+    "hat": "hæt",
+    "hate": "heɪt",
+    "hateful": "ˈheɪt.fəl",
+    "have": "hæv",
+    "having": "",
+    "hazelnut": "ˈheɪ.zəl.nʌt",
+    "he": "hiː",
+    "he'd": "hiːd",
+    "he'll": "hiːl",
+    "head": "hed",
+    "headache": "ˈhed.eɪk",
+    "heal": "hiːl",
+    "healer": "ˈhɪə.lər",
+    "healing": "ˈhiː.lɪŋ",
+    "healthy": "ˈhel.θi",
+    "heap": "hiːp",
+    "hear": "hɪər",
+    "heard": "hɪər",
+    "hearing": "ˈhɪə.rɪŋ",
+    "heart": "hɑːt",
+    "hearty": "ˈhɑː.ti",
+    "heat": "hiːt",
+    "heather": "ˈheð.ər",
+    "heaven": "ˈhev.ən",
+    "heavy": "ˈhev.i",
+    "heck": "hek",
+    "heep": "",
+    "height": "haɪt",
+    "held": "held",
+    "helf": "",
+    "helicopter": "ˈhel.ɪˌkɒp.tər",
+    "helium": "ˈhiː.li.əm",
+    "hell": "hel",
+    "hello": "heˈləʊ",
+    "help": "help",
+    "helped": "help",
+    "helping": "ˈhel.pɪŋ",
+    "her": "hɜːr",
+    "herbling": "",
+    "here": "hɪər",
+    "hero": "ˈhɪə.rəʊ",
+    "herself": "hɜːˈself",
+    "hey": "heɪ",
+    "hh": "",
+    "hi": "haɪ",
+    "hiccup": "ˈhɪk.ʌp",
+    "hiccuping": "ˈhɪk.ʌp",
+    "hid": "",
+    "hide": "haɪd",
+    "high": "haɪ",
+    "higher": "ˈhaɪ.ər",
+    "hill": "hɪl",
+    "him": "hɪm",
+    "himself": "hɪmˈself",
+    "hining": "",
+    "hiny": "",
+    "hip": "hɪp",
+    "hippo": "ˌhɪp.əˈpɒt.ə.məs",
+    "hirt": "",
+    "hit": "hɪt",
+    "hitting": "",
+    "hiya": "ˈhaɪ.jə",
+    "hmm": "həm",
+    "hocked": "hɒk",
+    "hoe": "həʊ",
+    "hogging": "hɒɡ",
+    "hold": "həʊld",
+    "holded": "",
+    "holder": "ˈhəʊl.dər",
+    "holding": "ˈhəʊl.dɪŋ",
+    "hole": "həʊl",
+    "holocaust": "ˈhɒl.ə.kɔːst",
+    "home": "həʊm",
+    "honey": "ˈhʌn.i",
+    "hook": "hʊk",
+    "hooked": "hʊkt",
+    "hooking": "hʊk",
+    "hoop": "huːp",
+    "hooray": "həˈreɪ",
+    "hoot": "huːt",
+    "hooted": "huːt",
+    "hooting": "huːt",
+    "hop": "hɒp",
+    "hope": "həʊp",
+    "hopele": "",
+    "hopkeeper": "",
+    "hopped": "",
+    "hopper": "ˈhɒp.ər",
+    "hopperoo": "",
+    "hopping": "",
+    "hore": "",
+    "horn": "hɔːn",
+    "horrified": "ˈhɒr.ɪ.faɪd",
+    "horse": "hɔːs",
+    "horsey": "ˈhɔː.si",
+    "hort": "",
+    "horter": "",
+    "hospital": "ˈhɒs.pɪ.təl",
+    "hot": "hɒt",
+    "hould": "",
+    "houlder": "",
+    "houldn't": "",
+    "hour": "aʊər",
+    "house": "haʊs",
+    "hout": "",
+    "houted": "",
+    "houting": "",
+    "hove": "həʊv",
+    "hovel": "ˈhɒv.əl",
+    "hoveled": "",
+    "hoveling": "",
+    "how": "haʊ",
+    "how'd": "",
+    "howed": "",
+    "hower": "",
+    "howing": "",
+    "hrimp": "",
+    "hrink": "",
+    "hrugged": "",
+    "hrugging": "",
+    "hrunk": "",
+    "hudder": "",
+    "huffle": "",
+    "hug": "hʌɡ",
+    "huge": "hjuːdʒ",
+    "hugged": "",
+    "hugging": "",
+    "huh": "hə",
+    "human": "ˈhjuː.mən",
+    "hump": "hʌmp",
+    "hundred": "ˈhʌn.drəd",
+    "hungrier": "ˈhʌŋ.ɡri",
+    "hungriest": "ˈhʌŋ.ɡri",
+    "hungry": "ˈhʌŋ.ɡri",
+    "hunt": "hʌnt",
+    "hunting": "ˈhʌn.tɪŋ",
+    "hurl": "hɜːl",
+    "hurried": "ˈhʌr.id",
+    "hurry": "ˈhʌr.i",
+    "hurt": "hɜːt",
+    "hurted": "",
+    "hurting": "hɜːt",
+    "husband": "ˈhʌz.bənd",
+    "hut": "hʌt",
+    "hutting": "",
+    "hy": "",
+    "hypnotized": "ˈhɪp.nə.taɪz",
+    "i": "aɪ",
+    "i'll": "aɪl",
+    "i'm": "aɪm",
+    "i've": "aɪv",
+    "ice": "aɪs",
+    "ick": "ɪk",
+    "ick.\"": "ɪk",
+    "icker": "",
+    "ide": "",
+    "idea": "aɪˈdɪə",
+    "identify": "aɪˈden.tɪ.faɪ",
+    "idetracked": "",
+    "idewalk": "",
+    "if": "ɪf",
+    "ighed": "",
+    "ight": "",
+    "ighted": "",
+    "ign": "",
+    "ignore": "ɪɡˈnɔːr",
+    "ignoring": "ɪɡˈnɔːr",
+    "ill": "ɪl",
+    "illier": "",
+    "illy": "",
+    "ilver": "",
+    "imitate": "ˈɪm.ɪ.teɪt",
+    "impact": "ˈɪm.pækt",
+    "imple": "",
+    "important": "ɪmˈpɔː.tənt",
+    "impossible": "ɪmˈpɒs.ə.bəl",
+    "impressed": "ɪmˈpres",
+    "in": "ɪn",
+    "in-charge": "",
+    "ince": "",
+    "inch": "ɪntʃ",
+    "indoor": "ˌɪnˈdɔːr",
+    "infant": "ˈɪn.fənt",
+    "infected": "ɪnˈfek.tɪd",
+    "ing": "-ɪŋ",
+    "inging": "",
+    "ingle": "",
+    "ink": "ɪŋk",
+    "inked": "ɪŋk",
+    "inking": "ɪŋk",
+    "inning": "ˈɪn.ɪŋ",
+    "innocence": "ˈɪn.ə.səns",
+    "innocent": "ˈɪn.ə.sənt",
+    "innocently": "ˈɪn.ə.sənt.li",
+    "inside": "ɪnˈsaɪd",
+    "install": "ɪnˈstɔːl",
+    "instead": "ɪnˈsted",
+    "interested": "ˈɪn.tres.tɪd",
+    "interesting": "ˈɪn.tres.tɪŋ",
+    "interrupt": "ˌɪn.təˈrʌpt",
+    "interruption": "ˌɪn.təˈrʌp.ʃən",
+    "into": "ˈɪn.tuː",
+    "involve": "ɪnˈvɒlv",
+    "ir": "",
+    "ir,\"": "",
+    "isn't": "ˈɪz.ənt",
+    "ister": "",
+    "it": "ɪt",
+    "it'd": "ˈɪt.əd",
+    "it'll": "ˈɪt.əl",
+    "ite": "-aɪt",
+    "item": "ˈaɪ.təm",
+    "itself": "ɪtˈself",
+    "itting": "",
+    "ix": "",
+    "ix-year-old": "",
+    "ize": "-aɪz",
+    "jack": "dʒæk",
+    "jackal": "ˈdʒæk.əl",
+    "jacket": "ˈdʒæk.ɪt",
+    "jam": "dʒæm",
+    "jar": "dʒɑːr",
+    "jason": "",
+    "jaw": "dʒɔː",
+    "jealou": "",
+    "jennifer": "",
+    "jet": "dʒet",
+    "jinkly": "",
+    "job": "dʒɒb",
+    "jock": "dʒɒk",
+    "joe": "dʒəʊ",
+    "joey": "ˈdʒəʊ.i",
+    "jog": "dʒɒɡ",
+    "jogged": "",
+    "jogging": "ˈdʒɒɡ.ɪŋ",
+    "john": "ˈdʒɒn",
+    "join": "dʒɔɪn",
+    "joined": "dʒɔɪn",
+    "joke": "dʒəʊk",
+    "journal": "ˈdʒɜː.nəl",
+    "joy": "dʒɔɪ",
+    "joyed": "",
+    "judge": "dʒʌdʒ",
+    "juggle": "ˈdʒʌɡ.əl",
+    "juggling": "ˈdʒʌɡ.lɪŋ",
+    "juice": "dʒuːs",
+    "jumbled": "ˈdʒʌm.bəl",
+    "jump": "dʒʌmp",
+    "jump-jump": "dʒʌmp",
+    "jumped": "dʒʌmp",
+    "jumper": "ˈdʒʌm.pər",
+    "jumping": "dʒʌmp",
+    "junk": "dʒʌŋk",
+    "just": "dʒʌst",
+    "kangaroo": "ˌkæŋ.ɡərˈuː",
+    "keep": "kiːp",
+    "keeper": "ˈkiː.pər",
+    "keeping": "ˈkiː.pɪŋ",
+    "kept": "kept",
+    "ki": "",
+    "kick": "kɪk",
+    "kicked": "kɪk",
+    "kid": "kɪd",
+    "kidding": "",
+    "kie": "",
+    "kill": "kɪl",
+    "killed": "kɪl",
+    "killer": "ˈkɪl.ər",
+    "killing": "ˈkɪl.ɪŋ",
+    "kind": "kaɪnd",
+    "kinda": "ˈkaɪ.ndə",
+    "kindergarten": "ˈkɪn.dəˌɡɑː.tən",
+    "king": "kɪŋ",
+    "kinned": "",
+    "kinny": "",
+    "kip": "kɪp",
+    "kipped": "kɪp",
+    "kit": "kɪt",
+    "kitten": "ˈkɪt.ən",
+    "kitty": "ˈkɪ.ti",
+    "kleenex": "ˈkliː.neks",
+    "knee": "niː",
+    "knee-haw": "",
+    "kneecap": "ˈniː.kæp",
+    "kneeled": "niːl",
+    "kneeling": "niːl",
+    "knew": "njuː",
+    "knife": "naɪf",
+    "knock": "nɒk",
+    "knocked": "nɒk",
+    "knocking": "nɒk",
+    "knot": "nɒt",
+    "know": "nəʊ",
+    "ky": "",
+    "ladder": "ˈlæd.ər",
+    "lady": "ˈleɪ.di",
+    "laid": "leɪd",
+    "lake": "leɪk",
+    "lala": "",
+    "lamb": "læm",
+    "lame": "leɪm",
+    "lammed": "",
+    "land": "lænd",
+    "landed": "ˈlæn.dɪd",
+    "lane": "leɪn",
+    "lard": "lɑːd",
+    "last": "lɑːst",
+    "late": "leɪt",
+    "later": "ˈleɪ.tər",
+    "laugh": "lɑːf",
+    "laughed": "lɑːf",
+    "laugher": "",
+    "laughing": "lɑːf",
+    "law": "lɔː",
+    "lawn": "lɔːn",
+    "lay": "leɪ",
+    "laying": "leɪ",
+    "lazy": "ˈleɪ.zi",
+    "lead": "liːd",
+    "leaned": "liːn",
+    "leapt": "",
+    "learn": "lɜːn",
+    "learned": "ˈlɜː.nɪd",
+    "learning": "ˈlɜː.nɪŋ",
+    "leash": "liːʃ",
+    "least": "liːst",
+    "leather": "ˈleð.ər",
+    "leave": "liːv",
+    "leaving": "",
+    "led": "led",
+    "leep": "",
+    "leeping": "",
+    "leepy": "",
+    "leeve": "",
+    "left": "left",
+    "leg": "leɡ",
+    "lemonade": "ˌlem.əˈneɪd",
+    "len": "",
+    "lend": "lend",
+    "lept": "",
+    "lesson": "ˈles.ən",
+    "let": "let",
+    "letter": "ˈlet.ər",
+    "letting": "ˈlet.ɪŋ",
+    "library": "ˈlaɪ.brər.i",
+    "lick": "lɪk",
+    "licked": "lɪk",
+    "lide": "",
+    "liding": "",
+    "lie": "laɪ",
+    "lied": "laɪ",
+    "life": "laɪf",
+    "lifeguard": "ˈlaɪf.ɡɑːd",
+    "lifted": "lɪft",
+    "light": "laɪt",
+    "like": "laɪk",
+    "liked": "laɪk",
+    "limo": "ˈlɪm.əʊ",
+    "line": "laɪn",
+    "liney": "",
+    "ling": "",
+    "lining": "ˈlaɪ.nɪŋ",
+    "lion": "ˈlaɪ.ən",
+    "lip": "lɪp",
+    "lipped": "ˌtaɪtˈlɪpt",
+    "lippery": "",
+    "lipping": "ˈlɪp.ɪŋ",
+    "listen": "ˈlɪs.ən",
+    "listening": "ˈlɪs.ən",
+    "lit": "lɪt",
+    "literally": "ˈlɪt.ər.əl.i",
+    "little": "ˈlɪt.əl",
+    "littler": "ˈlɪt.əl",
+    "live": "lɪv",
+    "lived": "lɪv",
+    "locked": "lɒk",
+    "lonely": "ˈləʊn.li",
+    "long": "lɒŋ",
+    "longer": "lɒŋ",
+    "look": "lʊk",
+    "looked": "lʊk",
+    "looking": "ˈlʊk.ɪŋ",
+    "loop": "luːp",
+    "loop-de-loop": "ˌluːp ðə ˈluːp",
+    "loose": "luːs",
+    "loosened": "ˈluː.sən",
+    "lop": "lɑp",
+    "lose": "luːz",
+    "losing": "",
+    "lost": "lɒst",
+    "lot": "lɒt",
+    "loud": "laʊd",
+    "louder": "laʊd",
+    "loudly": "ˈlaʊd.li",
+    "love": "lʌv",
+    "loved": "lʌv",
+    "lovely": "ˈlʌv.li",
+    "loving": "ˈlʌv.ɪŋ",
+    "low": "ləʊ",
+    "lower": "ˈləʊ.ər",
+    "lowing": "ləʊ",
+    "lowly": "ˈləʊ.li",
+    "luck": "lʌk",
+    "luckily": "ˈlʌk.əl.i",
+    "lucky": "ˈlʌk.i",
+    "lumen": "ˈluː.mən",
+    "lump": "lʌmp",
+    "lunch": "lʌntʃ",
+    "lunchboxe": "",
+    "lunche": "",
+    "lurred": "",
+    "lying": "ˈlaɪ.ɪŋ",
+    "machine": "məˈʃiːn",
+    "macked": "",
+    "mad": "mæd",
+    "madder": "",
+    "made": "meɪd",
+    "madly": "ˈmæd.li",
+    "magic": "ˈmædʒ.ɪk",
+    "maid": "meɪd",
+    "mailman": "ˈmeɪl.mæn",
+    "make": "meɪk",
+    "maker": "ˈmeɪ.kər",
+    "making": "ˈmeɪ.kɪŋ",
+    "mall": "mɔːl",
+    "maller": "",
+    "mama": "məˈmɑː",
+    "man": "mæn",
+    "manager": "ˈmæn.ɪ.dʒər",
+    "manned": "mænd",
+    "many": "ˈmen.i",
+    "marble": "ˈmɑː.bəl",
+    "mark": "mɑːk",
+    "market": "ˈmɑː.kɪt",
+    "mart": "mɑːt",
+    "mashed": "mæʃt",
+    "mask": "mɑːsk",
+    "master": "ˈmɑː.stər",
+    "mat": "mæt",
+    "matter": "ˈmæt.ər",
+    "may": "meɪ",
+    "maybe": "ˈmeɪ.bi",
+    "me": "miː",
+    "meal": "mɪəl",
+    "mean": "miːn",
+    "meant": "",
+    "meantime": "ˈmiːn.taɪm",
+    "meat": "miːt",
+    "med": "med",
+    "medic": "ˈmed.ɪk",
+    "medical": "ˈmed.ɪ.kəl",
+    "medicine": "ˈmed.ɪ.sən",
+    "medium": "ˈmiː.di.əm",
+    "meet": "miːt",
+    "meeting": "ˈmiː.tɪŋ",
+    "mell": "ˌpelˈmel",
+    "melt": "melt",
+    "melted": "ˈmel.tɪd",
+    "memory": "ˈmem.ər.i",
+    "men": "men",
+    "mercie": "",
+    "messed": "mes",
+    "messy": "ˈmes.i",
+    "met": "met",
+    "metal": "ˈmet.əl",
+    "mhm": "",
+    "mi": "miː",
+    "mice": "maɪs",
+    "microphone": "ˈmaɪ.krə.fəʊn",
+    "middle": "ˈmɪd.əl",
+    "might": "maɪt",
+    "mike": "maɪk",
+    "mikey": "",
+    "mile": "maɪl",
+    "miled": "",
+    "miley": "",
+    "miling": "",
+    "milk": "mɪlk",
+    "milker": "ˈmɪl.kər",
+    "milled": "mɪl",
+    "mind": "maɪnd",
+    "minding": "maɪnd",
+    "mine": "maɪn",
+    "minor": "ˈmaɪ.nər",
+    "mint": "mɪnt",
+    "minute": "ˈmɪn.ɪt",
+    "mirror": "ˈmɪr.ər",
+    "missed": "mɪs",
+    "missing": "ˈmɪs.ɪŋ",
+    "mistake": "mɪˈsteɪk",
+    "mister": "ˈmɪs.tər",
+    "misty": "ˈmɪs.ti",
+    "mix": "mɪks",
+    "mixed": "mɪkst",
+    "mixing": "mɪks",
+    "mm": "",
+    "mmm": "",
+    "model": "ˈmɒd.əl",
+    "moke": "",
+    "mold": "məʊld",
+    "molding": "ˈməʊl.dɪŋ",
+    "mole": "məʊl",
+    "mom": "mɒm",
+    "mommy": "ˈmɒm.i",
+    "money": "ˈmʌn.i",
+    "moneymaker": "ˈmʌn.iˌmeɪ.kər",
+    "monkey": "ˈmʌŋ.ki",
+    "monster": "ˈmɒn.stər",
+    "moon": "muːn",
+    "moose": "muːs",
+    "moothed": "",
+    "moother": "",
+    "more": "mɔːr",
+    "morning": "ˈmɔː.nɪŋ",
+    "mortar": "ˈmɔː.tər",
+    "most": "məʊst",
+    "mostly": "ˈməʊst.li",
+    "mother": "ˈmʌð.ər",
+    "motorcycle": "ˈməʊ.təˌsaɪ.kəl",
+    "mound": "maʊnd",
+    "mountain": "ˈmaʊn.tɪn",
+    "mouse": "maʊs",
+    "mouth": "maʊθ",
+    "move": "muːv",
+    "moved": "muːvd",
+    "movie": "ˈmuː.vi",
+    "moving": "ˈmuː.vɪŋ",
+    "mr": "ˈmɪs.tər",
+    "much": "mʌtʃ",
+    "mud": "mʌd",
+    "muller": "",
+    "mum": "mʌm",
+    "mummy": "ˈmʌm.i",
+    "munching": "mʌntʃ",
+    "munchy": "",
+    "muscle": "ˈmʌs.əl",
+    "mushed": "mʌʃ",
+    "music": "ˈmjuː.zɪk",
+    "must": "mʌst",
+    "mustache": "məˈstɑːʃ",
+    "mustard": "ˈmʌs.təd",
+    "my": "maɪ",
+    "myself": "maɪˈself",
+    "nack": "ˈnɪk.næk",
+    "nah": "næː",
+    "nail": "neɪl",
+    "naked": "ˈneɪ.kɪd",
+    "name": "neɪm",
+    "named": "neɪm",
+    "nap": "næp",
+    "napped": "",
+    "natche": "",
+    "natched": "",
+    "naughty": "ˈnɔː.ti",
+    "neak": "",
+    "neaked": "",
+    "near": "nɪər",
+    "nearby": "ˌnɪəˈbaɪ",
+    "nearly": "ˈnɪə.li",
+    "neat": "niːt",
+    "neck": "nek",
+    "necklace": "ˈnek.ləs",
+    "need": "niːd",
+    "needed": "ˈniː.dɪd",
+    "neezed": "",
+    "neither": "ˈnaɪ.ðər",
+    "nervou": "",
+    "net": "net",
+    "never": "ˈnev.ər",
+    "new": "njuː",
+    "next": "nekst",
+    "nice": "naɪs",
+    "nicely": "ˈnaɪs.li",
+    "niche": "niːʃ",
+    "nickel": "ˈnɪk.əl",
+    "night": "naɪt",
+    "nightly": "ˈnaɪt.li",
+    "nighttime": "ˈnaɪt.taɪm",
+    "nine": "naɪn",
+    "ninja": "ˈnɪn.dʒə",
+    "no": "nəʊ",
+    "nobody": "ˈnəʊ.bə.di",
+    "nod": "nɒd",
+    "noise": "nɔɪz",
+    "non-running": "",
+    "none": "nʌn",
+    "noodle": "ˈnuː.dəl",
+    "nope": "nəʊp",
+    "nor-ail": "",
+    "normal": "ˈnɔː.məl",
+    "north": "nɔːθ",
+    "nose": "nəʊz",
+    "not": "nɒt",
+    "note": "nəʊt",
+    "nothing": "ˈnʌθ.ɪŋ",
+    "notice": "ˈnəʊ.tɪs",
+    "noticed": "ˈnəʊ.tɪs",
+    "noticing": "ˈnəʊ.tɪs",
+    "now": "naʊ",
+    "nowy": "",
+    "nuggled": "",
+    "number": "ˈnʌm.bər",
+    "nurse": "nɜːs",
+    "nut": "nʌt",
+    "nutrition": "njuːˈtrɪʃ.ən",
+    "o": "əʊ",
+    "o'clock": "əˈklɒk",
+    "oaked": "",
+    "oaking": "",
+    "oap": "ˌəʊ.eɪˈpiː",
+    "oaring": "",
+    "object": "ˈɒb.dʒɪkt",
+    "observation": "ˌɒb.zəˈveɪ.ʃən",
+    "occer": "",
+    "ocean": "ˈəʊ.ʃən",
+    "octopu": "",
+    "of": "əv",
+    "off": "ɒf",
+    "offered": "ˈɒf.ər",
+    "office": "ˈɒf.ɪs",
+    "officer": "ˈɒf.ɪ.sər",
+    "oftly": "",
+    "oh": "əʊ",
+    "okay": "ˌəʊˈkeɪ",
+    "old": "əʊld",
+    "older": "əʊld",
+    "olid": "",
+    "olve": "",
+    "ome": "",
+    "omebody": "",
+    "omeone": "",
+    "omeplace": "",
+    "omething": "",
+    "ometime": "",
+    "omewhere": "",
+    "on": "ɒn",
+    "on-stand": "",
+    "once": "wʌns",
+    "one": "wʌn",
+    "only": "ˈəʊn.li",
+    "onto": "ˈɒn.tu",
+    "ooh": "uː",
+    "oon": "",
+    "oop": "ˌæl.iˈuːp",
+    "oops-a-daisy": "ˈʊp.səˌdeɪ.zi",
+    "oopsie": "",
+    "open": "ˈəʊ.pən",
+    "opened": "ˈəʊ.pən",
+    "opening": "ˈəʊ.pən.ɪŋ",
+    "operating": "ˈɒp.ər.eɪ.tɪŋ",
+    "or": "ɔːr",
+    "orange": "ˈɒr.ɪndʒ",
+    "order": "ˈɔː.dər",
+    "ordinary": "ˈɔː.dən.əri",
+    "ore": "ɔːr",
+    "orphan": "ˈɔː.fən",
+    "orry": "",
+    "ort": "ɔːt",
+    "orta": "",
+    "other": "ˈʌð.ər",
+    "ouch": "aʊtʃ",
+    "ound": "",
+    "ounding": "",
+    "oup": "",
+    "our": "aʊər",
+    "ource": "",
+    "out": "aʊt",
+    "outbreak": "ˈaʊt.breɪk",
+    "outdoor": "ˈaʊtˌdɔːr",
+    "outer": "ˈaʊ.tər",
+    "outfit": "ˈaʊt.fɪt",
+    "outside": "ˌaʊtˈsaɪd",
+    "oval": "ˈəʊ.vəl",
+    "over": "ˈəʊ.vər",
+    "ow": "aʊ",
+    "owie": "",
+    "owl": "aʊl",
+    "own": "əʊn",
+    "owned": "-əʊnd",
+    "owner": "ˈəʊ.nər",
+    "oy": "ɔɪ",
+    "pa": "pɑː",
+    "pace": "peɪs",
+    "pack": "pæk",
+    "package": "ˈpæk.ɪdʒ",
+    "packed": "pækt",
+    "packing": "ˈpæk.ɪŋ",
+    "paddle": "ˈpæd.əl",
+    "page": "peɪdʒ",
+    "paid": "peɪd",
+    "pail": "peɪl",
+    "pain": "peɪn",
+    "painkiller": "ˈpeɪnˌkɪl.ər",
+    "paint": "peɪnt",
+    "painted": "peɪnt",
+    "pair": "peər",
+    "panel": "ˈpæn.əl",
+    "panked": "",
+    "pant": "pænt",
+    "panting": "pænt",
+    "paper": "ˈpeɪ.pər",
+    "parade": "pəˈreɪd",
+    "paramedic": "ˌpær.əˈmed.ɪk",
+    "parent": "ˈpeə.rənt",
+    "park": "pɑːk",
+    "part": "pɑːt",
+    "party": "ˈpɑː.ti",
+    "passe": "pɑːˈseɪ",
+    "passed": "pɑːs",
+    "passing": "ˈpɑː.sɪŋ",
+    "past": "pɑːst",
+    "pat": "pæt",
+    "patch": "pætʃ",
+    "patched": "pætʃ",
+    "path": "pɑːθ",
+    "patient": "ˈpeɪ.ʃənt",
+    "patting": "",
+    "pavement": "ˈpeɪv.mənt",
+    "paw": "pɔː",
+    "pay": "peɪ",
+    "payer": "ˈpeɪ.ər",
+    "paying": "peɪ",
+    "peak": "piːk",
+    "peaking": "piːk",
+    "peanut": "ˈpiː.nʌt",
+    "pear": "peər",
+    "pearl": "pɜːl",
+    "pecial": "",
+    "pee": "piː",
+    "peel": "piːl",
+    "peeled": "piːl",
+    "pell": "ˌpelˈmel",
+    "pelled": "",
+    "pen": "pen",
+    "pennie": "",
+    "people": "ˈpiː.pəl",
+    "per": "pɜːr",
+    "percent": "pəˈsent",
+    "perfect": "ˈpɜː.fekt",
+    "perhap": "",
+    "person": "ˈpɜː.sən",
+    "personally": "ˈpɜː.sən.əl.i",
+    "pet": "pet",
+    "petting": "",
+    "phone": "fəʊn",
+    "phoned": "fəʊn",
+    "phrase": "freɪz",
+    "pic": "pɪk",
+    "pick": "pɪk",
+    "picked": "pɪk",
+    "picking": "ˈpɪkɪŋz",
+    "pickle": "ˈpɪk.əl",
+    "picnic": "ˈpɪk.nɪk",
+    "picnick": "",
+    "picnicking": "",
+    "picture": "ˈpɪk.tʃər",
+    "pie": "paɪ",
+    "piece": "piːs",
+    "pig": "pɪɡ",
+    "pig-ish": "",
+    "piggy": "ˈpɪɡ.i",
+    "pike": "paɪk",
+    "pile": "paɪl",
+    "piled": "paɪl",
+    "pill": "pɪl",
+    "pilled": "pɪl",
+    "pilling": "pɪl",
+    "pillow": "ˈpɪl.əʊ",
+    "pin": "pɪn",
+    "pinch": "pɪntʃ",
+    "pine": "paɪn",
+    "ping": "pɪŋ",
+    "pink": "pɪŋk",
+    "pinkie": "ˈpɪŋ.ki",
+    "pinkling": "",
+    "pinning": "",
+    "pinny": "ˈpɪn.i",
+    "pipe": "paɪp",
+    "pipeline": "ˈpaɪp.laɪn",
+    "pit": "pɪt",
+    "pitch": "pɪtʃ",
+    "pitting": "",
+    "pizza": "ˈpiːt.sə",
+    "place": "pleɪs",
+    "plague": "pleɪɡ",
+    "plain": "pleɪn",
+    "plan": "plæn",
+    "plane": "pleɪn",
+    "planning": "ˈplæn.ɪŋ",
+    "plant": "plɑːnt",
+    "plash": "plæʃ",
+    "plashed": "plæʃ",
+    "plashing": "ˈplæʃ.ɪŋ",
+    "plastic": "ˈplæs.tɪk",
+    "plate": "pleɪt",
+    "platform": "ˈplæt.fɔːm",
+    "play": "pleɪ",
+    "played": "pleɪ",
+    "player": "ˈpleɪ.ər",
+    "playful": "ˈpleɪ.fəl",
+    "playing": "pleɪ",
+    "playpen": "ˈpleɪ.pen",
+    "please": "pliːz",
+    "pleased": "pliːzd",
+    "pleasure": "ˈpleʒ.ər",
+    "plight": "plaɪt",
+    "plinch": "",
+    "plit": "",
+    "plop": "plɒp",
+    "plopped": "",
+    "plu": "",
+    "pluck": "plʌk",
+    "plugged": "",
+    "plushed": "",
+    "plying": "plaɪ",
+    "pocket": "ˈpɒk.ɪt",
+    "point": "pɔɪnt",
+    "pointed": "ˈpɔɪn.tɪd",
+    "pointing": "ˈpɔɪn.tɪŋ",
+    "pointy": "ˈpɔɪn.ti",
+    "poison": "ˈpɔɪ.zən",
+    "poke": "pəʊk",
+    "pole": "pəʊl",
+    "policy": "ˈpɒl.ə.si",
+    "politely": "pəˈlaɪt.li",
+    "polka": "ˈpɒl.kə",
+    "pond": "pɒnd",
+    "pony": "ˈpəʊ.ni",
+    "ponytail": "ˈpəʊ.ni.teɪl",
+    "poo": "puː",
+    "pooch": "puːtʃ",
+    "poochie": "",
+    "poof": "pʊf",
+    "poofing": "",
+    "pooing": "puː",
+    "pooky": "",
+    "pool": "puːl",
+    "poon": "",
+    "poop": "puːp",
+    "pooped": "puːpt",
+    "poor": "pɔːr",
+    "pop": "pɒp",
+    "popcorn": "ˈpɒp.kɔːn",
+    "popped": "",
+    "popping": "",
+    "popsicle": "ˈpɒp.sɪ.kəl",
+    "porch": "pɔːtʃ",
+    "port": "pɔːt",
+    "possible": "ˈpɒs.ə.b��l",
+    "pot": "pɒt",
+    "potato": "pəˈteɪ.təʊ",
+    "potatoe": "",
+    "potted": "ˈpɒt.ɪd",
+    "pour": "pɔːr",
+    "poured": "pɔːr",
+    "pouring": "pɔːr",
+    "power": "paʊər",
+    "practically": "ˈpræk.tɪ.kəl.i",
+    "practice": "ˈpræk.tɪs",
+    "practicing": "ˈpræk.tɪ.sɪŋ",
+    "prayed": "preɪ",
+    "pre": "priː-",
+    "pread": "",
+    "pregnant": "ˈpreɡ.nənt",
+    "prepare": "prɪˈpeər",
+    "prepared": "prɪˈpeəd",
+    "present": "ˈprez.ənt",
+    "pressed": "prest",
+    "pressing": "ˈpres.ɪŋ",
+    "pretend": "prɪˈtend",
+    "pretended": "prɪˈten.dɪd",
+    "pretending": "prɪˈtend",
+    "pretty": "ˈprɪt.i",
+    "preview": "ˈpriː.vjuː",
+    "pring": "",
+    "pringboard": "",
+    "pringtime": "",
+    "print": "prɪnt",
+    "prize": "praɪz",
+    "probably": "ˈprɒb.ə.bli",
+    "problem": "ˈprɒb.ləm",
+    "proceed": "prəˈsiːd",
+    "product": "ˈprɒd.ʌkt",
+    "propeller": "prəˈpel.ər",
+    "prototype": "ˈprəʊ.tə.taɪp",
+    "proud": "praʊd",
+    "puffer": "ˈpʌf.ər",
+    "puffing": "pʌf",
+    "puffy": "ˈpʌf.i",
+    "puke": "pjuːk",
+    "pull": "pʊl",
+    "pulled": "pʊld",
+    "puller": "ˈkraʊdˌpʊl.ər",
+    "pulling": "pʊl",
+    "pumped": "pʌmpt",
+    "pup": "pʌp",
+    "puppet": "ˈpʌp.ɪt",
+    "puppie": "",
+    "puppy": "ˈpʌp.i",
+    "purple": "ˈpɜː.pəl",
+    "purpose": "ˈpɜː.pəs",
+    "purse": "pɜːs",
+    "push": "pʊʃ",
+    "pushed": "pʊʃt",
+    "pushing": "ˈpʊʃ.ɪŋ",
+    "put": "pʊt",
+    "putting": "",
+    "puzzle": "ˈpʌz.əl",
+    "puzzled": "ˈpʌz.əld",
+    "pyramid": "ˈpɪr.ə.mɪd",
+    "quarter": "ˈkwɔː.tər",
+    "queasy": "ˈkwiː.zi",
+    "queeze": "",
+    "queezed": "",
+    "question": "ˈkwes.tʃən",
+    "quick": "kwɪk",
+    "quicker": "kwɪk",
+    "quickly": "ˈkwɪk.li",
+    "quiet": "ˈkwaɪ.ət",
+    "quietly": "ˈkwaɪət.li",
+    "quint": "kwɪnt",
+    "quirting": "",
+    "quite": "kwaɪt",
+    "ra-": "ɑː ˈeɪ",
+    "rabbi": "ˈræb.aɪ",
+    "rabbit": "ˈræb.ɪt",
+    "rabid": "ˈræb.ɪd",
+    "race": "reɪs",
+    "raced": "reɪs",
+    "racing": "ˈreɪ.sɪŋ",
+    "rack": "ræk",
+    "raft": "rɑːft",
+    "rain": "reɪn",
+    "rained": "reɪn",
+    "raining": "reɪn",
+    "raising": "reɪz",
+    "rake": "reɪk",
+    "raked": "reɪkt",
+    "ran": "ræn",
+    "rang": "ræŋ",
+    "raptor": "ˈræp.tər",
+    "raspberrie": "",
+    "rat": "ræt",
+    "rating": "ˈreɪ.tɪŋ",
+    "raved": "reɪv",
+    "raver": "ˈreɪ.vər",
+    "re-jump": "",
+    "reach": "riːtʃ",
+    "reache": "",
+    "reached": "riːtʃ",
+    "reaching": "riːtʃ",
+    "read": "riːd",
+    "reading": "ˈriː.dɪŋ",
+    "ready": "ˈred.i",
+    "real": "rɪəl",
+    "realize": "ˈrɪə.laɪz",
+    "realized": "ˈrɪəlaɪzd",
+    "really": "ˈrɪə.li",
+    "reason": "ˈriː.zən",
+    "rebuild": "ˌriːˈbɪld",
+    "rebuilt": "",
+    "rece": "",
+    "receipt": "rɪˈsiːt",
+    "recognize": "ˈrek.əɡ.naɪz",
+    "record": "rɪˈkɔːd",
+    "recorder": "rɪˈkɔː.dər",
+    "recording": "rɪˈkɔː.dɪŋ",
+    "recursion": "rɪˈkɜː.ʒən",
+    "red": "red",
+    "reef": "riːf",
+    "reform": "rɪˈfɔːm",
+    "release": "rɪˈliːs",
+    "relieved": "rɪˈliːvd",
+    "remember": "rɪˈmem.bər",
+    "remembered": "rɪˈmem.bər",
+    "remembering": "rɪˈmem.bər",
+    "remind": "rɪˈmaɪnd",
+    "reminded": "rɪˈmaɪnd",
+    "remote": "rɪˈməʊt",
+    "rent": "rent",
+    "repeating": "rɪˈpiːt",
+    "replace": "rɪˈpleɪs",
+    "rescue": "ˈres.kjuː",
+    "resistance": "rɪˈzɪs.təns",
+    "resource": "rɪˈzɔːs",
+    "rest": "rest",
+    "restaurant": "ˈres.tər.ɒnt",
+    "rested": "ˈres.tɪd",
+    "returned": "rɪˈtɜːnd",
+    "returning": "rɪˈtɜːn",
+    "rex": "reks",
+    "rhino": "ˈraɪ.nəʊ",
+    "ribbon": "ˈrɪb.ən",
+    "ride": "raɪd",
+    "riding": "ˈraɪ.dɪŋ",
+    "right": "raɪt",
+    "ring": "rɪŋ",
+    "rip": "rɪp",
+    "ripping": "",
+    "rise": "raɪz",
+    "river": "ˈrɪv.ər",
+    "road": "rəʊd",
+    "robert": "ˌhɜːb ˈrɒb.ət",
+    "rock": "rɒk",
+    "rocket": "ˈrɒk.ɪt",
+    "rod": "rɒd",
+    "rolled": "rəʊl",
+    "roof": "ruːf",
+    "room": "ruːm",
+    "rooster": "ˈruː.stər",
+    "rope": "rəʊp",
+    "rose": "rəʊz",
+    "rotten": "ˈrɒt.ən",
+    "rough": "rʌf",
+    "round": "raʊnd",
+    "row": "rəʊ",
+    "rub": "rʌb",
+    "rubber": "ˈrʌb.ər",
+    "rubbery": "ˈrʌb.ər.i",
+    "rubbing": "",
+    "ruin": "ˈruː.ɪn",
+    "ruined": "ˈruː.ɪnd",
+    "ruining": "ˈruː.ɪn",
+    "rule": "ruːl",
+    "rumble": "ˈrʌm.bəl",
+    "run": "rʌn",
+    "runned": "",
+    "running": "ˈrʌn.ɪŋ",
+    "rush": "rʌʃ",
+    "rushe": "",
+    "rushed": "rʌʃt",
+    "rushing": "ˈrʌʃ.ɪŋ",
+    "t-shirt": "ˈtiː.ʃɜːt",
+    "table": "ˈteɪ.bəl",
+    "tack": "tæk",
+    "tacked": "tæk",
+    "taff": "ˈtæf.i",
+    "tag": "tæɡ",
+    "tage": "",
+    "tail": "teɪl",
+    "tained": "",
+    "take": "teɪk",
+    "taken": "ˈteɪ.kən",
+    "taking": "",
+    "talk": "tɔːk",
+    "talked": "tɔːk",
+    "talking": "tɔːk",
+    "tall": "tɔːl",
+    "taller": "tɔːl",
+    "tamp": "tæmp",
+    "tand": "",
+    "tanded": "",
+    "tanding": "",
+    "tangled": "ˈtæŋ.ɡəld",
+    "tank": "tæŋk",
+    "tape": "teɪp",
+    "taped": "teɪp",
+    "taping": "teɪp",
+    "tapped": "",
+    "tapping": "",
+    "tar": "tɑːr",
+    "tarantula": "təˈræn.tjə.lə",
+    "tare": "teər",
+    "tared": "",
+    "taring": "",
+    "tart": "tɑːt",
+    "tarted": "tɑːt",
+    "tarting": "tɑːt",
+    "tassel": "ˈtæs.əl",
+    "taste": "teɪst",
+    "taught": "tɔːt",
+    "taxi": "ˈtæk.si",
+    "tay": "teɪˈsæks dɪˌziːz",
+    "tayed": "",
+    "taying": "",
+    "tea": "tiː",
+    "teach": "tiːtʃ",
+    "teacher": "ˈtiː.tʃər",
+    "teal": "tiːl",
+    "tealing": "",
+    "team": "tiːm",
+    "teamed": "tiːm",
+    "teaming": "ˈtiːmɪŋ",
+    "tear": "teər",
+    "teased": "tiːz",
+    "teasing": "tiːz",
+    "teen": "tiːn",
+    "teeple": "",
+    "teeter": "ˈtiː.tər",
+    "teeth": "tiːθ",
+    "telescope": "ˈtel.ɪ.skəʊp",
+    "tell": "tel",
+    "telling": "ˈtel.ɪŋ",
+    "temperature": "ˈtem.prə.tʃər",
+    "ten": "ten",
+    "tendency": "ˈten.dən.si",
+    "tenni": "",
+    "tep": "",
+    "tepped": "",
+    "tepping": "",
+    "terrible": "ˈter.ə.bəl",
+    "terrified": "ˈter.ə.faɪd",
+    "test": "test",
+    "tested": "test",
+    "testing": "ˈtes.tɪŋ",
+    "than": "ðæn",
+    "thank": "θæŋk",
+    "thanked": "θæŋk",
+    "thankful": "ˈθæŋk.fəl",
+    "thanking": "θæŋk",
+    "that": "ðæt",
+    "the": "ðiː",
+    "theater": "ˈθɪə.tər",
+    "their": "ðeər",
+    "theirself": "",
+    "them": "ðem",
+    "themselve": "",
+    "then": "ðen",
+    "there": "ðeər",
+    "thermometer": "θəˈmɒm.ɪ.tər",
+    "these": "ðiːz",
+    "they": "ðeɪ",
+    "they'd": "ðeɪd",
+    "they'll": "ðeɪl",
+    "they're": "ðeər",
+    "they've": "ðeɪv",
+    "thi": "",
+    "thick": "θɪk",
+    "thing": "θɪŋ",
+    "thingy": "ˈθɪŋ.i",
+    "think": "θɪŋk",
+    "thinking": "ˈθɪŋ.kɪŋ",
+    "third": "θɜːd",
+    "thirsty": "ˈθɜː.sti",
+    "those": "ðəʊz",
+    "though": "ðəʊ",
+    "thought": "θɔːt",
+    "three": "θriː",
+    "threw": "θruː",
+    "thrilled": "θrɪld",
+    "throat": "θrəʊt",
+    "through": "θruː",
+    "throw": "θrəʊ",
+    "throwed": "",
+    "throwing": "θrəʊ",
+    "tick": "tɪk",
+    "ticked": "tɪk",
+    "ticker": "ˈtɪk.ər",
+    "ticket": "ˈtɪk.ɪt",
+    "tickie": "",
+    "ticking": "ˈtɪk.ɪŋ",
+    "ticky": "",
+    "tie": "taɪ",
+    "tied": "taɪ",
+    "tight": "taɪt",
+    "tightened": "ˈtaɪ.tən",
+    "tighter": "taɪt",
+    "till": "tɪl",
+    "tim": "",
+    "time": "taɪm",
+    "timing": "ˈtaɪ.mɪŋ",
+    "timmy": "",
+    "ting": "tɪŋ",
+    "tink": "",
+    "tinky": "",
+    "tiny": "ˈtaɪ.ni",
+    "tipped": "",
+    "tiptoed": "ˈtɪp.təʊ",
+    "tire": "taɪər",
+    "tired": "taɪəd",
+    "tirring": "",
+    "titan": "ˈtaɪ.tən",
+    "titch": "tɪtʃ",
+    "titched": "",
+    "title": "ˈtaɪ.təl",
+    "to": "tuː",
+    "tocky": "",
+    "today": "təˈdeɪ",
+    "toe": "təʊ",
+    "together": "təˈɡeð.ər",
+    "toilet": "ˈtɔɪ.lət",
+    "toiling": "tɔɪl",
+    "told": "təʊld",
+    "tole": "",
+    "tolling": "ˈtəʊ.lɪŋ",
+    "tomach": "",
+    "tomachache": "",
+    "tomorrow": "təˈmɒr.əʊ",
+    "ton": "tʌn",
+    "toner": "ˈtəʊ.nər",
+    "tongue": "tʌŋ",
+    "tonight": "təˈnaɪt",
+    "too": "tuː",
+    "tood": "",
+    "took": "tʊk",
+    "tool": "tuːl",
+    "top": "tɒp",
+    "topped": "",
+    "toppled": "ˈtɒp.əl",
+    "torage": "",
+    "tore": "tɔːr",
+    "torekeeper": "",
+    "torie": "",
+    "torm": "",
+    "torn": "tɔːn",
+    "tory": "ˈtɔː.ri",
+    "torytelling": "",
+    "totally": "ˈtəʊ.təl.i",
+    "totter": "ˈtɒt.ər",
+    "touch": "tʌtʃ",
+    "touche": "tuːˈʃeɪ",
+    "touched": "tʌtʃt",
+    "touching": "ˈtʌtʃ.ɪŋ",
+    "tour": "tʊər",
+    "tove": "",
+    "toward": "tɔːrd",
+    "towel": "taʊəl",
+    "tower": "taʊər",
+    "town": "taʊn",
+    "toy": "tɔɪ",
+    "track": "træk",
+    "trade": "treɪd",
+    "trader": "ˈtreɪ.dər",
+    "traight": "",
+    "traightening": "",
+    "trail": "treɪl",
+    "trailer": "ˈtreɪ.lər",
+    "trailing": "ˈtreɪ.lɪŋ",
+    "train": "treɪn",
+    "trained": "treɪnd",
+    "trampoline": "ˈtræm.pəl.iːn",
+    "trange": "",
+    "trangely": "",
+    "tranger": "",
+    "trap": "træp",
+    "trash": "træʃ",
+    "travel": "ˈtræv.əl",
+    "trawberrie": "",
+    "tray": "treɪ",
+    "treat": "triːt",
+    "tree": "triː",
+    "treet": "",
+    "trength": "",
+    "tretch": "",
+    "tretched": "",
+    "trick": "trɪk",
+    "trie": "",
+    "tried": "traɪd",
+    "trifle": "ˈtraɪ.fəl",
+    "trike": "traɪk",
+    "tring": "",
+    "trip": "trɪp",
+    "tripped": "",
+    "tripper": "ˈtrɪp.ər",
+    "tripping": "",
+    "trolled": "trəʊl",
+    "troller": "",
+    "trouble": "ˈtrʌb.əl",
+    "troubled": "ˈtrʌb.əld",
+    "truck": "trʌk",
+    "tructure": "",
+    "true": "truː",
+    "truggled": "",
+    "trunk": "trʌŋk",
+    "truth": "truːθ",
+    "try": "traɪ",
+    "trying": "ˈtraɪ.ɪŋ",
+    "tub": "tʌb",
+    "tubby": "ˈtʌb.i",
+    "tube": "tʃuːb",
+    "tuck": "tʌk",
+    "tucked": "tʌk",
+    "tucking": "tʌk",
+    "tudent": "",
+    "tuff": "tʌf",
+    "tuffed": "",
+    "tugging": "",
+    "tumble": "ˈtʌm.bəl",
+    "tumbled": "ˈtʌm.bəl",
+    "tumbling": "ˈtʌm.bəl",
+    "tummie": "",
+    "tummy": "ˈtʌm.i",
+    "tunk": "",
+    "tunnel": "ˈtʌn.əl",
+    "tupid": "",
+    "turkey": "ˈtɜː.ki",
+    "turn": "tɜːn",
+    "turned": "tɜːn",
+    "turning": "ˈtɜː.nɪŋ",
+    "turtle": "ˈtɜː.təl",
+    "tweak": "twiːk",
+    "twirl": "twɜːl",
+    "twirled": "twɜːl",
+    "twirling": "ˈtwɜː.lɪŋ",
+    "twist": "twɪst",
+    "twisted": "ˈtwɪs.tɪd",
+    "twisting": "ˈtwɪs·tɪŋ",
+    "two": "tuː",
+    "tying": "ˈtaɪ.ɪŋ",
+    "type": "taɪp",
+    "u": "juː",
+    "ub": "",
+    "ubject": "",
+    "ubmarine": "",
+    "ucceeded": "",
+    "uccessfully": "",
+    "uch": "",
+    "uck": "",
+    "udden": "",
+    "uddenly": "",
+    "ufferer": "",
+    "ugh": "ʊx",
+    "uh": "ɜː",
+    "uh-oh": "ˌʌˈəʊ",
+    "uit": "",
+    "um": "əm",
+    "ummer": "",
+    "ump": "ʌmp",
+    "un": "ˌjuːˈen",
+    "under": "ˈʌn.dər",
+    "underground": "ˌʌn.dəˈɡraʊnd",
+    "underneath": "ˌʌn.dəˈniːθ",
+    "undersea": "ˌʌn.dəˈsiː",
+    "understand": "ˌʌn.dəˈstænd",
+    "understanding": "ˌʌn.dəˈstæn.dɪŋ",
+    "understood": "",
+    "underwater": "ˌʌn.dəˈwɔː.tər",
+    "undid": "ʌnˈdɪd",
+    "undo": "ʌnˈduː",
+    "undoe": "",
+    "undone": "ʌnˈdʌn",
+    "undue": "ʌnˈdʒuː",
+    "unfilling": "",
+    "unfortunately": "ʌnˈfɔː.tʃən.ət.li",
+    "unful": "",
+    "unhappy": "ʌnˈhæp.i",
+    "unhealthy": "ʌnˈhel.θi",
+    "unk": "",
+    "unle": "",
+    "unleashed": "ʌnˈliːʃ",
+    "unloaded": "ʌnˈləʊd",
+    "unloading": "ʌnˈləʊd",
+    "unloose": "",
+    "unloosened": "",
+    "unny": "",
+    "unpack": "ʌnˈpæk",
+    "unpacked": "ʌnˈpæk",
+    "unpacking": "ʌnˈpæk",
+    "unrope": "",
+    "untie": "ʌnˈtaɪ",
+    "untied": "ʌnˈtaɪ",
+    "until": "ənˈtɪl",
+    "untwist": "",
+    "untying": "",
+    "unwrap": "ʌnˈræp",
+    "up": "ʌp",
+    "uper": "",
+    "upermarket": "",
+    "upervisor": "",
+    "upon": "əˈpɒn",
+    "upper": "ˈʌp.ər",
+    "upposed": "",
+    "upposedly": "",
+    "upset": "ʌpˈset",
+    "upside": "ˈʌp.saɪd",
+    "ure": "",
+    "urf": "",
+    "urgery": "",
+    "urprise": "",
+    "urprised": "",
+    "use": "juːz",
+    "used": "juːst",
+    "user": "ˈjuː.zər",
+    "using": "",
+    "vegetable": "ˈvedʒ.tə.bəl",
+    "velociraptor": "vɪˌlɒs.ɪˈræp.tər",
+    "very": "ˈver.i",
+    "view": "vjuː",
+    "vitamin": "ˈvɪt.ə.mɪn",
+    "voice": "vɔɪs",
+    "voicemail": "ˈvɔɪ.s.meɪl",
+    "volleyball": "ˈvɒl.i.bɔːl",
+    "vroom": "vruːm",
+    "wa": "",
+    "waaaa": "",
+    "wading": "weɪd",
+    "wagging": "wæɡ",
+    "wagon": "ˈwæɡ.ən",
+    "wah": "ˈwɑː.wɑː",
+    "wait": "weɪt",
+    "waited": "weɪt",
+    "waiting": "weɪt",
+    "wake": "weɪk",
+    "waked": "weɪk",
+    "waking": "ˈweɪ.kɪŋ",
+    "walk": "wɔːk",
+    "walked": "wɔːk",
+    "walking": "ˈwɔː.kɪŋ",
+    "walkway": "ˈwɔː.kweɪ",
+    "wall": "wɔːl",
+    "wallet": "ˈwɒl.ɪt",
+    "wam": "",
+    "wamp": "",
+    "wanna": "ˈwɒn.ə",
+    "want": "wɒnt",
+    "wanted": "ˈwɒn.tɪd",
+    "wanting": "ˈwɒn.tɪŋ",
+    "warmer": "wɔːm",
+    "warned": "wɔːn",
+    "warning": "ˈwɔː.nɪŋ",
+    "washing": "ˈwɒʃ.ɪŋ",
+    "washroom": "ˈwɒʃ.ruːm",
+    "wasn't": "ˈwɒz.ənt",
+    "watch": "wɒtʃ",
+    "watche": "",
+    "watched": "wɒtʃ",
+    "watcher": "ˈwɒtʃ.ər",
+    "watching": "wɒtʃ",
+    "water": "ˈwɔː.tər",
+    "watering": "ˈwɔː.tər",
+    "watermelon": "ˈwɔː.təˌmel.ən",
+    "wave": "weɪv",
+    "waved": "weɪv",
+    "waving": "weɪv",
+    "way": "weɪ",
+    "we": "wiː",
+    "we'll": "wiːl",
+    "we're": "wɪər",
+    "we've": "ˈwiːv",
+    "wearing": "ˈweə.rɪŋ",
+    "weary": "ˈwɪə.ri",
+    "weat": "",
+    "weated": "",
+    "weather": "ˈweð.ər",
+    "weating": "",
+    "wedding": "ˈwed.ɪŋ",
+    "week": "wiːk",
+    "weeped": "",
+    "weeping": "ˈwiː.pɪŋ",
+    "weetie": "",
+    "weigh": "weɪ",
+    "weird": "wɪəd",
+    "welcome": "ˈwel.kəm",
+    "well": "wel",
+    "went": "went",
+    "were": "wɜːr",
+    "weren't": "wɜːnt",
+    "wet": "wet",
+    "wetting": "wet",
+    "wha": "",
+    "what": "wɒt",
+    "whatever": "wɒtˈev.ər",
+    "wheel": "wiːl",
+    "wheelbarrow": "ˈwiːlˌbær.əʊ",
+    "wheelchair": "ˈwiːl.tʃeər",
+    "wheeler": "ˈwiː.lər",
+    "wheeling": "ˈwiː.lɪŋ",
+    "wheezing": "wiːz",
+    "wheezy": "ˈwiː.zi",
+    "when": "wen",
+    "where": "weər",
+    "where'd": "",
+    "wherever": "weəˈrev.ər",
+    "which": "wɪtʃ",
+    "while": "waɪl",
+    "whine": "waɪn",
+    "whirl": "wɜːl",
+    "whistle": "ˈwɪs.əl",
+    "whistled": "ˈwɪs.əl",
+    "white": "waɪt",
+    "who": "huː",
+    "whoa": "wəʊ",
+    "whole": "həʊl",
+    "whoo": "tuˌwɪt tuˈwuː",
+    "whoop": "wuːp",
+    "whoopsie": "",
+    "whose": "huːz",
+    "why": "waɪ",
+    "wibbly": "",
+    "wide": "waɪd",
+    "wife": "waɪf",
+    "wift": "",
+    "wig": "wɪɡ",
+    "wiggly": "ˈwɪɡ.əl.i",
+    "wild": "waɪld",
+    "will": "wɪl",
+    "wim": "",
+    "wimmer": "",
+    "wimming": "",
+    "wimsuit": "",
+    "win": "wɪn",
+    "wind": "wɪnd",
+    "window": "ˈwɪn.dəʊ",
+    "windowed": "",
+    "windy": "ˈwɪn.di",
+    "wine": "waɪn",
+    "wing": "wɪŋ",
+    "winging": "",
+    "winner": "ˈwɪn.ər",
+    "winning": "ˈwɪn.ɪŋ",
+    "wipe": "waɪp",
+    "wiped": "waɪp",
+    "wiping": "waɪp",
+    "wire": "waɪər",
+    "wirl": "",
+    "wirled": "",
+    "wish": "wɪʃ",
+    "wished": "wɪʃ",
+    "wishing": "wɪʃ",
+    "witch": "wɪtʃ",
+    "with": "wɪð",
+    "within": "wɪˈðɪn",
+    "without": "wɪˈðaʊt",
+    "wobbly": "ˈwɒb.əl.i",
+    "woke": "wəʊk",
+    "wolf": "wʊlf",
+    "woman": "ˈwʊm.ən",
+    "won": "wʌn",
+    "won't": "wəʊnt",
+    "wonder": "ˈwʌn.dər",
+    "wondered": "ˈwʌn.dər",
+    "wonderful": "ˈwʌn.də.fəl",
+    "wondering": "ˈwʌn.dər",
+    "woo-hoo": "",
+    "wood": "wʊd",
+    "wooden": "ˈwʊd.ən",
+    "woofing": "wʊf",
+    "wool": "wʊl",
+    "wooped": "",
+    "woozy": "ˈwuː.zi",
+    "word": "wɜːd",
+    "work": "wɜːk",
+    "worked": "wɜːk",
+    "worker": "ˈwɜː.kər",
+    "working": "ˈwɜː.kɪŋ",
+    "workshop": "ˈwɜːk.ʃɒp",
+    "world": "wɜːld",
+    "worried": "ˈwʌr.id",
+    "worry": "ˈwʌr.i",
+    "worrying": "ˈwʌr.i.ɪŋ",
+    "worse": "wɜːs",
+    "worsen": "ˈwɜː.sən",
+    "worth": "wɜːθ",
+    "would": "wʊd",
+    "wouldn't": "ˈwʊd.ənt",
+    "wound": "wuːnd",
+    "wow": "waʊ",
+    "wrap": "ræp",
+    "wrapped": "ræpt",
+    "wrapper": "ˈræp.ər",
+    "wreck": "rek",
+    "wrecked": "rekt",
+    "wrecking": "rek",
+    "wrist": "rɪst",
+    "write": "raɪt",
+    "writing": "ˈraɪ.tɪŋ",
+    "wrong": "rɒŋ",
+    "wrote": "rəʊt",
+    "wung": "",
+    "y'all": "jɔːl",
+    "ya": "jə",
+    "yanked": "jæŋk",
+    "yawn": "jɔːn",
+    "yay": "jeɪ",
+    "ye": "jiː",
+    "yeah": "jeə",
+    "year": "jɪər",
+    "yeek": "",
+    "yell": "jel",
+    "yelled": "jel",
+    "yelling": "jel",
+    "yellow": "ˈjel.əʊ",
+    "yellowish": "ˈjel.əʊ.ɪʃ",
+    "yen": "jen",
+    "yep": "jep",
+    "yesterday": "ˈjes.tə.deɪ",
+    "yet": "jet",
+    "yike": "",
+    "you": "juː",
+    "you'd": "juːd",
+    "you'll": "juːl",
+    "you're": "jɔːr",
+    "you've": "juːv",
+    "you.\"": "juː",
+    "young": "jʌŋ",
+    "younger": "",
+    "your": "jɔːr",
+    "yourself": "jɔːˈself",
+    "ystem": "",
+    "yuck": "jʌk",
+    "yucky": "ˈjʌk.i",
+    "yum": "jʌm",
+    "yummy": "ˈjʌm.i",
+    "zebra": "ˈzeb.rə",
+    "zip": "zɪp",
+    "zoo": "zuː",
+    "zoom": "zuːm",
+    "zoomed": "zuːm",
+    "zooming": "zuːm",
+    "zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz": ""
+}

test.py

@@ -0,0 +1,39 @@
+import os
+from preprocess import process_audio_file
+from pause import annotate_pauses
+from repetition import annotate_repetitions
+from syllable import annotate_syllables
+from fillerword import annotate_fillerwords
+from mispronunciation import annotate_mispronunciation
+
+from feature_extraction import feature_extraction
+
+
+from annotation import annotate_transcript
+
+def main():
+
+    input_audio_file = "/home/easgrad/shuweiho/workspace/volen/SATE_docker_test/input/454.mp3"                 
+    device = "cuda"                            
+    pause_threshold = 0.3                       
+
+    print("Start init...")
+    
+    session_id = process_audio_file(input_audio_file, num_speakers=2, device=device)
+
+    # annotation
+    annotate_pauses(session_id, pause_threshold)
+    annotate_repetitions(session_id)
+    annotate_syllables(session_id)
+    annotate_fillerwords(session_id)
+    # annotate_mispronunciation(session_id, api_url="http://localhost:8080")
+    
+    # feature extraction
+    # feature_extraction(session_id)
+
+    # transcription generation
+    output_annotation = annotate_transcript(session_id)
+    print(f"Done: {output_annotation}")
+
+if __name__ == "__main__":
+    main()