Delete public-parquet/.ipynb_checkpoints/personal shopper demo-checkpoint.ipynb

public-parquet/.ipynb_checkpoints/personal shopper demo-checkpoint.ipynb

@@ -1,122 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "e255d061-b87e-41d1-bdec-75717c311031",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import glob, heapq, logging, os, time\n",
-    "import polars as pl\n",
-    "import pandas as pd\n",
-    "import numpy as np\n",
-    "from sentence_transformers import SentenceTransformer\n",
-    "from IPython.display import display\n",
-    "\n",
-    "# ─── logging config ───────────────────────────────────────────────────────────\n",
-    "logging.basicConfig(\n",
-    "    level=logging.INFO,\n",
-    "    format=\"[%(asctime)s] %(message)s\",\n",
-    "    datefmt=\"%H:%M:%S\"\n",
-    ")\n",
-    "\n",
-    "# ─── helper: cosine similarity ────────────────────────────────────────────────\n",
-    "def cosine_batch(q_vec: np.ndarray, mat: np.ndarray) -> np.ndarray:\n",
-    "    q_norm   = q_vec / np.linalg.norm(q_vec)\n",
-    "    mat_norm = mat / np.linalg.norm(mat, axis=1, keepdims=True)\n",
-    "    return mat_norm @ q_norm\n",
-    "\n",
-    "# ─── model & shards ───────────────────────────────────────────────────────────\n",
-    "MODEL_NAME = \"sentence-transformers/all-MiniLM-L6-v2\"\n",
-    "encoder    = SentenceTransformer(MODEL_NAME)\n",
-    "SHARDS     = sorted(glob.glob(\"tsmpd_public_*.parquet\"))\n",
-    "\n",
-    "# ─── streaming search ─────────────────────────────────────────────────────────\n",
-    "def search_streaming(query: str, top_k: int = 10) -> pd.DataFrame:\n",
-    "    q_vec = encoder.encode(query, convert_to_numpy=True).astype(\"float32\")\n",
-    "\n",
-    "    best = []               # heap of (score, uid, vendor, title, paragraph)\n",
-    "    heapq.heapify(best)\n",
-    "\n",
-    "    for idx, shard in enumerate(SHARDS, 1):\n",
-    "        t0 = time.perf_counter()\n",
-    "        logging.info(f\"Shard {idx}/{len(SHARDS)} → loading {os.path.basename(shard)}\")\n",
-    "\n",
-    "        tbl = pl.read_parquet(\n",
-    "            shard,\n",
-    "            columns=[\"uid\", \"vendor\", \"title\", \"paragraph\", \"embedding\"]\n",
-    "        )\n",
-    "        logging.info(f\"    rows = {tbl.height:,}  | load = {time.perf_counter() - t0:.1f}s\")\n",
-    "\n",
-    "        embs = np.vstack(tbl[\"embedding\"].to_list()).astype(\"float32\")\n",
-    "        sims = cosine_batch(q_vec, embs)\n",
-    "\n",
-    "        for i, score in enumerate(sims):\n",
-    "            item = (\n",
-    "                float(score),\n",
-    "                tbl[\"uid\"][i],\n",
-    "                tbl[\"vendor\"][i],\n",
-    "                tbl[\"title\"][i],\n",
-    "                tbl[\"paragraph\"][i],\n",
-    "            )\n",
-    "            if len(best) < top_k:\n",
-    "                heapq.heappush(best, item)\n",
-    "            elif score > best[0][0]:\n",
-    "                heapq.heapreplace(best, item)\n",
-    "\n",
-    "        logging.info(\n",
-    "            f\"    processed in {time.perf_counter() - t0:.1f}s   \"\n",
-    "            f\"best_min={best[0][0]:.4f}\"\n",
-    "        )\n",
-    "        del tbl, embs, sims  # free mem per shard\n",
-    "\n",
-    "    if not best:\n",
-    "        return pd.DataFrame()\n",
-    "\n",
-    "    best_sorted = sorted(best, key=lambda x: x[0], reverse=True)\n",
-    "    scores, uids, vendors, titles, paras = zip(*best_sorted)\n",
-    "\n",
-    "    return pd.DataFrame({\n",
-    "        \"score\": scores,\n",
-    "        \"uid\": uids,\n",
-    "        \"vendor\": vendors,\n",
-    "        \"title\": titles,\n",
-    "        \"paragraph\": paras\n",
-    "    })\n",
-    "\n",
-    "# ─── run a test query ─────────────────────────────────────────────────────────\n",
-    "query   = \"vintage brass floor lamp with Edison bulbs\"\n",
-    "start   = time.perf_counter()\n",
-    "hits    = search_streaming(query, top_k=5)\n",
-    "elapsed = time.perf_counter() - start\n",
-    "logging.info(f\"Finished query in {elapsed:.1f}s\")\n",
-    "\n",
-    "# display full text in each column\n",
-    "pd.set_option(\"display.max_colwidth\", None)\n",
-    "display(hits)\n"
-   ]
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.13.2"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 5
-}