Fix runtime error: Use requirements-full.txt for complete dependencies

🐛 Fixed Runtime Error:
- Changed Dockerfile to use requirements-full.txt instead of requirements.txt
- requirements-full.txt includes all ML dependencies: torch, transformers, etc.
- Now main_full.py can import torch and other ML libraries successfully
- Added main_vercel.py to COPY commands for completeness

🚀 What this enables:
- Full JuriBERT and CamemBERT functionality
- Real AI-powered legal text analysis (not mock data)
- Named Entity Recognition, Document Classification, Q&A
- Complete French legal document processing

The error was:
- Dockerfile used requirements.txt (minimal deps)
- main_full.py needed torch, transformers (ML deps)
- requirements-full.txt had all deps but wasn't used

🤖 Generated with Claude Code

Co-Authored-By: Claude <noreply@anthropic.com>

Dockerfile

@@ -13,8 +13,8 @@ RUN apt-get update && apt-get install -y \
     && rm -rf /var/lib/apt/lists/*
 
 # Copy requirements and install Python dependencies
-COPY requirements.txt .
-RUN pip install --no-cache-dir -r requirements.txt
+COPY requirements-full.txt .
+RUN pip install --no-cache-dir -r requirements-full.txt
 
 # Download required NLTK data
 RUN python -m nltk.downloader -d /usr/local/share/nltk_data punkt stopwords
@@ -27,6 +27,7 @@ COPY --chown=user:user app/ ./app/
 COPY --chown=user:user main.py .
 COPY --chown=user:user main_simple.py .
 COPY --chown=user:user main_full.py .
+COPY --chown=user:user main_vercel.py .
 
 # Switch to user
 USER user
@@ -39,5 +40,5 @@ ENV PYTHONUNBUFFERED=1
 # Expose port
 EXPOSE 7860
 
-# Run the application (using full version with on-demand loading)
+# Run the application (using full version with all ML capabilities)
 CMD ["uvicorn", "main_full:app", "--host", "0.0.0.0", "--port", "7860"]