requirements.txt

# ── Core Framework ────────────────────────────────────────────────────────────
fastapi==0.115.6
uvicorn[standard]==0.32.1
python-multipart==0.0.17          # File upload support
pydantic==2.9.2
pydantic-settings==2.6.1

# ── NLP & ML ──────────────────────────────────────────────────────────────────
transformers==4.46.3
torch==2.5.1
sentence-transformers==3.3.1
scikit-learn==1.5.2
safetensors>=0.4.3                # Faster, safer model serialisation (used by transformers)
spacy==3.8.2
langdetect==1.0.9
nltk==3.9.1

# ── Input Modules ─────────────────────────────────────────────────────────────
pytesseract==0.3.13               # OCR
Pillow==11.0.0                    # Image processing
# openai-whisper==20240930        # ASR (Filipino speech) — installed separately in Dockerfile (--no-build-isolation)
beautifulsoup4==4.12.3            # URL scraping
requests==2.32.3
lxml==5.3.0

# ── Evidence Retrieval ────────────────────────────────────────────────────────
newsapi-python==0.2.7

# ── Database ──────────────────────────────────────────────────────────────────
sqlalchemy==2.0.36
asyncpg==0.30.0                   # Async PostgreSQL driver
alembic==1.14.0

# ── Caching ───────────────────────────────────────────────────────────────────
redis==5.2.1
cachetools==5.5.0

# ── Dataset Pipeline ─────────────────────────────────────────────────────────
datasets>=2.21.0              # HuggingFace dataset loader (fake_news_filipino, liar, etc.)
pandas>=2.2.0                  # Parquet I/O for combined dataset
pyarrow>=17.0.0                # Parquet engine
datasketch>=1.6.4              # MinHash LSH for fast deduplication
kagglehub>=0.3.0               # Auto-download Kaggle datasets (ISOT)

# ── Utilities ─────────────────────────────────────────────────────────────────
python-dotenv==1.0.1
httpx==0.28.1                     # Async HTTP client
aiofiles==24.1.0
tqdm==4.67.1
numpy==1.26.4
facebook-scraper>=0.2.59          # Fallback scraper for Facebook post URLs (public posts, no login)

# ── Testing ───────────────────────────────────────────────────────────────────
pytest==8.3.4
pytest-asyncio==0.24.0
httpx==0.28.1                     # FastAPI TestClient