ddgs
requests
python-dotenv
trafilatura

# fetch fallbacks
readability-lxml
newspaper4k

# table extraction
pandas
lxml
tabulate

# ranking
rank-bm25

# optional: semantic reranking (bi-encoder + cross-encoder)
# sentence-transformers

# extract - install only the providers you use
# anthropic              # ClaudeExtractor
# openai                 # GPTExtractor
google-generativeai      # GeminiExtractor (free tier)
groq                     # GroqExtractor (free tier, fastest)

# PDF datasheets
pdfplumber

# optional: JS-rendered pages (pip install playwright separately, then: playwright install chromium)
# playwright
