Hugging Face's logo

Spaces:
Aniket2006
/
Chatbot
Sleeping

App Files Community
Chatbot

Commit History

  1. Security: Load API keys from environment variable

    Aniket2006
    026c0c9

  1. Fix: Pass user query to LLM context in Fast Lane and Deep Dive stages

    Aniket2006
    7f34cf9

  2. Add language matching: respond in same language as user query (Hindi/English/Hinglish)

    Aniket2006
    9c22892

  3. Fix: Handle list type in context_priority_used extraction

    Aniket2006
    882becc

  4. Upgrade to Hybrid Architecture v3 (Fast Lane + Deep Dive)

    Aniket2006
    e2d9873

  5. Add 21 Groq API keys with fallback mechanism

    Aniket2006
    f895e11

  6. Update Groq API keys with rotation logic

    Aniket2006
    3d6dd09

  1. Update Groq API key

    Aniket2006
    1190215

  2. Update Groq API key

    Aniket2006
    499a614

  3. feat: migrate from Gemini to Groq API for LLM calls

    Aniket2006
    f0c4cb2

  1. Increase max_output_tokens to 4096 to prevent response truncation

    Aniket2006
    dcb1582

  2. Complete rewrite: direct API fetch, proper vegetation indices + cluster stress extraction

    Aniket2006
    d6e5751

  3. Fix: format specifier error in SAR section

    Aniket2006
    65058ca

  4. Fix: extract Sentinel-2 vegetation_indices_summary and SAR crop_health

    Aniket2006
    15baaaf

  5. Improve prompt: force data-specific responses, add context logging

    Aniket2006
    f11af9a

  6. Fix: handle None values in SAR bands section

    Aniket2006
    c187541

  7. Fix: add defensive null checks in generate_response

    Aniket2006
    258275e

  8. Add comprehensive context: all indices, SAR, weather, persona-based responses

    Aniket2006
    c7df30e

  9. Switch to gemini-2.5-flash (2.0 quota exhausted)

    Aniket2006
    bec40f7

  10. Fix: use gemini-2.0-flash model

    Aniket2006
    ee7723b

  11. Add /chat/stream endpoint for Flutter compatibility

    Aniket2006
    b965fde

  12. Add google-generativeai package

    Aniket2006
    89ca989

  13. Restore original simple chatbot from initial commit

    Aniket2006
    cb547c1

  14. Simplify chatbot: single LLM call, remove complex reasoning

    Aniket2006
    aaf7391

  15. Fix: record_api_call once per query, not per key rotation

    Aniket2006
    83e9fb0

  16. Pass skip_llm=True to Sentinel2, enforce 5 calls/minute limit

    Aniket2006
    56a0fa6

  17. Add strict 5 calls/minute rate limit with smart fallback

    Aniket2006
    87f35a1

  18. Reduce API usage: default to simple reasoning (1 call), add backoff

    Aniket2006
    dd7cfa9

  19. Add SSE streaming endpoint /chat/stream for typewriter effect

    Aniket2006
    5d82df6

  20. Remove conversation history from LLM context to reduce tokens

    Aniket2006
    7b3dbf5

  21. Enhanced compact context: ALL 13 indices, clusters, prev_analysis, conv_history

    Aniket2006
    a8d7d80

  22. Add compact prompts and context builder - reduces token usage by ~50%

    Aniket2006
    cb3e125

  23. Fix: add 400 to rotation triggers for invalid API keys

    Aniket2006
    0f136d3

  24. Fix: pass context to _generate_response for persona/zone/trend/weather

    Aniket2006
    a6e4432

  25. Add 5-key Gemini API fallback, conversation memory, historical trends, and zone recommendations

    Aniket2006
    504a512

  1. Lower confidence thresholds for satellite data fetch and reasoning

    Aniket2006
    a7903f2

  2. Fix maybeSingle not supported - use execute and check results

    Aniket2006
    9d8c05a

  3. Fix user_profiles query - use address column instead of location

    Aniket2006
    b014757

  4. Full Developer Spec implementation: Priority-based Multi-Stage Reasoning⁠

    Aniket2006
    ee8289e

  5. Add satellite band data injection from HF Space APIs (SAR, Sentinel-2)

    Aniket2006
    2221290

  6. Auto-fetch field data (coordinates, crop_type) from Supabase

    Aniket2006
    3fad61c

  7. Fix: Copy all Python files in Dockerfile

    Aniket2006
    f3a54eb

  8. Implement multi-stage reasoning architecture (Claim→Validate→Contradict→Confirm)

    Aniket2006
    e2fa5a7

  9. Add auto model discovery to find working Gemini model

    Aniket2006
    07a23c0

  10. Try gemini-1.5-flash model

    Aniket2006
    c7be993

  11. Switch to gemini-1.5-flash-8b for higher rate limits

    Aniket2006
    c36a5b7

  12. Add retry logic for rate limits

    Aniket2006
    99bff2e

  13. Try gemini-2.0-flash-exp model

    Aniket2006
    126daae

  14. Fix: Use v1beta API and add better error logging

    Aniket2006
    fcbe917

  15. Fix: Use gemini-pro with v1 API endpoint

    Aniket2006
    0a05e0c