AI-Based Doubt Resolution Bot

An academic doubt-resolution assistant powered by Google Gemini (gemini-1.5-flash). The working implementation is a command-line chatbot (main.py) that keeps a local JSON conversation history and answers questions with a Gemini-API-key-help system prompt. Two HTML fragments (login.html, doubt.html) provide a student login page and a doubt-submission form, but no backend implementing their /login / /submit_doubt endpoints exists in this repo (see Known gaps).

Note on repo description: it mentions FastAPI, FAISS, and SentenceTransformers. None of those appear in the committed code — main.py uses only google-generativeai plus the standard library. They are treated below as roadmap, not implemented features.

Features

• CLI Q&A loop over Gemini 1.5 Flash with a focused system prompt (Gemini API-key setup, usage, troubleshooting, best practices)
• Last-5-exchanges context window injected into each prompt
• Persistent history in conversation_history.json (timestamped user/bot pairs), loaded at startup and saved after every turn
• Structured logging (logging module) for configuration, history, and generation errors
• Graceful degradation: friendly fallback message on generation failure; clean exit handling
• Web UI fragments: styled student login form and doubt-submission form wired to POST /login and POST /submit_doubt (backend not included)

Stack

Layer	Technology
LLM	Google Gemini 1.5 Flash via google-generativeai
Language	Python 3.x (os, json, logging, datetime)
Frontend fragments	Plain HTML + vanilla JS fetch (doubt.html, login.html)
Persistence	Local conversation_history.json (created at runtime, git-ignored)
Ignored at runtime	__pycache__/, *.pyc, conversation_history.json patterns belong in .gitignore

Repository structure

AI-Based-Doubt-Resolution-Bot-/
├── main.py          # CLI bot: Gemini config, history load/save, prompt builder, REPL loop
├── doubt.html       # HTML fragment: greeting + doubt form → POST /submit_doubt (expects {response})
├── login.html       # Standalone page: student login form → POST /login; RIT branding header
├── .gitignore       # Python / packaging / test ignores
├── LICENSE          # MIT (© 2025 SHYAMFRANCIS)
└── conversation_history.json  # Created at runtime (not committed)

Installation

Prerequisites: Python 3.8+, a Gemini API key (Google AI Studio).

git clone https://github.com/SHYAMFRANCIS/AI-Based-Doubt-Resolution-Bot-.git
cd AI-Based-Doubt-Resolution-Bot-
pip install google-generativeai
export GEMINI_API_KEY="your-key-here"   # PowerShell: $env:GEMINI_API_KEY="your-key-here"

⚠️ Security: the committed main.py currently hardcodes an API key as a module-level api_key string. Rotate that key, delete it from the source, and read it from the environment instead:

GEMINI_API_KEY = os.environ.get("GEMINI_API_KEY")

Usage

python main.py

Welcome to the Gemini API Key Doubt Resolution Bot!
Ask your questions about Gemini API key setup, usage, or troubleshooting.
Type 'exit' to quit.

You: How do I create a Gemini API key?
Bot: ...
You: exit
Goodbye!

• Empty input is rejected with a prompt to enter a valid question.
• Each exchange is appended to conversation_history.json with an ISO timestamp.

Examples

You ask	Bot context
"Where do I find my Gemini API key?"	System prompt + last 5 turns → concise, beginner-friendly setup steps
"Why do I get a 400/403 from the API?"	Troubleshooting guidance grounded in conversation history
exit	Saves history and quits

Web fragments (frontend only):

<!-- doubt.html posts urlencoded doubt=... and renders JSON field `response` -->
fetch('/submit_doubt', { method: 'POST',
  headers: {'Content-Type': 'application/x-www-form-urlencoded'},
  body: new URLSearchParams({ doubt }) })

<!-- login.html -->
<form action="/login" method="POST"> ... username / email / password ... </form>

Configuration

Setting	Location	Default
Model	initialize_model() in main.py	gemini-1.5-flash
History file	HISTORY_FILE	conversation_history.json
Context window	generate_response()	last 5 exchanges
System prompt	generate_response()	Gemini-API-key specialist, beginner-friendly
Log level	logging.basicConfig	INFO

Known gaps

• doubt.html / login.html reference POST /submit_doubt and POST /login, but no Flask/FastAPI server implements them — main.py is CLI-only (input() loop).
• No requirements.txt; install is a single pip install google-generativeai.
• No FAISS index, SentenceTransformers embeddings, voice input, or multilingual support in code despite the repo description — candidates for future work.

Contributing

1. Fork and create a feature branch.
2. Move the API key to GEMINI_API_KEY env (do not commit keys).
3. Add the missing web backend or requirements.txt if in scope.
4. Open a pull request describing the change and how it was tested (python main.py smoke test).

License

MIT — see LICENSE. © 2025 SHYAMFRANCIS.