# Task Management - [/] Research Data Sources - [ ] Analyze `research_report.json` structure - [x] Analyze `baymard_unified_database_v2.json` structure - [ ] Draft Product Requirements Document (PRD) - [x] Define User Personas and Goals - [x] detailed Feature Specifications (Search, RAG, Image Analysis) - [x] Define Technical Architecture (High-level) - [x] Define Metrics and Success Criteria - [x] Research File Linking - [x] Explore `research_pages` directory naming convention - [x] Create script to match JSON entries with file paths - [x] Update `research_report.json` with file paths - [x] Update File Structure - [x] Update script for new paths and relative linking - [x] Run script to fix `research_report.json` - [x] Verify Match Quality - [x] Create script to check similarity between JSON title and file path - [x] Report ambiguous matches to user - [x] Handle Unmatched Items - [x] Mark missing file paths as "파일 없음" - [/] Review PRD with User - [x] Review PRD with User - [x] Update PRD with finalized data structure (file_path) - [x] Request final PRD approval ## Chatbot PoC Development (Metadata Search) - [x] Setup Development Environment - [x] Create virtual environment and install dependencies (langchain, google-genai) - [ ] Verify Gemini API Key (`gemini-3-flash-preview`) - [x] Data Loader Implementation - [x] Implement `data_loader.py` for both JSON sources - [x] Search Logic Implementation - [x] Implement keyword extraction & source routing with LLM - [x] Implement filtering logic for JSON data - [x] Implement answer generation with Gemini 3 Flash - [x] Interface Implementation - [x] Build CLI for testing - [x] Verify specific used cases (Review, Internal Only, etc.) ## Web App Development (Streamlit) - [x] Setup Web Environment - [x] Update `requirements.txt` with `streamlit` - [x] Install new dependencies - [x] Implement Web Interface - [x] Create `app.py` using Streamlit chat elements - [x] Integrate `SearchEngine` with Streamlit session state - [x] Verification - [x] Run Streamlit app and verify chat flow ## Feature Enhancements - [x] Add Conversation Context - [x] Update `search_engine.py` to handle chat history - [x] Update `app.py` to pass session history - [x] Verify context-aware follow-up questions ## RAG Integration Phase - [x] Data Preparation: Assign Unique IDs - [x] Create script to assign IDs to JSON and sync with MD files - [x] Run ID assignment script - [x] Vector DB Setup - [x] Install search dependencies - [x] Implement search indexing for separate collections (Internal/Baymard) - [x] Hybrid Search Implementation - [x] Update `search_engine.py` to combine Metadata + Vector Search - [x] Verify RAG results - [x] Enhance Baymard Indexing (PDF Extraction) ## Multimodal Support - [x] Implement File Upload in Web App - [x] Add `st.file_uploader` to `app.py` - [x] Update `search_engine.py` to handle image input - [x] Verification - [x] Test image analysis flow - [x] Deployment - [x] Prepare for Hugging Face Space - [x] Update `.env` with HF Token - [x] Copy Data to Project Directory (Containerization) - [x] Update Path Logic in Code (Relative Paths) - [x] Create `Dockerfile` - [x] Create `deploy_to_hf.py` script - [x] Review PRD with User ## Slackbot Integration - [x] Slack App Setup - [x] Create Setup Guide - [x] Configure Scopes & Tokens - [x] Implement Slackbot Logic - [x] Add `slack_bolt` dependency - [x] Create `slack_bot.py` - [x] Handle File Uploads (Images/Videos) - [x] Handle DMs and Thread Replies - [x] Implement Thread Context Awareness - [x] Update `slack_bot.py` to fetch thread history - [x] Convert Slack history to SearchEngine format - [x] Deployment Updates - [x] Update Dockerfile for dual execution (or separate) - [x] Separate Web App and Slackbot Deployments - [x] Migrate Data to HF Dataset (Fix Storage Limit) - [x] Upload `data` to HF Dataset - [x] Update `Dockerfile` to download data on startup - [x] Create `Dockerfile.web` and `deploy_web.py` - [x] Create `Dockerfile.slack` and `deploy_slack.py` - [x] Deploy to separate Spaces - [x] Fix Slack Formatting Issues (Markdown to mrkdwn) - [x] Update search_engine for strict formatting instructions - [x] Implement deterministic Regex converter - [x] Redeploy Slackbot (Clean Install) ## Search & Context Improvements - [x] Improve Search Quality - [x] Add `thefuzz` for fuzzy matching logic - [x] Refactor `filter_internal_data` to use fuzzy scoring - [x] Enhance Context Awareness - [x] Update `keyword_extraction_prompt` to inherit context from history - [x] Verify multi-turn conversation flow