Update from backup state - 2026-05-06

ANTIGRAVITY_INSTRUCTIONS.md

@@ -0,0 +1,29 @@
+# Antigravity Setup & Instructions
+
+This backup was created on **2026-05-05**.
+It contains the full state of the Martechsol RAG Chatbot project.
+
+## Current Setup Info
+- **Main Workspace Path:** `D:\RAG_Backup_2026_05_05`
+- **Primary LLM:** Qwen 3 32B (via Groq)
+- **Deployment Target:** Hugging Face Spaces (`space` remote)
+- **Git Remote:** `https://huggingface.co/spaces/joedown11/chatRAG`
+
+## Instructions for Antigravity (Coding Assistant)
+
+If the user says **"push"**, you should perform the following actions to update the project and its deployment:
+
+1.  **Sync Local Changes**: Ensure all local files are saved.
+2.  **Commit Changes**:
+    - Run `git add .`
+    - Run `git commit -m "Update from backup state - [Current Date]"` (or a specific message if provided).
+3.  **Deploy to Space**:
+    - Run `git push space main` (or `master` if appropriate).
+4.  **Verify Configuration**:
+    - Ensure `.env` or `app/core/config.py` has the correct API keys if they were changed.
+    - If the user changed the API provider (e.g., from Groq to something else), verify that `app/core/config.py` reflects the new provider and model.
+
+## Note for next session:
+The user intends to change the API provider and other settings *after* this backup.
+If you are opening this folder, check `app/core/config.py` first to see the current provider.
+If the user says "push", they want the *latest* version of the code and setup to be updated everywhere as it is right now.

CONFIG_NOTES.md

@@ -1,29 +1,32 @@
 # Configuration Backup - Martechsol RAG Chatbot
-**Date:** 2026-04-28
+**Date:** 2026-05-05
 **Source:** C:\Users\DELL\Desktop\RAG_backup2
-**Destination:** D:\RAG_Backup_2026_04_28
+**Destination:** D:\RAG_Backup_2026_05_05
 
-## Core Settings (from app/core/config.py)
+## Core Settings (latest state)
 - **App Name:** Fast RAG Chatbot
-- **LLM Provider:** groq (default)
-- **Groq Model:** llama-3.1-8b-instant
+- **LLM Provider:** groq
+- **Main LLM Model:** qwen/qwen3-32b (Groq)
+- **Query Rewriter:** llama-3.1-8b-instant (Groq)
 - **HF Model:** meta-llama/Llama-3.1-8B-Instruct
 - **Embedding Model:** BAAI/bge-small-en-v1.5
+- **Reranker Model:** cross-encoder/ms-marco-MiniLM-L-6-v2
 - **Docs Directory:** docs/
 - **Index Directory:** data/index/
 - **Sessions Directory:** data/sessions/
-- **Chunk Size:** 420 tokens
+- **Chunk Size:** 350 tokens (Optimized for context window)
 - **Overlap:** 80 tokens
-- **Top K:** 4
+- **Top K:** 8
+- **Max Context Chunks:** 4
 
 ## Admin Credentials
 - **Username:** martech_admin
 - **Password:** martech_admin_303
 - **OTP Expiry:** 300 seconds (5 minutes)
 
-## Security
+## Security & Notifications
 - **Auth Method:** HTTP Basic Auth + OTP (email-based)
-- **Email for OTP:** randomjoedown@gmail.com
+- **Admin Email:** randomjoedown@gmail.com
 - **SMTP Server:** smtp.gmail.com (Port 465)
 
 ## API / Integration
@@ -32,7 +35,7 @@
 - **Admin Endpoint:** /admin
 - **CORS:** Allowed for all (*) for testing
 
-## Infrastructure
-- **Hugging Face Path:** /data (for persistent storage)
-- **Local Path:** data/ (for local storage)
-- **Dockerfile:** Based on python:3.10-slim, serves on port 7860
+## Optimization Notes
+- **Latency Optimization**: Skip thinking process for Qwen3 using deterministic intent detection.
+- **Conciseness Protocol**: Strict 25-30 word response limit enforced via System Prompt.
+- **Hybrid Search**: FAISS + BM25 with Reranker for high precision.

PROJECT_TECHNICAL_SPEC.md

@@ -0,0 +1,89 @@
+# Martechsol HR Assistant: Project Technical Specification & Workflow
+
+This document provides a comprehensive overview of the **Martechsol HR Assistant**, covering its architecture, the models used, API integrations, and the end-to-end data workflow.
+
+---
+
+## 1. Project Architecture Overview
+
+The system is a high-performance **Retrieval-Augmented Generation (RAG)** application designed to provide precise, grounded answers based on employee handbooks and HR documents.
+
+### High-Level Stack:
+- **Language**: Python 3.10+
+- **API Framework**: FastAPI
+- **Server**: Uvicorn
+- **UI**: Gradio (for direct testing/Hugging Face Spaces)
+- **Web Integration**: Custom HTML/CSS/JS "Floating Icon" for WordPress/External sites.
+- **Persistence**: FAISS (Vector DB) & Local Session Storage.
+
+---
+
+## 2. Models & API Infrastructure
+
+The project utilizes 4 distinct models to ensure a balance between speed, reasoning capability, and accuracy.
+
+| Model Component | Model Name / ID | Provider / Engine | Purpose |
+| :--- | :--- | :--- | :--- |
+| **Main LLM (Brain)** | `qwen/qwen3-32b` | **Groq** | Reasoning, formatting, and generating the final response. |
+| **Query Rewriter** | `llama-3.1-8b-instant` | **Groq** | Expanding user queries into multiple search terms (used as fallback). |
+| **Embedding Model** | `BAAI/bge-small-en-v1.5` | Sentence Transformers | Converting text chunks into mathematical vectors for search. |
+| **Reranker** | `ms-marco-MiniLM-L-6-v2` | Cross-Encoder | Deeply evaluating the relevance of retrieved document chunks. |
+
+### APIs Connected:
+1.  **Groq API**: Powers the LLM inference (Qwen3 and Llama 3.1). Groq is chosen for its extreme speed (LPUs).
+2.  **SMTP (Gmail/Custom)**: Used for sending OTPs or notifications (if configured).
+3.  **Hugging Face**: Used for hosting the Gradio interface and downloading model weights for Embeddings/Reranking.
+
+---
+
+## 3. The RAG Pipeline Workflow
+
+When a user sends a message, the system follows this exact sequence:
+
+### Phase 1: Query Processing
+1.  **Normalization**: The message is cleaned and converted to a cache key (MD5 hash).
+2.  **Cache Lookup**: If the exact question was asked recently, the answer is returned instantly from an in-memory LRU cache.
+3.  **Local Expansion**: The query is expanded into up to 3 targeted search queries using a deterministic keyword map (e.g., "timing" → "office hours", "workday schedule"). This saves 3-5 seconds of LLM latency.
+
+### Phase 2: Retrieval (Search)
+1.  **Vector Search**: The system uses **FAISS** to find the most mathematically similar text chunks in the `docs/` folder.
+2.  **Keyword Search (BM25)**: Complements vector search by finding exact word matches.
+3.  **Hybrid Filtering**: Chunks are initially filtered by a relevance threshold (RRF).
+
+### Phase 3: Intelligence & Reranking
+1.  **Cross-Encoding**: The top retrieved chunks are sent to the **Reranker model**. Unlike standard search, the reranker looks at the *actual meaning* of the chunk relative to the question.
+2.  **Context Construction**: The most relevant chunks (capped at 1500 words to avoid token limits) are assembled into a context block.
+
+### Phase 4: Generation (The Final Answer)
+1.  **Prompt Injection**: The question, context, and a pruned chat history (last 2 turns) are wrapped in a strict **Master System Prompt**.
+2.  **LLM Execution**: The request is sent to `qwen/qwen3-32b`.
+    - *Optimization*: The instruction `/no_think` is added to skip internal chain-of-thought, reducing response time by 40-60%.
+3.  **Post-Processing**:
+    - Strips `<think>` tags.
+    - Removes conversational filler (e.g., "Based on the documents provided...").
+    - Enforces HTML formatting (bolding key terms, `<br>` for lists).
+
+---
+
+## 4. Backend Components & Directory Structure
+
+### Core Files:
+- `app/main.py`: Entry point for the FastAPI application.
+- `app/core/config.py`: Centralized configuration (API keys, model names, chunk sizes).
+- `app/services/rag_pipeline.py`: Orchestrates the flow from query to final answer.
+- `app/services/llm.py`: Handles communication with Groq and prompt engineering.
+- `app/services/vector_store.py`: Manages FAISS indexing and retrieval.
+- `app/services/session_store.py`: Manages user history and OTP sessions.
+
+### Data Flow:
+1.  **Input**: User Query (HTTP POST to `/chat` or Gradio WebSocket).
+2.  **Processing**: FastAPI → RAG Pipeline Service → LLM Service.
+3.  **Output**: JSON Response with `reply` and `retrieved_chunks` (metadata).
+
+---
+
+## 5. Summary of Key Features
+- **Deterministic Intent Detection**: Instantly recognizes topics like "leaves" or "salary" to pull the right data without LLM delay.
+- **Thinking Model Suppression**: Uses the power of Qwen3/DeepSeek while disabling their slow "thinking" phase for the final user response.
+- **Hybrid Search**: Combines the precision of keywords with the "vibe" search of vectors.
+- **Strict Formatting**: Enforces list styles and word counts to ensure a professional HR tone.

robocopy_backup.log

@@ -1,81 +0,0 @@
-
--------------------------------------------------------------------------------
-   ROBOCOPY     ::     Robust File Copy for Windows                              
--------------------------------------------------------------------------------
-
-  Started : Wednesday, April 29, 2026 9:00:28 PM
-   Source : c:\Users\DELL\Desktop\RAG_backup2\
-     Dest : C:\Users\DELL\Desktop\RAG_backup_final\
-
-    Files : *.*
-	    
-Exc Files : .DS_Store
-	    
- Exc Dirs : .git
-	    .gemini
-	    __pycache__
-	    venv
-	    
-  Options : *.* /TEE /S /E /DCOPY:DA /COPY:DAT /PURGE /MIR /R:2 /W:5 
-
-------------------------------------------------------------------------------
-
-	                  13	c:\Users\DELL\Desktop\RAG_backup2\
-	  *EXTRA File 		      60	exclude_list.txt
-	    Newer     		     378	.env.example
-	    New File  		      46	.gitattributes
-	    Newer     		     194	app.py
-	    New File  		    1198	CONFIG_NOTES.md
-	    Newer     		     295	docker-compose.yml
-	    Newer     		     782	Dockerfile
-	    New File  		    6675	martech_sol_logo.jpg
 99%  
-	    Newer     		     161	requirements.txt
-	    New File  		       0	robocopy_backup.log
-	                   3	c:\Users\DELL\Desktop\RAG_backup2\app\
-	    Newer     		    8109	main.py
-	    Newer     		   10861	ui_gradio.py
-	                   2	c:\Users\DELL\Desktop\RAG_backup2\app\admin\
-	    Newer     		   14492	router.py
-	                   1	c:\Users\DELL\Desktop\RAG_backup2\app\admin\templates\
-	    Newer     		   35075	admin.html
-	                   2	c:\Users\DELL\Desktop\RAG_backup2\app\api\
-	    Newer     		     693	schemas.py
-	                   2	c:\Users\DELL\Desktop\RAG_backup2\app\core\
-	    Newer     		    2594	config.py
-	                  10	c:\Users\DELL\Desktop\RAG_backup2\app\services\
-	    Newer     		    3568	chunker.py
-	    Newer     		    8361	llm.py
-	    Newer     		    4989	rag_pipeline.py
-	    New File  		    1661	reranker.py
-	    Newer     		   10783	session_store.py
-	    Newer     		    4823	vector_store.py
-	                   0	c:\Users\DELL\Desktop\RAG_backup2\data\
-	                   1	c:\Users\DELL\Desktop\RAG_backup2\data\index\
-	                   1	c:\Users\DELL\Desktop\RAG_backup2\data\sessions\
-	                   2	c:\Users\DELL\Desktop\RAG_backup2\docs\
-	                   1	c:\Users\DELL\Desktop\RAG_backup2\floaitng-icon\
-	  *EXTRA File 		   24856	Artificial Intelligence Chatbot.lottie
-	    New File  		  462684	KhloeSAC.lottie
 99%  
-	  New Dir          3	c:\Users\DELL\Desktop\RAG_backup2\PHPMailer\
-	    New File  		    1256	Exception.php
 93%  
-	    New File  		  183229	PHPMailer.php
 99%  
-	    New File  		   49148	SMTP.php
 99%  
-	                   3	c:\Users\DELL\Desktop\RAG_backup2\static\
-	    Newer     		   19723	addon.html
-	    New File  		  462684	bot-icon.lottie
 99%  
-	    Newer     		    2010	chat-loader.js
-	                   3	c:\Users\DELL\Desktop\RAG_backup2\wordpress code\
-
-------------------------------------------------------------------------------
-
-               Total    Copied   Skipped  Mismatch    FAILED    Extras
-    Dirs :        21         1        20         0         0         0
-   Files :        47        28        19         0         0         2
-   Bytes :    5.00 m    1.23 m    3.76 m         0         0    24.3 k
-   Times :   0:00:00   0:00:00                       0:00:00   0:00:00
-
-
-   Speed :           20,257,375 Bytes/sec.
-   Speed :            1,159.136 MegaBytes/min.
-   Ended : Wednesday, April 29, 2026 9:00:29 PM
-