Spaces:
Running
Running
Hasan-Atris3
commited on
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
title: CedroPM Bot
|
| 3 |
+
emoji: π€
|
| 4 |
+
colorFrom: blue
|
| 5 |
+
colorTo: indigo
|
| 6 |
+
sdk: docker
|
| 7 |
+
pinned: false
|
| 8 |
+
app_port: 7860
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# PM-RAG-ChatBot (CedroPM)
|
| 12 |
+
|
| 13 |
+
A specialized RAG (Retrieval Augmented Generation) chatbot designed for Project Managers and Software Architects. It processes Software Requirements Documents (SRDs) and provides expert-level analysis by combining project-specific data with a pre-seeded "Gold Standard" knowledge base of industry best practices.
|
| 14 |
+
|
| 15 |
+
## π Key Features
|
| 16 |
+
|
| 17 |
+
* **π Intelligent SRD Analysis**: Upload PDF requirements and get answers grounded strictly in your document.
|
| 18 |
+
* **π§ Global Expert Knowledge**: Pre-seeded with 30+ industry standards (ISO, OWASP, NIST, Agile) to provide "Senior Architect" advice even when your PDF is silent.
|
| 19 |
+
* **π¨ Diagram Vision**: Understands and explains architecture diagrams using Claude Vision or Qwen2-VL.
|
| 20 |
+
* **π Hybrid Search**: Uses a dual-scope retrieval engine to search your **Project PDF** and **Global Best Practices** simultaneously.
|
| 21 |
+
* **π¬ Interactive Chat**: Built with Chainlit for a clean, chat-like interface with history retention.
|
| 22 |
+
* **π Secure & Private**: Multi-user isolation ensures one project's data never leaks to another.
|
| 23 |
+
|
| 24 |
+
## ποΈ System Architecture
|
| 25 |
+
|
| 26 |
+
1. **Ingestion Layer**:
|
| 27 |
+
* **PDFs**: Processed with `pdfplumber` and `SmartSRDSplitter` (Section-aware chunking).
|
| 28 |
+
* **Tables**: Extracted via `camelot` and converted to Markdown.
|
| 29 |
+
* **Diagrams**: Processed via OCR (`Tesseract`) or Vision LLM (`Claude 3.5 Sonnet`).
|
| 30 |
+
2. **Knowledge Store**:
|
| 31 |
+
* **ChromaDB**: vector storage for semantic search.
|
| 32 |
+
* **Dual-Scope**: Queries filter by `(User + Project)` OR `(Global_Expert_Knowledge)`.
|
| 33 |
+
3. **Reasoning Engine**:
|
| 34 |
+
* **Claude 3.5 Sonnet**: Generates detailed, structured responses.
|
| 35 |
+
* **Seed Data**: A Python script injects "Golden Rules" for Security, DevOps, and PM methodologies.
|
| 36 |
+
|
| 37 |
+
## π Prerequisites
|
| 38 |
+
|
| 39 |
+
* **Python 3.10+**
|
| 40 |
+
* **System Tools**:
|
| 41 |
+
* `Poppler` (for PDF rendering)
|
| 42 |
+
* `Tesseract OCR` (for image text extraction)
|
| 43 |
+
* **API Key**: Anthropic API Key (for Claude).
|
| 44 |
+
|
| 45 |
+
## π§ Installation & Setup
|
| 46 |
+
|
| 47 |
+
### 1. Clone the Repository
|
| 48 |
+
```bash
|
| 49 |
+
git clone <your-repo-url>
|
| 50 |
+
cd PM-RAG-ChatBot
|