Spaces:
Running
Running
Upload 3 files
Browse files- README.md +35 -6
- app.py +20 -0
- requirements.txt +15 -0
README.md
CHANGED
|
@@ -1,14 +1,43 @@
|
|
| 1 |
---
|
| 2 |
-
title: Project
|
| 3 |
-
emoji:
|
| 4 |
-
colorFrom:
|
| 5 |
-
colorTo:
|
| 6 |
sdk: gradio
|
| 7 |
sdk_version: 6.3.0
|
| 8 |
app_file: app.py
|
| 9 |
pinned: false
|
| 10 |
license: mit
|
| 11 |
-
short_description:
|
| 12 |
---
|
| 13 |
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
title: Project Intelligence Hub
|
| 3 |
+
emoji: π
|
| 4 |
+
colorFrom: blue
|
| 5 |
+
colorTo: purple
|
| 6 |
sdk: gradio
|
| 7 |
sdk_version: 6.3.0
|
| 8 |
app_file: app.py
|
| 9 |
pinned: false
|
| 10 |
license: mit
|
| 11 |
+
short_description: PDF reports to queryable knowledge graph
|
| 12 |
---
|
| 13 |
|
| 14 |
+
# Project Intelligence Hub
|
| 15 |
+
|
| 16 |
+
Transform PDF project reports into a queryable Neo4j knowledge graph with natural language queries.
|
| 17 |
+
|
| 18 |
+
## Features
|
| 19 |
+
|
| 20 |
+
- **PDF Ingestion** β Extract entities and relationships from project reports
|
| 21 |
+
- **Knowledge Graph** β Build structured graph with Projects, Budgets, Locations, Milestones, Challenges
|
| 22 |
+
- **Hybrid Search** β Vector embeddings + BM25 keyword search
|
| 23 |
+
- **Natural Language Q&A** β Query the graph using plain English
|
| 24 |
+
|
| 25 |
+
## Setup (Required Secrets)
|
| 26 |
+
|
| 27 |
+
Configure these in **Settings > Repository secrets**:
|
| 28 |
+
|
| 29 |
+
| Secret | Description |
|
| 30 |
+
|--------|-------------|
|
| 31 |
+
| `TOGETHER_API_KEY` | Together AI API key |
|
| 32 |
+
| `NEO4J_URI` | Neo4j connection URI (e.g., `neo4j+s://xxx.databases.neo4j.io`) |
|
| 33 |
+
| `NEO4J_USERNAME` | Neo4j username (usually `neo4j`) |
|
| 34 |
+
| `NEO4J_PASSWORD` | Neo4j password |
|
| 35 |
+
| `NEO4J_DATABASE` | Neo4j database name (usually `neo4j`) |
|
| 36 |
+
|
| 37 |
+
## Tech Stack
|
| 38 |
+
|
| 39 |
+
- **LLM**: Together AI (DeepSeek-V3)
|
| 40 |
+
- **Embeddings**: m2-bert-80M-8k-retrieval
|
| 41 |
+
- **Graph DB**: Neo4j Aura
|
| 42 |
+
- **Reranker**: ms-marco-MiniLM-L-6-v2
|
| 43 |
+
- **UI**: Gradio
|
app.py
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Project Intelligence Hub - Hugging Face Spaces Entry Point."""
|
| 2 |
+
|
| 3 |
+
import os
|
| 4 |
+
|
| 5 |
+
# Ensure environment variables are loaded from HF Secrets
|
| 6 |
+
# Users set these in Settings > Repository secrets on HF Spaces
|
| 7 |
+
|
| 8 |
+
from src.ui.gradio_app import GradioApp
|
| 9 |
+
|
| 10 |
+
if __name__ == "__main__":
|
| 11 |
+
app = GradioApp()
|
| 12 |
+
app.launch(
|
| 13 |
+
server_name="0.0.0.0",
|
| 14 |
+
server_port=7860,
|
| 15 |
+
share=False,
|
| 16 |
+
)
|
| 17 |
+
else:
|
| 18 |
+
# For Hugging Face Spaces - module-level app object
|
| 19 |
+
app = GradioApp()
|
| 20 |
+
demo = app.build()
|
requirements.txt
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
gradio==6.3.0
|
| 2 |
+
python-dotenv==1.0.1
|
| 3 |
+
pydantic==2.12.5
|
| 4 |
+
neo4j==5.22.0
|
| 5 |
+
pypdf==4.2.0
|
| 6 |
+
together==1.3.14
|
| 7 |
+
langchain==0.2.17
|
| 8 |
+
langchain-core==0.2.43
|
| 9 |
+
langchain-community==0.2.19
|
| 10 |
+
langchain-experimental==0.0.60
|
| 11 |
+
langchain-together==0.1.5
|
| 12 |
+
langchain-text-splitters==0.2.4
|
| 13 |
+
tiktoken==0.7.0
|
| 14 |
+
sentence-transformers
|
| 15 |
+
httpx>=0.24.1
|