https://github.com/ranfysvalle02/hello-docling
"Hello world" but for docling.
https://github.com/ranfysvalle02/hello-docling
Last synced: 7 months ago
JSON representation
"Hello world" but for docling.
- Host: GitHub
- URL: https://github.com/ranfysvalle02/hello-docling
- Owner: ranfysvalle02
- Created: 2024-11-08T03:33:16.000Z (11 months ago)
- Default Branch: main
- Last Pushed: 2024-11-08T03:43:14.000Z (11 months ago)
- Last Synced: 2025-01-26T07:08:57.471Z (9 months ago)
- Language: Python
- Homepage:
- Size: 748 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# hello-docling
[Docling](https://ds4sd.github.io/docling/) is an open-source Python library for efficient document conversion and analysis.
## How It Works
1. **PDF Parsing**: Extracts text tokens and renders page images
2. **AI Model Application**: Applies specialized models to each page
3. **Post-processing**: Augments metadata, detects language, corrects reading order
4. **Assembly**: Combines results into a structured document object
5. **Output**: Generates JSON or Markdown output
Docling runs entirely locally on commodity hardware, making it a powerful tool for document analysis and conversion without relying on external services.
## Key Features
### 1. PDF Conversion
- Converts PDF documents to JSON or Markdown format
- Fast and stable processing
### 2. Advanced Layout Analysis
- Utilizes AI models (DocLayNet) for detailed page layout understanding
- Recovers reading order and document structure
### 3. Table Structure Recognition
- Employs TableFormer AI model to extract and reconstruct table structures
### 4. Metadata Extraction
- Automatically extracts document metadata:
- Title
- Authors
- References
- Language
### 5. OCR Capability
- Optional OCR for scanned PDFs
### 6. Post-processing
- Removes overlapping bounding-box proposals
- Intersects predictions with text tokens for complete content units
### 7. Integration
- Works with Quackling for optimized vector embedding and chunking
- Compatible with LLM frameworks like LlamaIndex