axa-group/Parsr
Parsr
Parsr is an unstructured data extractor and document parsing pipeline that converts raw files and images into cleaned, machine-readable formats. It functions as a document layout analyzer and a pipeline for extracting structured data and labels using large language models.
The system includes a document parsing visualizer, providing a graphical interface to upload documents and inspect the resulting structured data output.
The project covers document digitization workflows, including layout analysis to detect headings, tables, and lists, and automated data entry through the cleaning and enrichment of unstructured content.
Features
- Information Extraction - Uses large language models to map unstructured text segments into predefined structured data schemas.
- Document Layout Analysis - Detects structural elements like tables and headings to reconstruct the original hierarchy of a document.
- Optical Character Recognitions - Transforms image-based documents into machine-readable formats using optical character recognition and structural parsing.
- Document Digitization Tools - Converts physical or digital image files into clean text and organized data for downstream applications.
- Document Layout Analyzers - Systematically detects structural elements to reconstruct the original hierarchy of a document.
- Semantic Document Structuring - Detects headings, tables, and lists to regenerate the original semantic hierarchy of a document.
- Document Parsing Pipelines - Implements a pipeline using large language models to extract structured data and labels from documents.
- Document and Unstructured Extraction - Turns raw documents and images into structured data formats for automated processing and machine reading.
- Schema-Driven Extraction - Organizes extracted document fragments into a structured hierarchy based on target data definitions.
- Structured Data Extraction - Transforms unstructured files and images into enriched data formats for machine reading.
- Unstructured Data Transformation Tools - Converts raw files and images into cleaned, machine-readable formats for data entry and automation.
- Automated Data Entry Tools - Prepares unstructured content for direct entry into databases by cleaning and enriching extracted data.
- Multi-Stage Text Normalizers - Implements a sequential pipeline of normalization and noise removal passes to improve document data quality.
- Data Cleaning Pipelines - Processes raw document inputs to produce label-enriched and cleaned information for automation.
- Document Format Converters - Transforms images and files into organized text or data formats for downstream applications.
- Visual Document Parsing - Provides visual parsing capabilities to verify extracted structured data against original images.
- Parsing Result Inspectors - Displays a graphical interface to upload documents and visually inspect the resulting structured data.
- Data Extraction Visualizers - Provides a graphical tool to compare original document layouts against the extracted structured data output.
- Parsing Visualizers - Offers a graphical interface for uploading documents and inspecting the resulting structured data output.
- OCR - Document-to-structured-data conversion tool.
Star history
How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.
