Skip to main content

What is GovDoc Scanner?

Project Overview​

Problem Statement:​

In Greece, vital public company data is often locked in unstructured PDF files, making it challenging for citizens, researchers, and policymakers to access and analyze this information. The current state of these documents limits transparency and hinders efficient data use. The Flexible GovDoc Scanner project seeks to bridge this gap by transforming these PDFs into a structured, searchable database, thereby democratizing access to important corporate information.

What I am Making:​

The Flexible GovDoc Scanner is an open-source tool designed to convert unstructured GEMI portal PDFs into a fully searchable database accessible via a REST API. The solution comprises the following steps:

  • Fetch: Systematically retrieve documents from Greece’s official Open Data Portal, ensuring full compliance with legal and ethical standards.
  • Extract: Leverage Google Gemini's generative AI capabilities to accurately extract metadata and text from the documents, eliminating the need for traditional OCR.
  • Organize: Index and store the extracted data in OpenSearch, taking advantage of its built-in language analyzers for Greek text and fast querying capabilities.
  • Share: Develop a robust REST API that allows users to query the database by various parameters such as company name, incorporation date, and key individuals, with features like pagination, filtering, and rate limiting.

Tools Used​

  • Node.js: The runtime environment for all scripts and applications.
  • Nx: A smart, extensible build framework for monorepo management.
  • Google Gemini: Used for its generative AI capabilities in document processing and metadata extraction.

Something Missing?​

If something is missing in the documentation or if you found some part confusing, please click on edit this Page and create a PR for improvement. We love your contribution!