Analyze a PDF to find relevant meta and content details

Hangul is an NLP-based assistant for digital curators at ReliefWeb envisioned to enable them to handle three to four times the number of documents currently being processed. Once a text PDF is uploaded to the platform, relevant metadata is extracted from it.

Current metadata includes the document title, the date the document was published and modified, the number of pages in the document, the word length of the document, the language of the document, and the entities in the document. More complex features like extraction of abstract, conclusion, executive summary, and recognizing the theme (cluster) of the document are also in scope.