- Primarily intended to provide a document management platform for text based NLP.
- This is a multi-phase project:
- Text document management backend and API layer.
- Text analytics layer using AllenNLP, Pythorch, and similar frameworks.
- Front end using React.
- Cloud and on-premise hosting.
- Can be leveraged as a foundation for building a commercial product as well as a learning tool.
- Create/update/delete/fetch of text documents, both embeded and file based.
- Document search using date range, name, category, and tag criteria applying 'in', exact match, and range criteria.
- Association of document with tags and category for classification and search.
- Annotation of document.
- Linking of documents.
- Support for user, group, client, and client user abstractions which govern document access.
- Token based authentication.
- Group based access authorization to documents and other abstractions.
- Basic administration portal.
- Auto-generated OpenAPI compliant API.
- Ability to select model family and implementation.
- perform reading comprehension analysis using AllenNLP using BDAF and BDAFNAQNAET.
- Ability to cache and save to db the analysis results.