Automatic-Text-Summarizer
Automatic Text Summarizer with various approaches in NLP and Django
Approahces used
- LexRank
- LSA (Latent semantic analysis)
- KL-Sum
- Summarization with T5 Transformers
- Summarization with BART Transformers
- Summarization with GPT-2 Transformers
- Summarization with XLM Transformers
- Luhn
Technologies
- Django Framework
- NLTK library for NLP
- Postgresql ( currently using Sqlite )
- Django REST Framework
Improvements for the future
- Inputs not taken for the document text
- Integrating with Postgresql
- API creation with DRF
- Analysing the performance of the results of all the algos
- Manually Hypertuning the algos to increase performance
- Adding animations and proper styling
- A Dashboard to compare the results of all the algos
- ( optional ) Analysis of the text data using elasticsearch ?
Issues
- Output file is not automatically created ( might use callback, asynchronous ? )
- The form field of the input algo and text area is not styled properly
- About Page is blank
- Slow loading of the algo's ( add Redis for caching ? )
New Ideas
- Uploading PDF file, extracting text from the file and summarizing it
- Uploading Image file, extracting text from the file and summarizing it ( optional )
- Obtaining Abbreviation from the document
- Knowledge Graphs ( https://www.analyticsvidhya.com/blog/2019/10/how-to-build-knowledge-graph-text-using-spacy/ )
- Words Cloud
- Text Evaluation ( metrics )