sibtainrazajamali / comment_generation

Generate text based on entities extracted by "ner-english-ontonotes-large" model

Comment_Generation

Generate text based on entities extracted by "ner-english-ontonotes-large" model from input text.

The recognized entities are below

          | entity      | meaning              |
          |-------------+----------------------|
          | CARDINAL    | cardinal value       |
          | DATE        | date value           |
          | EVENT       | event name           |
          | FAC         | building name        |
          | GPE         | geo-political entity |
          | LANGUAGE    | language name        |
          | LAW         | law name             |
          | LOC         | location name        |
          | MONEY       | money name           |
          | NORP        | affiliation          |
          | ORDINAL     | ordinal value        |
          | ORG         | organization name    |
          | PERCENT     | percent value        |
          | PERSON      | person name          |
          | PRODUCT     | product name         |
          | QUANTITY    | quantity value       |
          | TIME        | time value           |
          | WORK_OF_ART | name of work of art  |

Extract Topic of input text by facebook/bart-large-mnli' model, the input text is classified into these categories ["artifacts", "animals", "food", "sport","technology",'travel', 'exploration', 'dancing', 'cooking']

'gpt2' is used as pretrained model for generating text

PERSON,GPE,ORG,DATE entities extracted from input text are combined and combination of (PERSON+" "+GPE),(PERSON+" "+ORG),(PERSON+" "+DATE) are passed to gpt2 text generation model Finally new data in the form of {"_id":generated text,"topic":extracted topic} saved in database

Requirements

transformers

flair

Usage

Pass the parameters of Comment_Generator Class to create a database connection for example: coll="my_coll", conn=pymongo.MongoClient("mongodb://localhost:27017/"), db_name='my_db'

Run main.py