Knowledge Extraction for COVID-19 Publications (KEP)

Content

This package provides the functionality of extracting and displaying key knowledge to enable a rapid understanding of COVID-19 publications. Current focuses include key topics and top disease and location mentions of the input COVID-19 publication. More functions are on the way.

Prerequisites

Install packages before use (with tested versions):

• Python >=3.8
• spacy (3.0.8)
• scispacy (0.4.0)
• gensim (4.1.2)
• nltk (3.7)
• en-core-web-sm (https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.0.0/en_core_web_sm-3.0.0.tar.gz)
• en-ner-bc5cdr-md (https://s3-us-west-2.amazonaws.com/ai2-s2-scispacy/releases/v0.4.0/en_ner_bc5cdr_md-0.4.0.tar.gz)
• bs4 (BeautifulSoup4)
• wordcloud
• pandas
• matplotlib

Default:

• re
• string
• os
• urllib
• xml

Local package:

• nltk_data (https://github.com/nltk/nltk_data/tree/gh-pages/packages) (rename the "packages" folder as "nltk_data" and put it in your own python project)

Instruction

Load the KEP package

pip install KEP

Sample code

from KEP import To_Generate_Disease,To_Generate_Key_Word,To_Generate_Location,To_Generate_All 

To_Generate_All(7824075)



# If multiple publications at a time, uncomment the following lines and replace the PMC ID:
# To_Generate_All(7824075)
# To_Generate_All(7824470)
# To_Generate_All(6988269)
# ...

Input

Input the PMC ID of a publication (or a set of PMC IDs)

(For example: input "7824075" representing the publication https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7824075/)

Output

1. The bar graph and word cloud of the key topics
2. The bar graph of the top disease mentions
3. The bar graph of the top location mentions

References

1. Chen, Q., Allot, A., & Lu, Z. (2020). Keep up with the latest coronavirus research. Nature, 579(7798), 193-194.
2. Comeau, D. C., Wei, C. H., Islamaj Doğan, R., & Lu, Z. (2019). PMC text mining subset in BioC: about three million full-text articles and growing. Bioinformatics, 35(18), 3533-3535.
3. BioC API for PMC: https://www.ncbi.nlm.nih.gov/research/bionlp/APIs/BioC-PMC/
4. Part of the functionality is based on the en-ner-bc5cdr-md, en-core-web-sm, and gensim packages.

Citation

TBA.

Report an issue

Should you have any questions or comments, please feel free to contact the author.