Signature extraction

[baderdean]

1. Fetch only the non-html part of the email and send it to a queue
2. Read emails from Redis Stream
3. Detect signature from non-HTML part of emails in the last 20 lines only
4. Extract automatically all information from signature using regexp and ENR including:

• fullname
• avatar
• title
• company
• address
• phone
• social networks: linkedin, X at least
• other URLS

6. Insert them to a temporary table
7. Group them by email to build the most complete and recent data for every contact
8. Insert the results to
9. Leadminer endpoint will verify the task-id, the token and insert them to the database, then notify the user that his data has been enriched

Inspiration:

• https://github.com/mailgun/talon
• https://towardsdatascience.com/named-entity-recognition-using-the-enron-email-dataset-in-python-part-2-fb7ce69bc473
• competitors: https://sigparser.com/ and https://parsio.io/email-signature-parser/
• https://medium.com/aimonks/ner-with-nltk-and-spac-f147ff8accbf

Others:
inspiration:
https://github.com/disklosr/spotd
https://github.com/mlightner/contact_parser
https://github.com/inders/signature-in-email
https://github.com/thrustlabs/contact-parser
https://github.com/convosense/email_signature_remover
https://github.com/Statosco/Email-Signature-Extractor-using-BERT-and-Gemini
https://github.com/KimathiNewton/emails-signature-info
https://github.com/ruslyrossi47/email_signature
https://github.com/captamericadevs/ContactBuilder
https://github.com/edogola4/Email-Signature-Extraction-with-LLMs
https://github.com/leomarcel/email_parser
https://github.com/kartiikthakur/Email-Zoning--ML-Project