Improve DOI extraction by refining regex to handle legacy and alternative citation formats

[Aakash-pal]

Summary

This PR improves DOI extraction by refining the regex and fallback detection logic to better handle real-world formats.

Changes

• Improved regex to prevent over-capturing (e.g., avoiding cases like 10.xxxx/abc)DOI)
• Ensured proper handling of punctuation and edge cases
• Validated extraction against multiple DOI formats including:
  • doi:10.xxxx/xxxx
  • https://doi.org/...
  • Bare DOIs
  • DOIs with punctuation
  • DOIs with addition space (DOI : 10.xxx)

Result

• More robust DOI detection
• Reduced false positives
• Improved handling of messy real-world text inputs

Testing

Tested using multiple DOI patterns and edge cases to ensure correctness and consistency.

Full sample input file attached in PR for reference.
sample.txt

The output:

Closes #470