issues-search

use whoosh to search github issues.

Implemented in Python using Flask, Whoosh and Mistune.

notes

had to update how the update process worked:

• seems inefficient, but re-adding every document every index update
• works fine if not updating all the time
• gonna need to grab the content every time anyway
• need to add a field that's the content's hash
• compare indexed content's hash with fetched content's hash

when changing the schema:

• the schema changes, so the document addition changes (args)
• but also,
  • in issues_app.py we call the search() method on our Search object
  • in issues_search.py, search() method calls create_search_result()
  • create_search_result() extracts particular fields and bundles them up
  • the search results are ultimately returned to the app's /search endpoint, which passes them to a template as-is: search.html
  • keyword: entries = result
  • therefore, search.html and create_search_result() must also change to reflect changes in field names

summary of how to change the schema:

• issues_app.py:
  • just combines the schema with the template, nothing to update
• issues_search.py:
  • main place where schema is defined/used
  • Search class defines Schema object, main definition
  • add_issue() (equally important) which defines how to extract the fields defined in the schema from the document
  • create_search_result() (also important) which packages up the search results for the template to deal with
• search.html: the search.html template uses a different variable namespace than the Python file issues_search.py or the flask app
  • The create_search-result() method of issues_search.py defines how search results are parsed and packaged for the search.html template
  • Jinja variables used in search.html should be defined in create_search_result() method of issues_search.py