This PR adds a full text search function to the research site. This fixes a problem where lots of our long form research is searchable within a document, but not between documents.
Technically, this is using haystack and whoosh to keep it simple (no extra solr service), given low volume of traffic.
Index creation is handled in the normal process when documents are added or removed. An admin action for ResearchItem can do this as a batch.
To start with, this indexes mysociety.org blog posts and stringprint documents referenced by the site.
In principle could be extended to search within tags or specific documents, but the full search seems the most useful thing. Can always add that next Christmas.
This PR adds a full text search function to the research site. This fixes a problem where lots of our long form research is searchable within a document, but not between documents.
Technically, this is using haystack and whoosh to keep it simple (no extra solr service), given low volume of traffic.
Index creation is handled in the normal process when documents are added or removed. An admin action for ResearchItem can do this as a batch.
To start with, this indexes mysociety.org blog posts and stringprint documents referenced by the site.
In principle could be extended to search within tags or specific documents, but the full search seems the most useful thing. Can always add that next Christmas.