0.5.8

What's Changed

• Compile sensitive terms up front for performance by @olaughter in #170

Full Changelog: v0.5.7...v0.5.8

0.5.7

What's Changed

• Fix/sensitive queries word boundaries by @kdutia in #167
• Add document only searches by @olaughter in #169

Full Changelog: v0.5.6...v0.5.7

0.5.6

• Refactors the search request parameters to work better with web searches from the backend

0.5.5

This Release:

• Integrates a method on the ParserOutput model to allow instantiation of the object from flat json.
• This is as we flatten our json structures when pushing to hugging face.

0.5.4

This Pull Request:

• Updates the numpy sub dependency to allow slightly lower versions.

0.5.3

• Adds a previous_continuation_token to allow navigating backwards through grouped searches
• Switches sorting to use vespa instead of running a sort in post processing

0.5.2

• Enable all results based searches

0.5.1

• Makes continuation tokens optional for searches again, following a validation issue that effectively meant they where a required field

0.5.0

• Adds the ability to get the next set of family or passage level results in searches via continuation tokens. Note that this is a minor breaking change, as the field name has changed from continuation_token to continuation_tokens, it also no longer accepts a string, only a list

0.4.9

• Fixed issue with search where different filter types would be searched as an OR, rather than an AND
• Adds counts for families and passages to query result