What kind of change does this PR introduce? (Bug fix, feature, docs update, ...)
Bug fix
What is the current behavior? (You can also link to an open issue here)
As previously implemented, users who have "embed", "add_documents", and "query" requests on old indexes with e5 models will no longer receive the same output as before, since they will not have known/implemented about the “content_type” field in embed, and will not have set their own prefixes on the (1) request level, and (2) index settings. Therefore, marqo prefixes their documents with the model default prefixes.
What is the new behavior (if this is a feature change)?
Cases:
Old indexes, when read by marqo 2.6 will have the new fields text_query_prefix and text_chunk_prefix populated by None, as is the default in MarqoIndex and IndexSettings
New indexes user populated with values in text_query_prefix and text_chunk_prefix will use these as the prefixes for the model, unless defined otherwise at the request level. This includes empty strings ""
New indexes without user-populated values in text_query_prefix and text_chunk_prefix
Does this PR introduce a breaking change? (What changes might users need to make in their application due to this PR?)
This PR fixes the breaking change from the original prefix feature implementation
Have unit tests been run against this PR? (Has there also been any additional testing?)
Yes. Also ran against marqo-api-tests
Related Python client changes (link commit/PR here)
Related documentation changes (link commit/PR here)
Other information:
Please check if the PR fulfills these requirements
[x] The commit message follows our guidelines
[x] Tests for the changes have been added (for bug fixes/features)
[ ] Docs have been added / updated (for bug fixes / features)
What kind of change does this PR introduce? (Bug fix, feature, docs update, ...) Bug fix
What is the current behavior? (You can also link to an open issue here) As previously implemented, users who have "embed", "add_documents", and "query" requests on old indexes with e5 models will no longer receive the same output as before, since they will not have known/implemented about the “content_type” field in embed, and will not have set their own prefixes on the (1) request level, and (2) index settings. Therefore, marqo prefixes their documents with the model default prefixes.
What is the new behavior (if this is a feature change)? Cases:
text_query_prefix
andtext_chunk_prefix
populated byNone
, as is the default inMarqoIndex
andIndexSettings
text_query_prefix
andtext_chunk_prefix
will use these as the prefixes for the model, unless defined otherwise at the request level. This includes empty strings""
text_query_prefix
andtext_chunk_prefix
Does this PR introduce a breaking change? (What changes might users need to make in their application due to this PR?) This PR fixes the breaking change from the original prefix feature implementation
Have unit tests been run against this PR? (Has there also been any additional testing?) Yes. Also ran against
marqo-api-tests
Related Python client changes (link commit/PR here)
Related documentation changes (link commit/PR here)
Other information:
Please check if the PR fulfills these requirements
[x] The commit message follows our guidelines
[x] Tests for the changes have been added (for bug fixes/features)
[ ] Docs have been added / updated (for bug fixes / features)