Updates to integrated vectorization #2045
Conversation
| if self.embeddings is None: | ||
| raise ValueError("Expecting Azure Open AI instance") | ||
|
|
||
| await search_manager.create_index( |
There was a problem hiding this comment.
Now we always make a vectorizer, even when not using integrated vectorization.
| else SearchField( | ||
| name="id", | ||
|
|
||
| if self.search_info.index_name not in [name async for name in search_index_client.list_index_names()]: |
There was a problem hiding this comment.
This code change looks larger than it is, because I moved all the index creation logic inside the if statement. Before, we would create the index object even if we didn't use it, which seemed an odd thing to do!
| sortable=False, | ||
| facetable=False, | ||
| vector_search_dimensions=1024, | ||
| vector_search_dimensions=self.embedding_dimensions, |
There was a problem hiding this comment.
Git is not diffing this field correctly, the imageEmbedding field is still 1024.
| prioritized_fields=SemanticPrioritizedFields( | ||
| title_field=None, content_fields=[SemanticField(field_name="content")] | ||
| vectorizers = [] | ||
| if self.embeddings and isinstance(self.embeddings, AzureOpenAIEmbeddingService): |
There was a problem hiding this comment.
This is new code, adding the vectorizer
| existing_index.vector_search.vectorizers is None | ||
| or len(existing_index.vector_search.vectorizers) == 0 | ||
| ): | ||
| if self.embeddings is not None and isinstance(self.embeddings, AzureOpenAIEmbeddingService): |
There was a problem hiding this comment.
This is new code, adding a vectorizer if one doesnt exist yet, to make it easy for folks to upgrade
| azure-ai-documentintelligence | ||
| azure-cognitiveservices-speech | ||
| azure-search-documents==11.6.0b1 | ||
| azure-search-documents==11.6.0b6 |
There was a problem hiding this comment.
I didn't get any errors about code changes on the .search() side so I'm assuming that's largely unchanged in the latest versions?
| filterable=False, | ||
| sortable=False, | ||
| facetable=False, | ||
| vector_search_dimensions=1024, |
Purpose
Fixes #1825
Updates to latest Azure AI search package and specify embedding model name and dimensions
Does this introduce a breaking change?
When developers merge from main and run the server, azd up, or azd deploy, will this produce an error?
If you're not sure, try it out on an old environment.
Does this require changes to learn.microsoft.com docs?
This repository is referenced by this tutorial
which includes deployment, settings and usage instructions. If text or screenshot need to change in the tutorial,
check the box below and notify the tutorial author. A Microsoft employee can do this for you if you're an external contributor.
Type of change
Code quality checklist
See CONTRIBUTING.md for more details.
python -m pytest).python -m pytest --covto verify 100% coverage of added linespython -m mypyto check for type errorsruffandblackmanually on my code.