Azure AI Service: Cannot import name 'SemanticSettings' from 'azure.search.documents.indexes.models'

You may encounter the following exception while testing Semantic Ranking with the Azure AI service. I was exploring the 'Quickstart: Semantic Ranking' guide on Azure AI Search, available on Microsoft Learn, using the Python library. 

The exception occurred during the execution of the following code snippet:

%pip install azure-search-documents --pre

%pip show azure-search-documents

%pip install python-dotenv

import os

from azure.core.credentials import AzureKeyCredential

from azure.search.documents.indexes import SearchIndexClient

from azure.search.documents import SearchClient

from azure.search.documents.indexes.models import (  

    SearchIndex,  

    SearchField,  

    SearchFieldDataType,  

    SimpleField,  

    SearchableField,

    ComplexField,

    SearchIndex,  

    SemanticConfiguration,  

    PrioritizedFields,  

    SemanticField,  

    SemanticSettings,  

)

ImportError: cannot import name 'SemanticSettings' from 'azure.search.documents.indexes.models'

Additionally, you might encounter the following exception when executing the sample code from the aforementioned document.

ImportError: cannot import name 'PrioritizedFields' from 'azure.search.documents.indexes.models'

To resolve the issue, replace 'PrioritizedFields' with 'SemanticPrioritizedFields' in every occurrence. Additionally, remove 'prioritized_' from the following code snippet.

semantic_config = SemanticConfiguration(

    name="my-semantic-config",

    prioritized_fields=SemanticPrioritizedFields(

        title_field=SemanticField(field_name="HotelName"),

        prioritized_keywords_fields=[SemanticField(field_name="Category")],

        prioritized_content_fields=[SemanticField(field_name="Description")]

    )

)

Updated

semantic_config = SemanticConfiguration(

    name="my-semantic-config",

    prioritized_fields=SemanticPrioritizedFields(

        title_field=SemanticField(field_name="HotelName"),

        keywords_fields=[SemanticField(field_name="Category")],

        content_fields=[SemanticField(field_name="Description")]

    )

)

The SemanticSettings class is not supported in the latest version of the azure-search-documents library. To resolve the issue, replace SemanticSettings with SemanticSearch.

%pip install azure-search-documents --pre

%pip show azure-search-documents

%pip install python-dotenv

import os

from azure.core.credentials import AzureKeyCredential

from azure.search.documents.indexes import SearchIndexClient

from azure.search.documents import SearchClient

from azure.search.documents.indexes.models import (  

    SearchIndex,  

    SearchField,  

    SearchFieldDataType,  

    SimpleField,  

    SearchableField,

    ComplexField,

    SearchIndex,  

    SemanticConfiguration,  

    SemanticPrioritizedFields,  

    SemanticField,  

    SemanticSearch,  

)

Also, replace SemanticSettings with SemanticSearch throughout your code.

semantic_settings = SemanticSettings(configurations=[semantic_config])

Updated

semantic_settings = SemanticSearch(configurations=[semantic_config])

Also, replace semantic_settings with semantic_search in the code snippet provided below.

index = SearchIndex(

    name=name,

    fields=fields,

    semantic_settings=semantic_settings,

    scoring_profiles=scoring_profiles,

    suggesters = suggester)

Updated

index = SearchIndex(

    name=name,

    fields=fields,

    semantic_search=semantic_settings,

    scoring_profiles=scoring_profiles,

    suggesters = suggester)

Now you can execute the Semantic Search demo without any issues, you may receive a excpetion related to enabling the semantic ranking for your AI Search service, refer to Azure AI Search: Semantic search is not enabled for this service to resolve the issue.

Azure AI Search: Semantic search is not enabled for this service.

You may encounter the following exception while testing Semantic Ranking with the Azure AI service. I was exploring the 'Quickstart: Semantic Ranking' guide on Azure AI Search, available on Microsoft Learn, using the Python library. The exception occurred during the execution of the following code snippet:

results =  search_client.search

(query_type='semantic', semantic_configuration_name='my-semantic-config',

    search_text="what hotel has a good restaurant on site",

    select='HotelName,Description,Category', query_caption='extractive')

for result in results:

    print(result["@search.reranker_score"])

    print(result["HotelName"])

    print(f"Description: {result['Description']}")

    captions = result["@search.captions"]

    if captions:

        caption = captions[0]

        if caption.highlights:

            print(f"Caption: {caption.highlights}\n")

        else:

            print(f"Caption: {caption.text}\n")

The exception encountered is likely due to the Semantic Ranking feature not being enabled for the Azure AI Search Service. To resolve this issue, ensure that Semantic Ranking is activated for the AI Search Service instance you are using for semantic searches. It's important to note that the Semantic Ranking feature is not available for the free tier of Azure AI Search.

Ensure that your Azure AI Search Service is set up with a pricing tier of Basic or higher.

Now, with the appropriate pricing tier selected, you should be able to enable the Smart Ranking plan, allowing your code to execute without any issues.