[AutoPR azure-ai-contentsafety]-generated-from-SDK Generation - Python-6652880

[azure-sdk-automation]

Configurations: 'specification/cognitiveservices/ContentSafety/tspconfig.yaml', SDK Release Type: beta, and CommitSHA: '926fc28f53c318fe3cf68eb8be1ea38e658c0879' in SpecRepo: 'https://github.com/Azure/azure-rest-api-specs' Pipeline run: https://dev.azure.com/azure-sdk/internal/_build/results?buildId=6652880 Refer to https://eng.ms/docs/products/azure-developer-experience/develop/sdk-release/sdk-release-prerequisites to prepare for SDK release. Release plan link: https://azsdk-releaseplan-dashboard-hveph5aqhhcfhtgu.westus-01.azurewebsites.net/?releaseplan=2242 Submitted by: Connor.Wood@microsoft.com

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).
9 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

[Copilot]

Pull request overview

Note

Copilot could not run the full agentic suite for this review because it was automatically requested on a bot-authored pull request. Request a review from Copilot under Reviewers to retry with the full agentic suite. Improved support for bot-authored pull requests is coming soon.

Updates the azure-ai-contentsafety Python SDK to a newer REST spec commit and regenerates the client to include new preview APIs (Content Provenance, Shield Prompt, protected material), along with modernized packaging metadata.

Changes:

• Bumps generated code/spec commit and package version to 1.1.0b1, adding Content Provenance Detection, Shield Prompt, and protected-material APIs.
• Migrates packaging from setup.py/sdk_packaging.toml to PEP 621 pyproject.toml and updates dependencies / supported Python versions.
• Adds generated samples/tests plus new sanitizer setup for recordings.

Reviewed changes

Copilot reviewed 62 out of 63 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
sdk/contentsafety/azure-ai-contentsafety/tsp-location.yaml	Updates swagger/spec commit reference; adjusts additional directory config.
sdk/contentsafety/azure-ai-contentsafety/tests/conftest.py	Minor formatting change in test proxy fixture file.
sdk/contentsafety/azure-ai-contentsafety/setup.py	Removes legacy setuptools build script (migration to pyproject).
sdk/contentsafety/azure-ai-contentsafety/sdk_packaging.toml	Removes old packaging config (moved into pyproject).
sdk/contentsafety/azure-ai-contentsafety/samples/sample_manage_blocklist.py	Adds pylint disables for generated-sample linting.
sdk/contentsafety/azure-ai-contentsafety/pyproject.toml	Introduces PEP 621 metadata / build-system; moves packaging config.
sdk/contentsafety/azure-ai-contentsafety/generated_tests/testpreparer_async.py	Adds generated async test base/preparer helpers.
sdk/contentsafety/azure-ai-contentsafety/generated_tests/testpreparer.py	Adds generated sync test base/preparer helpers.
sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_safety_async.py	Adds generated async tests (currently skipped).
sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_safety.py	Adds generated sync tests (currently skipped).
sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_provenance_async.py	Adds generated async tests for provenance APIs (skipped).
sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_provenance.py	Adds generated sync tests for provenance APIs (skipped).
sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_blocklist_async.py	Adds generated async tests for blocklist APIs (skipped).
sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_blocklist.py	Adds generated sync tests for blocklist APIs (skipped).
sdk/contentsafety/azure-ai-contentsafety/generated_tests/conftest.py	Adds sanitizer/test-proxy setup for generated recordings.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/shield_prompt.py	Adds generated sample for Shield Prompt API.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/remove_blocklist_items.py	Adds generated sample for removing blocklist items.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/provenance_get_operation_status.py	Adds generated sample for querying provenance LRO status.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/provenance_detect.py	Adds generated sample for starting provenance detection LRO.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/list_text_blocklists.py	Adds generated sample for listing blocklists.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/list_text_blocklist_items.py	Adds generated sample for listing blocklist items.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/get_text_blocklist_item.py	Adds generated sample for fetching a blocklist item.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/get_text_blocklist.py	Adds generated sample for fetching a blocklist.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/detect_text_protected_material.py	Adds generated sample for protected-material detection API.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/delete_text_blocklist.py	Adds generated sample for deleting a blocklist.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/create_or_update_text_blocklist.py	Adds generated sample for creating/updating a blocklist.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/analyze_text.py	Adds generated sample for text analysis API.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/analyze_image.py	Adds generated sample for image analysis API.
sdk/contentsafety/azure-ai-contentsafety/generated_samples/add_or_update_blocklist_items.py	Adds generated sample for adding/updating blocklist items.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/types.py	Adds TypedDict request/shape types for the public surface.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/models/_patch.py	Updates patch header and __all__ typing style.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/models/_models.py	Regenerates models, adds provenance + shield/protected-material models, migrates model base/util imports.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/models/_enums.py	Adds provenance/polling enums; improves enum docstrings.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/models/init.py	Exposes new models/enums and updates import/patch mechanics.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_vendor.py	Removes legacy typing-only vendor mixins (replaced by shared utils).
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_patch.py	Updates patch header and __all__ typing style.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_operations/_patch.py	Updates patch header and __all__ typing style.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_operations/_operations.py	Adds provenance/shield/protected-material ops; refactors mixins and streaming/decompression handling.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_operations/init.py	Updates exports/import pattern for operations mixins.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_configuration.py	Adds ContentProvenance config; updates api version defaults.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py	Adds async ContentProvenanceClient and refactors mixin inheritance; adds Self typing.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/init.py	Exposes ContentProvenanceClient from aio namespace.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_version.py	Bumps package version to 1.1.0b1.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_vendor.py	Removes legacy typing-only vendor mixins (replaced by shared utils).
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_validation.py	Adds API-version gating decorator used by new preview operations.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_utils/utils.py	Adds shared internal ClientMixin ABC to replace vendor files.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_utils/init.py	Adds utils package init for internal helpers.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_patch.py	Updates patch header and __all__ typing style.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_operations/_patch.py	Updates patch header and __all__ typing style.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_operations/init.py	Updates exports/import pattern for operations mixins.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_model_base.py	Removes legacy model base implementation (replaced by new utils model base).
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_configuration.py	Adds ContentProvenance config; updates api version defaults.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py	Adds sync ContentProvenanceClient and refactors mixin inheritance; adds Self typing.
sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/init.py	Exposes ContentProvenanceClient from sync namespace.
sdk/contentsafety/azure-ai-contentsafety/apiview-properties.json	Adds APIView cross-language metadata mappings for new surface area.
sdk/contentsafety/azure-ai-contentsafety/api.metadata.yml	Adds API parser metadata captured during generation.
sdk/contentsafety/azure-ai-contentsafety/api.md	Adds regenerated API signature document for review/validation.
sdk/contentsafety/azure-ai-contentsafety/_metadata.json	Adds generation metadata (api version, emitter version, spec commit).
sdk/contentsafety/azure-ai-contentsafety/MANIFEST.in	Minor formatting update.
sdk/contentsafety/azure-ai-contentsafety/CHANGELOG.md	Adds 1.1.0b1 entry placeholder text.

Comments suppressed due to low confidence (5)

sdk/contentsafety/azure-ai-contentsafety/tsp-location.yaml:1

• additionalDirectories: is now an explicit null value, whereas previously it was an empty list ([]). If tooling expects a list, this can break spec resolution; set this back to additionalDirectories: [] (or provide a proper YAML list).
  sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:1
• PEP 621 project.license is not a plain string; it should be a table (e.g., license = { text = \"MIT\" }) or license = { file = \"LICENSE\" }. As-is, some build backends may reject this metadata.
  sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:1
• [tool.setuptools.package-data] keys must be package names (or \"*\"), not an arbitrary label. This likely prevents py.typed from being included in the built wheel. Use the actual package (e.g. azure.ai.contentsafety) or \"*\" to include py.typed.
  sdk/contentsafety/azure-ai-contentsafety/generated_samples/analyze_text.py:1
• Using bool as a sample value will serialize as a Python type object and will fail JSON serialization (TypeError: Object of type type is not JSON serializable) if the sample is run. Replace with a real boolean placeholder (e.g. True/False) to keep samples runnable.
  sdk/contentsafety/azure-ai-contentsafety/generated_samples/shield_prompt.py:1
• These samples construct clients with credential as a plain string, but the SDK expects an AzureKeyCredential or token credential type. Update the sample to show correct credential creation (e.g. AzureKeyCredential(...) or DefaultAzureCredential()), otherwise the sample will fail at runtime.

[Copilot]

Pull request overview

Copilot reviewed 61 out of 62 changed files in this pull request and generated no new comments.

Suppressed comments (6)

sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:16

• The release version is 1.1.0b1, but this metadata advertises the distribution as Production/Stable. Beta package versions in this repository use Development Status :: 4 - Beta (for example, sdk/search/azure-search-documents/pyproject.toml:20), so PyPI metadata would otherwise misrepresent this release.
  sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:26
• This raises the minimum Python version from 3.7 to 3.10, but the new changelog entry does not disclose that 3.7–3.9 support is dropped, and README.md:21 still says Python 3.7 is supported. Update both user-facing documents so affected users are warned before upgrading.
  sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_provenance.py:14
• This class-level skip leaves both new provenance operations entirely untested; the existing package tests cover only Content Safety and Blocklist clients, while the async generated counterpart is also skipped. Add runnable sync and async coverage for starting, polling, and retrieving a provenance operation before releasing the new client.
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/types.py:201
• blocklistItemId is service-generated and read-only (the generated model also gives it visibility=["read"]), so requiring it in this request TypedDict rejects valid request dictionaries during type checking and led the generated sample to send a fabricated ID. Exclude the read-only field from the input shape and update the generated API artifacts/sample accordingly.
  sdk/contentsafety/azure-ai-contentsafety/CHANGELOG.md:12
• The release also exposes the new DetectOutcome enum, but this enum list omits it. Include it so the release notes accurately describe the provenance API surface.

- Added enums `DetectedProvenanceType` and `ProvenanceOperationKind`.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/models/_models.py:786

• The public model documentation uses the plural “properties” after the singular article “An.” Fix the source description to “An optional property…” and regenerate so the duplicate generated docstrings stay consistent.

[Copilot]

Pull request overview

Copilot reviewed 61 out of 62 changed files in this pull request and generated no new comments.

Suppressed comments (5)

sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:1

• [tool.setuptools.package-data] keys are package names, but pytyped is not a package. This will likely prevent py.typed from being included in the built wheel/sdist. Use the actual package name key (e.g., azure.ai.contentsafety) for the py.typed entry so typing markers are shipped correctly.
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/models/_models.py:1
• ODataV4Format is referenced in docstrings/types as azure.core.ODataV4Format, but it's imported from azure.core.exceptions. This is likely an import-path bug that can raise ImportError at runtime. Import ODataV4Format from the correct azure.core module/path that actually defines it (and keep the annotation consistent).
  sdk/contentsafety/azure-ai-contentsafety/tsp-location.yaml:1
• additionalDirectories: is now present without a value, which parses as null (not a list). Tooling that expects a list (previously []) can break. Set it explicitly to an empty list (additionalDirectories: []) or provide the intended list values.
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/types.py:1
• The TypedDict field is named blocklistItems, but the docstring documents blocklist_items. This mismatch makes the public typing/docs confusing. Update the docstring to document the actual key name(s) supported by the TypedDict (blocklistItems, and optionally any snake_case alias if intentionally supported).
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_validation.py:36
• api_version_validation always wraps with a sync def wrapper, even when decorating async def APIs. When stacked with @distributed_trace_async, this can break coroutine detection/async tracing behavior because the wrapped callable is no longer an iscoroutinefunction. Consider generating an async def wrapper when func is async (e.g., via inspect.iscoroutinefunction(func)) and await func(...) to preserve async semantics and tracing compatibility.

    def decorator(func):
        @functools.wraps(func)
        def wrapper(*args, **kwargs):
            try:
                # this assumes the client has an _api_version attribute
                client = args[0]
                client_api_version = client._config.api_version  # pylint: disable=protected-access

[Copilot]

Pull request overview

Copilot reviewed 61 out of 62 changed files in this pull request and generated no new comments.

Suppressed comments (3)

sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:16

• This release is explicitly beta and _version.py sets 1.1.0b1, but this classifier advertises the distribution as production/stable on package indexes. Use the Beta classifier, as other beta packages do (for example sdk/search/azure-search-documents/pyproject.toml:20 and sdk/identity/azure-identity/pyproject.toml:15).
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/types.py:201
• blocklistItemId is server-generated and read-only in the source contract, yet this request TypedDict makes it required. Consequently, the dictionary overload cannot type-check the valid create shape used by the existing model-based tests without supplying a read-only value, and plain dictionaries will serialize that value instead of filtering it. Generate a request shape that omits this read-only property (and refresh the derived API snapshot/samples).
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:120
• The maintained test suite covers the existing clients but has no executable coverage for this new provenance client or its LRO behavior. Add sync coverage for begin_detect/polling and get_operation_status (and async coverage where supported), including response deserialization and failure paths, before releasing the new API.

class ContentProvenanceClient(_ContentProvenanceClientOperationsMixin):

[Copilot]

Pull request overview

Copilot reviewed 63 out of 64 changed files in this pull request and generated no new comments.

Suppressed comments (9)

sdk/contentsafety/azure-ai-contentsafety/tsp-location.yaml:1

• additionalDirectories: is currently set to an empty/null value, which is likely invalid for the tooling that expects a list (previously []). Set it explicitly to an empty list (additionalDirectories: []) or remove the key entirely to avoid YAML parsing/consumer errors.
  sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:1
• The PEP 621 license field is expected to be a table (e.g., license = { text = \"MIT\" } or license = { file = \"LICENSE\" }). Using a bare string may be rejected by strict build backends/validators and can break packaging in some environments.
  sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:1
• [tool.setuptools.package-data] keys should be importable package names (or *), not an arbitrary label. As written, py.typed likely will not be included in the wheel for azure.ai.contentsafety, which breaks PEP 561 typing. Use a package key like azure.ai.contentsafety = [\"py.typed\"] (or equivalent) so the marker file is packaged.
  sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_safety.py:1
• All newly added generated tests are marked @pytest.mark.skip(...), so the new/updated API surface is not exercised in CI. If these tests are intended to validate this PR, remove the skip and replace placeholder inputs/assertions with real test cases (or add separate non-skipped tests covering the new operations).
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/types.py:1
• The docstring refers to blocklist_items, but the actual TypedDict key is blocklistItems. Similar snake_case vs camelCase mismatches appear elsewhere in types.py and can easily lead users to send the wrong JSON keys. Update the :ivar/:vartype docs to match the actual keys (or add explicit snake_case aliases + a conversion layer if snake_case is intended to be supported).
  sdk/contentsafety/azure-ai-contentsafety/CHANGELOG.md:5
• This entry reads like an internal instruction rather than user-facing release notes. Replace it with a short bullet list of customer-relevant changes (e.g., new APIs added: Content Provenance, Shield Prompt, protected material detection; packaging migration; etc.).

## 1.1.0b1 (2026-08-03)

skip changelog generation for data-plane package and please add changelog manually.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_operations/_operations.py:1274

• On error, the response body is not consumed/closed. In the Azure Core pipeline, not reading the body can prevent connection reuse and cause resource leaks under load. Align this with other methods in this file by reading the response on failures (ideally handling StreamConsumedError/StreamClosedError as done elsewhere).

        response = pipeline_response.http_response

        if response.status_code not in [204]:
            map_error(status_code=response.status_code, response=response, error_map=error_map)
            raise HttpResponseError(response=response)

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_operations/_operations.py:1743

• Same issue as delete_text_blocklist: on error the response body isn't consumed/closed, which can hurt connection pooling and reliability. Read the body before raising to ensure the transport can reuse the connection.

        response = pipeline_response.http_response

        if response.status_code not in [204]:
            map_error(status_code=response.status_code, response=response, error_map=error_map)
            raise HttpResponseError(response=response)

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_validation.py:38

• api_version_validation wraps methods with a synchronous def wrapper(...), even when decorating async def functions. While returning a coroutine still works, it breaks inspect.iscoroutinefunction(func) and can confuse tooling/middleware that relies on coroutine detection. Consider emitting an async def wrapper(...) when func is a coroutine function (and a sync wrapper otherwise).

    def decorator(func):
        @functools.wraps(func)
        def wrapper(*args, **kwargs):
            try:
                # this assumes the client has an _api_version attribute
                client = args[0]
                client_api_version = client._config.api_version  # pylint: disable=protected-access
            except AttributeError:
                return func(*args, **kwargs)

[Copilot]

Pull request overview

Copilot reviewed 61 out of 62 changed files in this pull request and generated no new comments.

Suppressed comments (7)

sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:20

• The release version is 1.1.0b1 and the PR is explicitly a beta release, but this classifier publishes it as Production/Stable. Beta packages use Development Status :: 4 - Beta (for example, sdk/cognitivelanguage/azure-ai-language-conversations/pyproject.toml:20 with version 2.0.0b2), so this metadata should be aligned to avoid misrepresenting the release on package indexes.
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_operations/_operations.py:812
• AsyncNoPolling uses this callback with the initial 202 response when polling=False. The service's initial payload has no result, so this returns an invalid DetectProvenanceResult missing required outcome; handle the initial status explicitly before deserializing the final nested result.

            deserialized = _deserialize(_models.DetectProvenanceResult, response.json().get("result", {}))

sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_provenance.py:13

• This skips the only added synchronous tests for ContentProvenanceClient, and the bodies contain no assertions. Convert these placeholders into executable recorded tests so the new initial 202 response, polling, final result deserialization, and status operation are covered.
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/types.py:124
• ImageData.content is a bytes property in the source TypeSpec and in the generated model, and the generated dictionary sample passes bytes, but this public TypedDict declares it as str. Type checkers will therefore reject the supported dictionary form of analyze_image; regenerate/customize this annotation as bytes and refresh the API snapshot.
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_operations/_operations.py:1146
• When polling=False, NoPolling invokes this callback on the initial 202 response. That response contains id and status but no result, so this produces an invalid DetectProvenanceResult with its required outcome missing. Handle the 202 initial response explicitly (normally returning no final result) before deserializing the nested result, and cover the no-polling path.

            deserialized = _deserialize(_models.DetectProvenanceResult, response.json().get("result", {}))

sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_provenance_async.py:14

• The async provenance test class is entirely skipped and has no assertions, leaving the new async LRO and status operation without executable coverage. Replace the generated placeholders with recorded tests, including the initial and final polling responses.
  sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_safety_async.py:14
• This class contains the only added tests for detect_text_protected_material and shield_prompt, but the class-level skip and placeholder assertions mean neither new operation is exercised. Add executable recorded coverage for their request serialization and response models before release.

[Copilot]

Pull request overview

Copilot reviewed 62 out of 63 changed files in this pull request and generated no new comments.

Suppressed comments (6)

sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:20

• The PR is a beta release (1.1.0b1), but this classifier publishes it as Production/Stable. This contradicts the release type and misrepresents the package on package indexes; use the Beta classifier.
  sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:30
• Raising the package requirement to Python 3.10 leaves the user-facing prerequisite stale: README.md:21 still says Python 3.7 or later. Update the README in the same release so installation guidance matches package metadata.
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/types.py:200
• blocklistItemId is documented as service-generated and the source contract marks it read-only, yet this request TypedDict requires callers to provide it. That makes the dictionary overload reject valid add requests during type checking; regenerate the request shape so this field is optional or omitted.
  sdk/contentsafety/azure-ai-contentsafety/CHANGELOG.md:8
• This description overstates the feature as determining whether media is AI-generated. The service only detects supported Microsoft-issued C2PA or watermark signals, and absence of a signal is not proof that content was not AI-generated; use the narrower wording from the service contract.

- Added `ContentProvenanceClient`, which detects whether media was generated or modified by an AI system:

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/init.py:16

• The new ContentProvenanceClient LRO workflow has no handwritten recorded or unit coverage in tests/; the existing suite covers only ContentSafetyClient and BlocklistClient. Add sync and async tests for the initial 202 response, successful polling/result deserialization, failure handling, and get_operation_status.

from ._client import ContentProvenanceClient  # type: ignore

sdk/contentsafety/azure-ai-contentsafety/api.md:276

• The handwritten tests/test_content_safety.py covers the existing analyze operations, but neither of the newly exposed detect_text_protected_material and shield_prompt methods has executable coverage. Add recorded sync/async tests that validate request serialization and response models for both operations.

        def detect_text_protected_material(

[Copilot]

Pull request overview

Copilot reviewed 62 out of 63 changed files in this pull request and generated no new comments.

Suppressed comments (8)

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:133

• The documented default is incorrect: ContentProvenanceClientConfiguration sets api_version to "2026-07-01-preview" when omitted, not None. Update the public client docstring to match the constructor behavior.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:216

• The documented default is incorrect: BlocklistClientConfiguration sets api_version to "2026-07-01-preview" when omitted, not None. Update the public client docstring to match the constructor behavior.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py:137

• The documented default is incorrect: the async ContentProvenanceClientConfiguration sets api_version to "2026-07-01-preview" when omitted, not None. Update the public async client docstring to match the constructor behavior.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py:224

• The documented default is incorrect: the async BlocklistClientConfiguration sets api_version to "2026-07-01-preview" when omitted, not None. Update the public async client docstring to match the constructor behavior.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:20

• This beta release is classified as Production/Stable, which makes the published package metadata contradict both VERSION = "1.1.0b1" and the PR's beta release type. Preview data-plane packages use the Beta classifier (for example, sdk/contentunderstanding/azure-ai-contentunderstanding/pyproject.toml:20).
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:52
• The documented default is incorrect: ContentSafetyClientConfiguration sets api_version to "2026-07-01-preview" when omitted, not None. Update the public client docstring so generated reference documentation describes the request version users actually receive.

This issue also appears in the following locations of the same file:

• line 130
• line 213

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py:52

• The documented default is incorrect: the async ContentSafetyClientConfiguration sets api_version to "2026-07-01-preview" when omitted, not None. Update the public async client docstring to describe the actual default.

This issue also appears in the following locations of the same file:

• line 134
• line 221

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:120

• The new provenance client and its LRO/status paths have no executable coverage: the only generated sync and async provenance tests are class-level skipped, while the maintained tests/ suite contains no provenance tests. Add runnable sync and async tests that validate request serialization, polling to the final DetectProvenanceResult, and status/error deserialization before exposing this client.

class ContentProvenanceClient(_ContentProvenanceClientOperationsMixin):

[Copilot]

Pull request overview

Copilot reviewed 62 out of 63 changed files in this pull request and generated no new comments.

Suppressed comments (15)

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:133

• This documents None as supported and as the default, but the configuration defaults to "2026-07-01-preview"; explicitly passing None also fails when the required query parameter is serialized. Document the actual default instead.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py:137

• This documents None as supported and as the default, but the async configuration defaults to "2026-07-01-preview"; explicitly passing None fails when serializing the required query parameter. Document the actual default instead.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py:224

• This documents None as supported and as the default, but the async configuration defaults to "2026-07-01-preview"; explicitly passing None fails when serializing the required query parameter. Document the actual default instead.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:216

• This documents None as supported and as the default, but the configuration defaults to "2026-07-01-preview"; explicitly passing None also fails when the required query parameter is serialized. Document the actual default instead.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/pyproject.toml:30

• The package now requires Python 3.10, but README.md:21 still tells users that Python 3.7 is supported. Update the README prerequisite so installation guidance matches the package metadata.
  sdk/contentsafety/azure-ai-contentsafety/CHANGELOG.md:12
• This list omits the newly exported DetectOutcome and OperationState enums, so the release notes do not describe the full new public enum surface shown in models/init.py and api.md.

- Added enums `DetectedProvenanceType` and `ProvenanceOperationKind`.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/types.py:201

• blocklistItemId is service-generated and read-only (the TypeSpec marks it Lifecycle.Read, and the model constructor intentionally does not accept it), but this request TypedDict makes it required. Type checkers therefore reject valid add-item dictionaries unless callers supply a field that should not be sent. Regenerate this request type without the read-only key.
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_operations/_operations.py:1146
• The initial 202 response contains only id and status, with no result. Because NoPolling.resource() invokes this callback for polling=False, defaulting the missing field to {} returns an invalid DetectProvenanceResult whose required outcome is absent. Preserve a missing result as None instead of constructing an empty final model.

            deserialized = _deserialize(_models.DetectProvenanceResult, response.json().get("result", {}))

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_operations/_operations.py:812

• The initial 202 response contains only id and status, with no result. AsyncNoPolling invokes this callback for polling=False, so the {} fallback produces an invalid DetectProvenanceResult missing its required outcome. Preserve the absent result as None.

            deserialized = _deserialize(_models.DetectProvenanceResult, response.json().get("result", {}))

sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_provenance.py:13

• All new synchronous provenance tests are skipped and contain no assertions, while the maintained tests/ suite has no provenance coverage. Add executable tests for the initial 202 response, polling completion, status/error deserialization, and polling=False behavior before releasing the new LRO.
  sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_provenance_async.py:14
• The asynchronous provenance test class is entirely skipped and its methods have no assertions, leaving the new async client and poller behavior untested. Add executable async tests, including the initial 202 response and polling=False path.
  sdk/contentsafety/azure-ai-contentsafety/generated_tests/test_content_safety_async.py:14
• This class-level skip disables all generated coverage for shield_prompt and detect_text_protected_material, and there are no corresponding tests in the maintained tests/ directory. Add runnable sync and async tests with response assertions for both new operations.
  sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:52
• This documents None as supported and as the default, but the configuration defaults to "2026-07-01-preview"; explicitly passing None also fails when the required query parameter is serialized. Document the actual default instead.

This issue also appears in the following locations of the same file:

• line 130
• line 213

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py:52

• This documents None as supported and as the default, but the async configuration defaults to "2026-07-01-preview"; explicitly passing None fails when serializing the required query parameter. Document the actual default instead.

This issue also appears in the following locations of the same file:

• line 134
• line 221

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/init.py:32

• The new public ContentProvenanceClient is not mentioned in README.md's feature list, authentication examples, or examples section, which still documents only ContentSafetyClient and BlocklistClient. Add getting-started guidance for constructing this client and running begin_detect so the new API is discoverable and usable.

    "ContentSafetyClient",
    "ContentProvenanceClient",
    "BlocklistClient",

[Copilot]

Pull request overview

Copilot reviewed 64 out of 65 changed files in this pull request and generated no new comments.

Suppressed comments (8)

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:133

• The documented default is inconsistent with the implementation: ContentProvenanceClientConfiguration defaults this argument to "2026-07-01-preview", not None. Update the public docstring so users know which API version is sent when they omit the argument.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:216

• The documented default is inconsistent with the implementation: BlocklistClientConfiguration defaults this argument to "2026-07-01-preview", not None. Update the public docstring so users know which API version is sent when they omit the argument.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py:137

• The documented default is inconsistent with the async configuration, which defaults this argument to "2026-07-01-preview", not None. Update the public docstring so users know which API version is sent when they omit the argument.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py:224

• The documented default is inconsistent with the async configuration, which defaults this argument to "2026-07-01-preview", not None. Update the public docstring so users know which API version is sent when they omit the argument.

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_client.py:52

• The documented default is inconsistent with the implementation: ContentSafetyClientConfiguration defaults this argument to "2026-07-01-preview", not None. Update the public docstring so users know which API version is sent when they omit the argument.

This issue also appears in the following locations of the same file:

• line 130
• line 213

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_client.py:52

• The documented default is inconsistent with the async configuration, which defaults this argument to "2026-07-01-preview", not None. Update the public docstring so users know which API version is sent when they omit the argument.

This issue also appears in the following locations of the same file:

• line 134
• line 221

    :keyword api_version: The API version to use for this operation. Known values are
     "2026-07-01-preview" and None. Default value is None. If not set, the operation's default API
     version will be used. Note that overriding this default value may result in unsupported
     behavior.

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/_operations/_operations.py:1146

• The initial 202 response contains the operation envelope (id and status) but no result. Because NoPolling.resource() invokes this callback on that initial response, begin_detect(..., polling=False).result() currently returns an empty DetectProvenanceResult, which incorrectly looks like a final result. Return None when the envelope has no result, and deserialize only completed responses.

            deserialized = _deserialize(_models.DetectProvenanceResult, response.json().get("result", {}))

sdk/contentsafety/azure-ai-contentsafety/azure/ai/contentsafety/aio/_operations/_operations.py:812

• The initial 202 response contains the operation envelope (id and status) but no result. Because AsyncNoPolling.resource() invokes this callback on that initial response, awaiting begin_detect(..., polling=False).result() currently returns an empty DetectProvenanceResult, which incorrectly looks like a final result. Return None when the envelope has no result, and deserialize only completed responses.

            deserialized = _deserialize(_models.DetectProvenanceResult, response.json().get("result", {}))

[github-actions]

[Pilot] PR Pipeline Failure Analysis

A CI pipeline failed on this pull request. Here is an automated analysis of what went wrong and how to get the build green.

What failed

A single test — sdk.contentsafety.azure-ai-contentsafety.tests.test_content_provenance — failed consistently across all test matrix platforms and packaging modes (sdist, whl, mindependency):

• macOS (Python 3.11)
• Ubuntu 24.04 (Python 3.10 coverage, 3.10, 3.13, 3.14)
• Windows 2022 (Python 3.12)

The failure is reproducible across every OS and Python version, which strongly suggests a code or test issue rather than infrastructure flakiness.

Recommended next steps

• Investigate the test_content_provenance test in sdk/contentsafety/azure-ai-contentsafety/tests/ — review what the test asserts and verify that the new SDK generation produces the expected content provenance behavior/API surface.
• Run the test locally: pytest sdk/contentsafety/azure-ai-contentsafety/tests/test_content_provenance.py -v to see the full failure message.
• If the test expectations are outdated relative to the new API spec, update the test recordings or test assertions accordingly.
• See the CI troubleshooting guide: https://aka.ms/ci-fix
• Push new commits to address the failures; this comment updates automatically on the next failing run.

Raw pipeline analysis (azsdk ci analyze)

Analyzing pipeline https://github.com/Azure/azure-sdk-for-python/pull/48375...
--------------------------------------------------------------------------------
Failed Tests
--------------------------------------------------------------------------------
{
  "LLM Artifacts - macos311 - 1/contentsafety-azure-ai-contentsafety-test-junit-sdist.xml": [
    "sdk.contentsafety.azure-ai-contentsafety.tests.test_content_provenance"
  ],
  "LLM Artifacts - macos311 - 1/contentsafety-azure-ai-contentsafety-test-junit-whl.xml": [
    "sdk.contentsafety.azure-ai-contentsafety.tests.test_content_provenance"
  ],
  "LLM Artifacts - macos311 - 1/contentsafety-azure-ai-contentsafety-test-junit-mindependency.xml": [
    "sdk.contentsafety.azure-ai-contentsafety.tests.test_content_provenance"
  ],
  "LLM Artifacts - ubuntu2404_310_coverage - 1/contentsafety-azure-ai-contentsafety-test-junit-sdist.xml": [
    "sdk.contentsafety.azure-ai-contentsafety.tests.test_content_provenance"
  ],
  "LLM Artifacts - ubuntu2404_310 - 1/contentsafety-azure-ai-contentsafety-test-junit-sdist.xml": [
    "sdk.contentsafety.azure-ai-contentsafety.tests.test_content_provenance"
  ],
  "LLM Artifacts - Ubuntu2404_314 - 1/contentsafety-azure-ai-contentsafety-test-junit-sdist.xml": [
    "sdk.contentsafety.azure-ai-contentsafety.tests.test_content_provenance"
  ],
  "LLM Artifacts - Ubuntu2404_313 - 1/contentsafety-azure-ai-contentsafety-test-junit-sdist.xml": [
    "sdk.contentsafety.azure-ai-contentsafety.tests.test_content_provenance"
  ],
  "LLM Artifacts - windows2022_312 - 1/contentsafety-azure-ai-contentsafety-test-junit-sdist.xml": [
    "sdk.contentsafety.azure-ai-contentsafety.tests.test_content_provenance"
  ]
  // ... (same test failing across all whl and mindependency variants)
}

Copilot detected the failing pipeline and generated the analysis above. To have it attempt a fix automatically, reply with @copilot please fix the failing pipeline on this PR.

Generated by Pipeline Analysis - Next Steps · 20.7 AIC · ⌖ 8.84 AIC · ⊞ 6.6K · ◷