Skip to content

[AutoPR azure-ai-projects]-generated-from-SDK Generation - Python-6070083#45950

Draft
azure-sdk wants to merge 44 commits intomainfrom
dargilco/test-sdk-generation-pipeline-2
Draft

[AutoPR azure-ai-projects]-generated-from-SDK Generation - Python-6070083#45950
azure-sdk wants to merge 44 commits intomainfrom
dargilco/test-sdk-generation-pipeline-2

Conversation

@azure-sdk
Copy link
Collaborator

Configurations: 'specification/ai-foundry/data-plane/Foundry/tspconfig.yaml', API Version: v1, SDK Release Type: stable, and CommitSHA: 'bd70f1fbb6f1691a3107d791c7aa5d1dede46f01' in SpecRepo: 'https://github.com/Azure/azure-rest-api-specs' Pipeline run: https://dev.azure.com/azure-sdk/internal/_build/results?buildId=6070083 Refer to https://eng.ms/docs/products/azure-developer-experience/develop/sdk-release/sdk-release-prerequisites to prepare for SDK release.

dargilco and others added 30 commits March 6, 2026 10:29
…#45611)

* marking finetuning pause and resume operations as live extended tests

* updating recording

---------

Co-authored-by: Jayesh Tanna <jatanna@microsoft.com>
* rename env vars

* rename env var

* resolved comments

* remove chat completion

* resolved comment
* Add CSV and synthetic data generation evaluation samples

Add two new evaluation samples under sdk/ai/azure-ai-projects/samples/evaluations/:

- sample_evaluations_builtin_with_csv.py: Demonstrates evaluating pre-computed
  responses from a CSV file using the csv data source type. Uploads a CSV file
  via the datasets API, runs coherence/violence/f1 evaluators, and polls results.

- sample_synthetic_data_evaluation.py: Demonstrates synthetic data evaluation
  (preview) that generates test queries from a prompt, sends them to a model
  target, and evaluates responses with coherence/violence evaluators.

Also adds:
- data_folder/sample_data_evaluation.csv: Sample CSV data file with 3 rows
- README.md: Updated sample index with both new samples

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Update synthetic eval sample: agent target + cleaner dataset ID retrieval

- Switch from model target to agent target (azure_ai_agent)
- Create agent version via agents.create_version() before evaluation
- Simplify output_dataset_id retrieval using getattr instead of nested hasattr/isinstance checks
- Add AZURE_AI_AGENT_NAME env var requirement
- Remove input_messages (not needed for agent target)

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Add model target synthetic eval sample, cross-reference both

- Add sample_synthetic_data_model_evaluation.py for model target with
  input_messages system prompt
- Update sample_synthetic_data_evaluation.py docstring with cross-reference
- Update README.md with both synthetic samples (agent and model)

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Rename synthetic agent sample, clarify README, add prompt/files comments

- Rename sample_synthetic_data_evaluation.py to sample_synthetic_data_agent_evaluation.py
- Clarify README: JSONL dataset vs CSV dataset descriptions
- Remove (preview) from synthetic sample descriptions in README
- Add comments about prompt and reference_files options in both synthetic samples

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Skip new eval samples in recording tests

Add sample_evaluations_builtin_with_csv.py, sample_synthetic_data_agent_evaluation.py,
and sample_synthetic_data_model_evaluation.py to samples_to_skip list since they
require file upload prerequisites or are long-running preview features.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Rename env vars per PR review: FOUNDRY_PROJECT_ENDPOINT, FOUNDRY_MODEL_NAME

Address review comments from howieleung:
- AZURE_AI_PROJECT_ENDPOINT -> FOUNDRY_PROJECT_ENDPOINT
- AZURE_AI_MODEL_DEPLOYMENT_NAME -> FOUNDRY_MODEL_NAME
Updated in all 3 new samples and README.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Rename AZURE_AI_AGENT_NAME to FOUNDRY_AGENT_NAME per review

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Update changelog with new sample entries

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* New samples

* resolved comments
* LLM validation use 5.2 and chat completion

* change log

* Resolved comments
* Adding-Upload-Evaluator

* Adding-Upload-Evaluator

* Adding-Upload-Evaluator

* Adding-Upload-Evaluator-aio

* rename

* added - eval and eval run

* fix

* adding tests

* updated as per review
howieleung and others added 14 commits March 23, 2026 11:59
* added toolset samples and more

* improved samples

* resolved comment
…lity gates (#45854)

Additional Pyright and Pylint errors will need to be fixed in another PR.
* Fix pyright and pylint errors in evaluator upload operations

- Replace hasattr(e, 'error_code') with getattr(e, 'error_code', None) to
  fix pyright reportAttributeAccessIssue on HttpResponseError
- Narrow union types with isinstance checks for EvaluatorVersion and
  CodeBasedEvaluatorDefinition to fix pyright type errors on definition/blob_uri
- Extract _upload_folder_to_blob() and _set_blob_uri() helpers to fix pylint
  R0914 (too many local variables) in upload()
- Update tests to use MagicMock(spec=CodeBasedEvaluatorDefinition) for
  definition mocks

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Add missing docstring params, return, and rtype for Azure SDK lint rules

Fix C4739/C4741/C4742 violations from azure-pylint-guidelines-checker:
- _set_blob_uri: add :param and :type for evaluator_version, blob_uri
- _start_pending_upload_and_get_container_client: add :param/:type/:return/:rtype
- _get_next_version: add :param/:type/:return/:rtype

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Add missing param docs for _upload_folder_to_blob (C4739)

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…uild of package @azure-tools/typespec-client-generator-core) (#45887)
…tspconfig.yaml', API Version: v1, SDK Release Type: stable, and CommitSHA: '57e7a9ebf3c779e4f8c90b9e140bf09c14aa0830' in SpecRepo: 'https://github.com/Azure/azure-rest-api-specs' Pipeline run: https://dev.azure.com/azure-sdk/internal/_build/results?buildId=6069658 Refer to https://eng.ms/docs/products/azure-developer-experience/develop/sdk-release/sdk-release-prerequisites to prepare for SDK release."

This reverts commit 39fb228.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

7 participants