Improving ConnectorClient and UserTokenClient error handling

[Rodrigo Brandão (rodrigobr-msft)]

This pull request refactors error handling for HTTP responses in the Microsoft Agents Hosting Core connector clients. It centralizes and standardizes response error handling by introducing a shared utility function, and updates all relevant client methods to use it. Additionally, it updates data serialization for attachment and transcript uploads to use model-based serialization, and removes a redundant class.

Error handling improvements:

• Added a new utility function _handle_request_error in _utils.py to centralize and standardize HTTP response error handling for all connector client methods.
• Updated all client methods in agent_sign_in.py and connector_client.py to replace ad-hoc error handling and logging with calls to _handle_request_error, ensuring consistent error reporting and exception raising across the codebase. [1] [2] Fe969ef8L89, Fe969ef8L129, Fe969ef8L216, Fe969ef8L239, Fe969ef8L286, Fe969ef8L342, Fe969ef8L388, Fe969ef8L426, Fe969ef8L455, Fe969ef8L505, Fe969ef8L546, Fe969ef8L582, Fe969ef8L616, Fe969ef8L664, Fe969ef8L695)

Data serialization and model usage:

• Changed attachment and transcript upload methods to use .model_dump() for serializing AttachmentData and Transcript objects, ensuring proper field mapping and cleaner code. (Fe969ef8L455, Fe969ef8L695)
• Removed the redundant local AttachmentData class from connector_client.py in favor of using the imported model. (Fe969ef8L43)

Other updates:

• Improved handling of HTTP 301/302 redirects in get_attachment to log a warning and return an empty BytesIO object instead of raising an error. (Fe969ef8L129)
• Updated method signatures to use the correct model types, such as changing the body parameter of send_conversation_history to use the Transcript model. (Fe969ef8L664)

These changes improve code maintainability, consistency, and correctness in handling HTTP responses and data serialization.

[Copilot]

Pull request overview

This PR refactors the Microsoft Agents Hosting Core connector clients to centralize HTTP error handling via a shared _handle_request_error helper, and updates several connector operations to use consistent model-based serialization (notably for attachments and transcripts).

Changes:

• Added a shared _handle_request_error utility and updated connector client methods to route non-success HTTP responses through it.
• Switched attachment upload and transcript/history upload payload serialization to model-based .model_dump(...).
• Expanded tests to assert externally observable REST contract behavior (requests made, typed responses, and error raising).

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tests/hosting_core/connector/test_user_token_client.py	Adds contract-style tests for token operations and sign-in state encoding.
tests/hosting_core/connector/test_connector_client.py	Adds contract-style tests for conversation lifecycle, members, attachments, and unexpected statuses.
libraries/microsoft-agents-hosting-core/microsoft_agents/hosting/core/connector/mcs/mcs_connector_client.py	Routes MCS connector error handling through the shared helper.
libraries/microsoft-agents-hosting-core/microsoft_agents/hosting/core/connector/client/user_token.py	Replaces ad-hoc response checks with _handle_request_error and standardizes 404 behavior for missing token.
libraries/microsoft-agents-hosting-core/microsoft_agents/hosting/core/connector/client/connector_client.py	Centralizes response error handling and switches attachment/transcript request bodies to .model_dump(...).
libraries/microsoft-agents-hosting-core/microsoft_agents/hosting/core/connector/client/agent_sign_in.py	Uses _handle_request_error for non-200 responses.
libraries/microsoft-agents-hosting-core/microsoft_agents/hosting/core/connector/_utils.py	Introduces _handle_request_error helper for consistent error raising/logging.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.