feat(gooddata-sdk): [AUTO] Add ToolCallEventResult schema and toolCallEvents to chat response

[yenkins-admin]

Summary

Added ToolCallEventResult SDK wrapper class for the new ToolCallEventResult OpenAPI schema and the toolCallEvents array field added to ChatResult. The generated API client already contained both changes; the SDK wrapper was added in compute/model/chat.py and exported from gooddata_sdk/__init__.py. A unit test covering construction, from_dict/to_dict round-trip, and client_class() was added in tests/compute/test_chat_models.py.

Impact: new_feature | Services: gooddata-afm-client
Tickets: GDAI-1438

Files changed

• packages/gooddata-sdk/src/gooddata_sdk/compute/model/chat.py
• packages/gooddata-sdk/src/gooddata_sdk/__init__.py
• packages/gooddata-sdk/tests/compute/test_chat_models.py

Agent decisions

Decisions (3)

Class placement — Placed ToolCallEventResult in compute/model/chat.py (compute module, no Catalog prefix)

• Alternatives: catalog//entity_model/ with Catalog prefix, Inline in compute/service.py
• Why: The AI chat feature lives in compute/service.py. Existing compute model classes (Attribute, Metric, Filter) use no Catalog prefix. Placing it here keeps the domain consistent.

Naming to avoid collision with generated client — Import generated class as ApiToolCallEventResult alias, define SDK class as ToolCallEventResult

• Alternatives: CatalogToolCallEventResult (Catalog prefix would be inconsistent with compute module), Re-export the generated class directly (no SDK wrapper)
• Why: Gives users a clean SDK name (ToolCallEventResult) while avoiding confusion; the alias makes the generated-vs-SDK distinction explicit inside the module.

Wrapper vs. pass-through — Created a proper attrs wrapper class extending Base

• Alternatives: Re-export generated ToolCallEventResult directly (no from_dict/to_dict), No wrapper at all
• Why: Gives users from_dict/to_dict/to_api serialization support and a stable SDK type for type annotations. Other response types (ChatResult, SearchResult) are returned raw from service methods, but ToolCallEventResult is a new leaf schema that benefits from SDK ownership.

Assumptions to verify (3)

• The generated API client (gooddata-api-client) already includes ToolCallEventResult and ChatResult.tool_call_events — confirmed by reading the files directly.
• The Base.from_dict / to_dict serialization path works for ToolCallEventResult because the generated ApiToolCallEventResult has from_dict (it inherits from ModelNormal). This was not executable-verified due to missing urllib3 in the test environment.
• All three fields (function_arguments, function_name, result) are non-optional (required in the OpenAPI spec), so they are declared as required positional attrs fields.

Risks (1)

• The Base.to_dict() path calls self.client_class().from_dict(snake_case_dict, camel_case=False). If the generated ApiToolCallEventResult.from_dict does not accept camel_case=False cleanly for all three fields, the round-trip test will fail.

Layers touched (3)

• entity_model — New ToolCallEventResult wrapper class in compute model layer
  • packages/gooddata-sdk/src/gooddata_sdk/compute/model/chat.py
• public_api — Exported ToolCallEventResult from the public SDK surface
  • packages/gooddata-sdk/src/gooddata_sdk/__init__.py
• tests — Unit tests for ToolCallEventResult construction, round-trip serialization, and client_class()
  • packages/gooddata-sdk/tests/compute/test_chat_models.py

Source commits (gdc-nas)

• 0e847c5 Merge pull request #21287 from gooddata/zovi/GDAI-1438-e2e-skill-tests

OpenAPI diff

diff --git a/microservices/afm-exec-api/src/test/resources/openapi/open-api-spec.json
@@ -1208,6 +1208,13 @@
           "threadIdSuffix" : { "description" : "Chat History thread suffix.", "type" : "string" },
+          "toolCallEvents" : {
+            "description" : "Tool call events emitted during the agentic loop (only present when GEN_AI_YIELD_TOOL_CALL_EVENTS is enabled).",
+            "items" : { "$ref" : "#/components/schemas/ToolCallEventResult" },
+            "type" : "array"
+          }
@@ -4485,6 +4492,25 @@
+      "ToolCallEventResult" : {
+        "description" : "Tool call events emitted during the agentic loop.",
+        "properties" : {
+          "functionArguments" : { "description" : "JSON-encoded arguments passed to the tool function.", "type" : "string" },
+          "functionName" : { "description" : "Name of the tool function that was called.", "type" : "string" },
+          "result" : { "description" : "Result returned by the tool function.", "type" : "string" }
+        },
+        "required" : [ "functionArguments", "functionName", "result" ],
+        "type" : "object"
+      },

Workflow run

---

Generated by SDK OpenAPI Sync workflow

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.68%. Comparing base (37d0593) to head (3cdaea7).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1556      +/-   ##
==========================================
+ Coverage   78.66%   78.68%   +0.02%     
==========================================
  Files         230      231       +1     
  Lines       15405    15421      +16     
==========================================
+ Hits        12118    12134      +16     
  Misses       3287     3287              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.