Open
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #475 +/- ##
==========================================
+ Coverage 82.61% 82.71% +0.09%
==========================================
Files 96 96
Lines 9147 9159 +12
==========================================
+ Hits 7557 7576 +19
+ Misses 1590 1583 -7 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
| ) | ||
|
|
||
|
|
||
| def _set_tool_span_attributes(span: trace.Span, galileo_span: ToolSpan) -> None: |
Contributor
There was a problem hiding this comment.
Please add:
gen_ai.tool.name
gen_ai.tool.description
gen_ai.tool.call.arguments
gen_ai.tool.call.result
Reference: https://opentelemetry.io/docs/specs/semconv/gen-ai/gen-ai-spans/#execute-tool-span
Contributor
There was a problem hiding this comment.
Commit 3f3a0f9 addressed this comment by adding several of the requested GenAI tool span semantic convention attributes in _set_tool_span_attributes, including gen_ai.tool.name, gen_ai.tool.call.arguments, and gen_ai.tool.call.result, and by updating the tool call id attribute to gen_ai.tool.call.id.
This change adds support for setting additional attributes on OpenTelemetry spans when processing ToolSpan objects from the galileo-core library. Specifically, the following attributes are now set on the span: - gen_ai.input.messages: JSON array with the tool's input data - gen_ai.output.messages: JSON array with the tool's output data (if available) - gen_ai.tool.call_id: the unique identifier for the tool call (if available) This provides more detailed observability into the usage of AI/ML tools within the Galileo platform.
This change adds two new test cases for the start_galileo_span context manager: 1. Test that start_galileo_span correctly sets the expected attributes on the generated OpenTelemetry span when provided a fully populated ToolSpan. 2. Test that start_galileo_span handles a ToolSpan with a None output and tool_call_id gracefully. The tests ensure the context manager is properly dispatching the ToolSpan attributes to the underlying OpenTelemetry span, which is crucial for providing meaningful tracing data.
- Renames `gen_ai.tool.call_id` to `gen_ai.tool.call.id` for consistency - Adds `gen_ai.tool.call.arguments` and `gen_ai.tool.call.result` attributes - Updates tests to cover the new attribute set
This change adds the 'gen_ai.operation.name' attribute to the OpenTelemetry spans created for tool execution. The operation name is set to 'execute_tool' to provide more context about the purpose of the span.
calebe94
force-pushed
the
feature/54704-support-tool-spans
branch
from
3f3a0f9 to
94a180c
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Shortcut:
Description:
This change adds support for setting additional attributes on OpenTelemetry spans when processing ToolSpan objects from the galileo-core library.
Specifically, the following attributes are now set on the span:
This provides more detailed observability into the usage of AI/ML tools within the Galileo platform.
The following
ToolSpanfields are now mapped to OpenTelemetry semantic convention attributes:inputgen_ai.input.messagesoutputgen_ai.output.messagestool_call_idgen_ai.tool.call_idAttribute Format Details:
gen_ai.input.messages(always set):[{"role": "tool", "content": "<input_string>"}]"tool"(identifies this as a tool/function call)ToolSpan.inputgen_ai.output.messages(conditional):[{"role": "tool", "content": "<output_string>"}]ToolSpan.outputis not None"tool"indicates this is tool outputToolSpan.outputgen_ai.tool.call_id(conditional):ToolSpan.tool_call_idis not NoneTests:
Evidences
I've created a demo script called tool_span_demo.py to test this changes. To run this script we need to export a couple of credentials:
Inside the galileo-python directory I've started a new
virtualenvironmentso we can install the project locally:Then I installed the dependencies:
After all this, I've ran the
tool_span_demo.pyPython script and got this as a result:What it demonstrates:
✅ All
ToolSpanfields correctly map to OTel semantic conventions✅ Manual span creation works via
start_galileo_span()✅ Conditional attribute setting for None values
✅ Comprehensive demo script provided with 5 test cases
✅ Unit tests covering all scenarios (3 test cases, all passing)
✅ Documentation and troubleshooting guide created
✅ OpenTelemetry semantic compliance verified