Fix AWS Bedrock ARN parsing issue in converse methods

[devin-ai-integration]

Description

Fixes AWS Bedrock instrumentation crash when using ARN or cross-region model IDs that contain multiple dots. The issue occurred in patch_converse_stream and patch_converse functions where modelId.split(".") assumed exactly 2 parts, but ARNs and cross-region IDs have multiple dots, causing ValueError: too many values to unpack (expected 2).

Problem:

# This fails with ARNs like: arn:aws:bedrock:us-east-1:123456789012:inference-profile/us.anthropic.claude-3-haiku-20240307-v1:0
(vendor, _) = modelId.split(".")  # ValueError: too many values to unpack

Solution:
Replaced the problematic split calls with the existing robust parse_vendor_and_model_name_from_model_id() function that already handles ARNs correctly and is used consistently in other methods like patch_invoke_model.

Changes Made

• src/langtrace_python_sdk/instrumentation/aws_bedrock/patch.py:
  • Line 110: Replace (vendor, _) = modelId.split(".") with vendor, _ = parse_vendor_and_model_name_from_model_id(modelId) in patch_converse_stream
  • Line 145: Same replacement in patch_converse function

Key Review Points

⚠️ Critical for reviewer to verify:

1. Test with actual ARN formats: The existing test suite covers the parser function, but please test the full instrumentation pipeline with:

   • arn:aws:bedrock:us-east-1:<account_id>:inference-profile/us.anthropic.claude-3-haiku-20240307-v1:0
   • us.anthropic.claude-sonnet-4-20250514-v1:0

2. Behavioral consistency: Confirm that parse_vendor_and_model_name_from_model_id() returns compatible values for all model ID formats that the original split logic handled

3. Run existing test suite: I couldn't run the full test suite locally due to missing dependencies, so please verify no regressions with:

   python -m pytest src/tests/aws_bedrock/ -v

Testing Evidence

Created test script that reproduces the original error:

Testing model_id: arn:aws:bedrock:us-east-1:123456789012:inference-profile/us.anthropic.claude-3-haiku-20240307-v1:0
  ERROR with original method: too many values to unpack (expected 2)

Testing model_id: us.anthropic.claude-sonnet-4-20250514-v1:0  
  ERROR with original method: too many values to unpack (expected 2)

Checklist for bug fix:

• Identified root cause of the ValueError
• Used existing robust parser instead of creating new logic
• Made changes consistent with other methods in the same file
• Preserved original variable naming and unpacking pattern
• ⚠️ Needs reviewer: Full test suite execution with dependencies
• ⚠️ Needs reviewer: End-to-end testing with actual AWS Bedrock calls

---

Link to Devin run: https://app.devin.ai/sessions/9b326f6380504040bf8b711ce512ab94
Requested by: karthik@scale3labs.com

Risk Assessment: Low-risk change (2 lines) but affects core functionality. The existing parser is well-tested, but human verification of compatibility is essential since I couldn't run the full test suite locally.

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring