feat: custom worker count for IBM models in OpenSearch

[edwinjosechittilappilly]

This pull request introduces a targeted update to the embedding concurrency logic in opensearch_multimodal.py. The main change is the restriction of parallel embedding requests for IBM/Watsonx models to avoid hitting rate limits.

Concurrency and rate limit handling:

• In embed_chunk, the max_workers value is set to 1 for IBM/Watsonx models, ensuring only one embedding request is processed at a time and thus reducing the risk of rate limit errors. For other models, the previous logic of up to 8 workers is retained.

Summary by CodeRabbit

• Bug Fixes
  • Improved handling of IBM Watsonx embedding operations by optimizing concurrency settings to prevent rate-limiting issues. Non-IBM embedding models continue to use parallel processing for improved performance.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

Important

Review skipped

Auto incremental reviews are disabled on this repository.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Note

Other AI code review bot(s) detected

CodeRabbit has detected other AI code review bot(s) in this pull request and will avoid duplicating their findings in the review comments. This may lead to a less comprehensive review.

Walkthrough

Adds conditional concurrency control for embeddings by detecting IBM Watsonx models and restricting parallelism to a single worker when identified. Otherwise, standard parallelism calculation min(max(len(texts), 1), 8) is used.

Changes

Cohort / File(s)	Change Summary
Embedding Concurrency Control src/lfx/src/lfx/components/elastic/opensearch_multimodal.py	Adds model detection logic in embedding operations to conditionally set max_workers. When embedding_model name contains "ibm" or selected_embedding type name contains "watsonx", max_workers is restricted to 1. Otherwise, standard parallelism formula is applied.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

• Review the conditional logic for model name detection (IBM/Watsonx strings)
• Verify that max_workers assignment doesn't inadvertently affect other embedding models
• Confirm the impact on the threading/retry loop behavior remains as intended

Possibly related PRs

• fix: Add IBM watsonx.ai support to EmbeddingModel #10677: Adds IBM Watsonx embedding support with model name detection—directly related to the concurrency control change that targets these newly added models.
• feat: Add OpenSearch multimodal multi-embedding component #10714: Introduces threaded embedding support in the same file—the current PR modifies the concurrency behavior of that threaded embedding mechanism for specific model types.

Suggested labels

enhancement

Suggested reviewers

• phact

Pre-merge checks and finishing touches

Important

Pre-merge checks failed

Please resolve all errors before merging. Addressing warnings is optional.

❌ Failed checks (1 warning, 1 inconclusive)

Check name	Status	Explanation	Resolution
Test Quality And Coverage	⚠️ Warning	IBM/Watsonx concurrency control implementation (lines 881-885) lacks any corresponding tests in the test suite.	Add test cases validating IBM/Watsonx detection logic, max_workers=1 assignment, and ThreadPoolExecutor behavior with single-worker concurrency.
Test Coverage For New Implementations	❓ Inconclusive	Cannot verify test coverage; the modified opensearch_multimodal.py file and test files are not accessible in the current repository context.	Provide access to the PR changes with the modified source file and corresponding test files to verify test coverage for the max_workers concurrency control implementation.

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately reflects the main change: introducing custom worker count configuration for IBM models in OpenSearch embedding operations.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Test File Naming And Structure	✅ Passed	PR includes comprehensive test file (test_opensearch_multimodal.py) with robust pytest structure, descriptive test functions, proper fixtures, and coverage of unit, integration, positive and negative test cases.
Excessive Mock Usage Warning	✅ Passed	PR only modifies source code for concurrency control with no test file changes. Existing tests show minimal mock usage (0.03 @patch per test, 10 total calls across 31 tests) with appropriate fixture reuse and real component testing.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[Copilot]

Pull request overview

This PR introduces concurrency control for IBM/Watsonx embedding models to mitigate rate limiting issues during document ingestion in OpenSearch. The change restricts parallel embedding requests to a single worker for IBM models while maintaining the existing multi-threaded approach (up to 8 workers) for other embedding providers.

Key Changes:

• Added detection logic to identify IBM/Watsonx models based on model name or embedding class name
• Set max_workers=1 for IBM/Watsonx models to prevent rate limit errors during parallel embedding generation
• Preserved existing concurrency logic for non-IBM models

---

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

[Copilot]

The detection logic for IBM/Watsonx models relies on substring matching which could be fragile. Consider checking the exact class type instead:

from langchain_ibm import WatsonxEmbeddings

is_ibm = isinstance(selected_embedding, WatsonxEmbeddings)

This would be more robust than checking for "watsonx" in the class name, as it would work correctly even if the class is renamed or extended. If you need to support multiple IBM embedding classes, you could check against a tuple of types.

[edwinjosechittilappilly]

could be a future use case but currently only for ingestion hence selected embeddings is fine

[Copilot]

The new concurrency restriction logic for IBM/Watsonx models lacks test coverage. Consider adding tests to verify:

1. When an IBM/Watsonx embedding is used, max_workers is set to 1
2. When a non-IBM embedding is used, max_workers follows the original logic (min(max(len(texts), 1), 8))
3. The detection logic correctly identifies IBM models by checking the embedding_model string
4. The detection logic correctly identifies Watsonx models by checking the class name

This would help prevent regressions if the detection logic or concurrency settings change in the future.

[coderabbitai]

Actionable comments posted: 0

🧹 Nitpick comments (1)

src/lfx/src/lfx/components/elastic/opensearch_multimodal.py (1)

878-887: IBM/Watsonx concurrency guard looks correct; consider small readability tweaks

The is_ibm check and max_workers = 1 branch achieve the PR goal without changing behavior for other providers. A couple of minor, optional nits:

• is_ibm and max_workers don’t depend on the retry attempt; you could compute them once before the while loop to avoid repeating the same logic and make intent clearer.
• If you expect more IBM/Watsonx-specific handling later, extracting the detection into a tiny helper (e.g. _is_ibm_embedding(embedding_model, selected_embedding)) would centralize the heuristic and make it easier to extend (e.g. other IBM model name patterns).

No functional changes are required here; these would just tighten clarity and future maintainability.

📜 Review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 17af7fb and 7e36fce.

📒 Files selected for processing (1)

• src/lfx/src/lfx/components/elastic/opensearch_multimodal.py (1 hunks)

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (4)

• GitHub Check: CodeQL analysis (python)
• GitHub Check: Agent
• GitHub Check: Update Component Index
• GitHub Check: Update Starter Projects

[github-actions]

Frontend Unit Test Coverage Report

Coverage Summary

Lines	Statements	Branches	Functions
	15.43% (4244/27499)	8.61% (1811/21013)	9.69% (587/6057)

Unit Test Results

Tests	Skipped	Failures	Errors	Time
1671	0 💤	0 ❌	0 🔥	21.413s ⏱️

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 32.54%. Comparing base (17af7fb) to head (3a3a17c).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #10866      +/-   ##
==========================================
- Coverage   32.82%   32.54%   -0.29%     
==========================================
  Files        1367     1370       +3     
  Lines       64017    63518     -499     
  Branches     9553     9391     -162     
==========================================
- Hits        21014    20672     -342     
+ Misses      41943    41806     -137     
+ Partials     1060     1040      -20     

Flag	Coverage Δ
backend	51.53% <ø> (-0.42%)	⬇️
frontend	14.28% <ø> (+0.15%)	⬆️
lfx	39.99% <ø> (-0.29%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.
see 34 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.