test_csv_output_parser_openai

[Sqiyuan]

When running the integration test in src/backend/tests/integration/components/output_parsers/test_output_parser.py, it first reports the error that OpenAIModelComponent does not have an output_parser named (but the test wants to pass the value to OpenAImodel through this parameter); after modifying this test code, the test again reports ValueError: Input field input_value not found in target vertex OutputParserComponent-rwv8N. This Is it because this test is faulty? @dosu

[dosubot]

Yes, the test is faulty. The current implementation of OpenAIModelComponent does not have an input parameter named output_parser, nor does OutputParserComponent have an input field named input_value. The errors you’re seeing happen because the test tries to pass these non-existent fields to the components, which isn’t supported by their interfaces. You can see in the code that OpenAIModelComponent only accepts parameters like max_tokens, model_kwargs, json_mode, model_name, api_key, etc., and OutputParserComponent only takes parser_type as input, not input_value (OpenAIModelComponent, OutputParserComponent).

Recent code and PRs also show that output_parser support is deprecated and will be removed soon (PR 8434). To fix the test, update it to match the actual component interfaces—remove the output_parser and input_value fields from the test inputs and use only the supported parameters.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}   

[Sqiyuan]

But the test_output_parser.py test was not removed in the latest code @dosu

[dosubot]

You're right—the test_output_parser.py file is still present in the latest code, even though the components it targets no longer support the parameters used in the test. There's no documentation, comments, or open issues explaining why this test remains, and no evidence that it's intentionally being kept for legacy reasons. Its retention is likely an oversight, and as written, the test will fail due to mismatched inputs and deprecated features in the components it tries to test (source).

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}   

[Sqiyuan]

And the src/backend/base/langflow/components/helpers/output_parser.py file was not deleted either @dosu

[dosubot]

The output_parser.py file is still in the codebase because it was likely overlooked during recent refactors that moved other legacy components to a "deactivated" directory. It's marked as legacy and isn't used anywhere in production code—it's only imported in the helpers module and referenced in a now-broken integration test. There are no open issues, comments, or PRs discussing its removal or future, so its presence appears to be technical debt rather than intentional retention reference refactor context.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}