feat: added unit tests for decomposer.py

[madsysharma]

feat: Add unit tests for decomposer.py

Closes #852

Description

The layer-1 query decomposer (nlp-orchestrator/decomposer.py) had only a thin smoke test. This PR replaces it with a comprehensive pytest suite that fully isolates the logic from the Groq LPU by mocking the async chat-completion call: no API key and no network are required to run it.

Cases covered:

• Happy path: parses a JSON array into sub-questions; preserves order and a single-item result; strips surrounding whitespace/newlines.
• 5-question cap: returns the first 5 when the model emits more; leaves exactly-5 untouched.
• Domain guardrail: a non-legal query (model returns []) yields an empty list.
• Fallback branches: every malformed/wrong-shape response falls back to [query]: invalid JSON, ```json markdown fences (the decomposer does not strip them), a JSON object, a JSON scalar string, a list of non-strings, and a mixed-type list.
• Transport failures: generic Exception, TimeoutError, and RuntimeError from the client all fall back to [query]; the result is always a list.
• Request construction: asserts the call uses GROQ_MODEL_FAST, temperature=0.3, max_tokens=512, the "user" role, and that the user's query is interpolated into DECOMPOSE_PROMPT.

A clean _mock_completion(content) helper replaces the previous nested type("obj", ...) mock construction. The module sets a dummy GROQ_API_KEY before import so the suite is runnable standalone (the config module calls sys.exit(1) when the key is absent).

Testing

PS C:\Users\madhu\Documents\nyay-setu-working> pytest -v nlp-orchestrator/tests/test_decomposer.py
================================================= test session starts =================================================
platform win32 -- Python 3.11.7, pytest-9.0.3, pluggy-1.6.0 -- C:\Program Files\Python311\python.exe
cachedir: .pytest_cache
rootdir: C:\Users\madhu\Documents\nyay-setu-working
plugins: anyio-4.13.0, deepeval-4.0.5, Faker-40.19.1, langsmith-0.8.8, asyncio-1.4.0, cov-4.1.0, mock-3.15.1, repeat-0.9.4, rerunfailures-16.3, xdist-3.8.0
asyncio: mode=Mode.STRICT, debug=False, asyncio_default_fixture_loop_scope=None, asyncio_default_test_loop_scope=function
collected 19 items

nlp-orchestrator/tests/test_decomposer.py::test_returns_parsed_sub_questions PASSED                              [  5%]
nlp-orchestrator/tests/test_decomposer.py::test_result_is_list_of_strings PASSED                                 [ 10%]
nlp-orchestrator/tests/test_decomposer.py::test_single_sub_question_preserved PASSED                             [ 15%]
nlp-orchestrator/tests/test_decomposer.py::test_surrounding_whitespace_is_stripped PASSED                        [ 21%]
nlp-orchestrator/tests/test_decomposer.py::test_caps_at_five_sub_questions PASSED                                [ 26%]
nlp-orchestrator/tests/test_decomposer.py::test_exactly_five_sub_questions_unchanged PASSED                      [ 31%]
nlp-orchestrator/tests/test_decomposer.py::test_non_legal_query_returns_empty_list PASSED                        [ 36%]
nlp-orchestrator/tests/test_decomposer.py::test_invalid_json_falls_back_to_original_query PASSED                 [ 42%]
nlp-orchestrator/tests/test_decomposer.py::test_markdown_fenced_json_falls_back PASSED                           [ 47%]
nlp-orchestrator/tests/test_decomposer.py::test_json_object_instead_of_list_falls_back PASSED                    [ 52%]
nlp-orchestrator/tests/test_decomposer.py::test_json_scalar_string_falls_back PASSED                             [ 57%]
nlp-orchestrator/tests/test_decomposer.py::test_list_with_non_string_items_falls_back PASSED                     [ 63%]
nlp-orchestrator/tests/test_decomposer.py::test_list_with_mixed_types_falls_back PASSED                          [ 68%]
nlp-orchestrator/tests/test_decomposer.py::test_api_exception_falls_back_to_original_query PASSED                [ 73%]
nlp-orchestrator/tests/test_decomposer.py::test_timeout_error_falls_back_to_original_query PASSED                [ 78%]
nlp-orchestrator/tests/test_decomposer.py::test_result_is_always_a_list_on_failure PASSED                        [ 84%]
nlp-orchestrator/tests/test_decomposer.py::test_calls_groq_once_with_expected_parameters PASSED                  [ 89%]
nlp-orchestrator/tests/test_decomposer.py::test_user_query_is_embedded_in_prompt PASSED                          [ 94%]
nlp-orchestrator/tests/test_decomposer.py::test_message_uses_user_role PASSED                                    [100%]Running teardown with pytest sessionfinish...


================================================= 19 passed in 0.65s ==================================================

Type of change

• Tests — adds/expands test coverage (non-breaking)

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• Documentation update — N/A (test-only change)
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[vercel]

@madsysharma is attempting to deploy a commit to the CodeBlooded's projects Team on Vercel.

A member of the Team first needs to authorize it.

[github-actions]

Hi @madsysharma, thanks for contributing to Nyay Setu! 🎉

I have automatically:

• 👤 Assigned this PR to you.
• 🏷️ Applied the gssoc:approved label.

Our workflows will now analyze your changes to classify:

• 📈 PR Difficulty: level:*
• 🧩 PR Type: type:*
• 🌟 PR Quality: quality:*

Tip

Ensure your PR description references the issue it resolves (e.g. Closes #123). This allows the bot to inherit any additional labels from that issue!

Happy coding! 🚀

[madsysharma]

Hi @viru0909-dev , please review this PR. Thank you.

[viru0909-dev]

Review Summary

This PR has been reviewed and requires changes before it can be approved. The issues below must be resolved.

Issues to Resolve

Result	Check	Notes
FAIL	Title	Title should follow the conventional commit format (e.g., 'feat:', 'fix:', 'refactor:') or start with a capital letter.
FAIL	Issue Link	An issue number is referenced but without a closing keyword. Use 'Closes #NNN', 'Fixes #NNN', or 'Resolves #NNN' so the issue closes automatically on merge.

Passing Checks

Result	Check	Notes
PASS	Description	Description is complete and well-structured.
PASS	Checklist	All 8 checklist item(s) are checked.
PASS	Merge Conflicts	Conflict status is not yet computed by GitHub — skipping.
PASS	Branch Freshness	Branch was last updated 0.1 days ago.
PASS	CI / Tests	No CI checks are configured or results are not yet available.
PASS	Code Quality	No code quality issues detected in the diff.

---

To proceed: push the required fixes to test/decomposer and the PR will be re-evaluated on the next review run.

[madsysharma]

Hi @viru0909-dev , have made the changes. Please review them and let me know if they're good to go. Thank you.