Test infra follow-ups #2: library invariants, helpers.py coverage 45→86%, cov in CI

[dokson]

Summary

Closes the test-infra follow-ups deferred in #140. Five logically-distinct improvements bundled because they all touch tests/ together and share the same audit pass (pytest --cov=financedatabase --cov-report=term-missing).

Coverage 78% → 92%. Tests 32 → 49. No snapshot file modified — only test code, config, and the workflow step.

Improvement	Files
1. test_exchange_market_one_to_one uses the library	tests/test_equities.py
2. Cover the helpers.py 45% → 86% gap (+ asset-class invalid-value tests)	tests/test_*.py
3. Behavioural smoke asserts in every test_select	all 7 asset-class test files
4. Coverage reporting in CI	.github/workflows/testing.yml
5. Deferred items called out explicitly (xdist, parametrize)	—

1. test_exchange_market_one_to_one now uses the library

After #140, the library reads local data (use_local_location=True) and the conftest regenerates compression artifacts from database/*.csv at import time, so equities.select() is in sync with the checked-out source of truth for the test session. The previous pd.read_csv("database/equities.csv") workaround is no longer needed:

-    import pandas as pd
-    from pathlib import Path
-    csv_path = Path(__file__).resolve().parents[1] / "database" / "equities.csv"
-    df = pd.read_csv(csv_path, dtype=str)
+    df = equities.select()

The invariant is identical (each exchange code maps to exactly one market label); only the data source is now the library, consistent with every other test in the file.

2. Coverage of helpers.py and asset-class validation paths

Targeted tests for every uncovered branch identified by pytest --cov. All pure asserts (no recorder), following the test_sec_enrichment_controller.py precedent. New tests added in tests/test_equities.py:

New test	Branch covered (helpers.py / Equities.py)
test_search_with_list_of_index	search(index=[…]) list-membership
test_search_case_sensitive	search(case_sensitive=True)
test_search_case_sensitive_with_list	list + case-sensitive combo (helpers.py:124-141)
test_search_invalid_column_is_ignored	key not in data_filter.columns warning
test_to_toolkit_raises_without_financetoolkit	to_toolkit ImportError path (mocked)
test_to_toolkit_success_path	to_toolkit success path (mocked Toolkit, helpers.py:236-265)
test_init_raises_on_request_failure	__init__ RequestException path (mocked)
test_module_show_options_raises_on_invalid_selection	module-level show_options ValueError paths
test_module_show_options_raises_on_request_failure	module-level show_options network failure (mocked)
test_module_show_options_local_location	module-level show_options(use_local_location=True) (helpers.py:320, 336-341)

Plus one parallel test per asset class in tests/test_<asset>.py:

New test	Branch covered
test_select_with_invalid_value_raises	select(<filter>=...) ValueError for unknown values — covers the validate-and-raise blocks unique to each asset class (e.g. ETFs.py:83, 96, 109, 115-126)

The asset-class test exercises every filter column declared on that class's select() signature, so adding a new filter to the API naturally requires the test to be updated.

3. Behavioural smoke assert in every asset-class test_select

Snapshot tests pass even when the library returns an empty DataFrame (the snapshot, regenerated under the same broken state, would just be empty too — and the test would pass silently). One cheap assertion per file catches that class of regression:

smoke = equities.select()
assert not smoke.empty
assert "country" in smoke.columns

Applied to all 7 asset-class test files. Trivial runtime cost; meaningful safety net against the empty-DataFrame failure mode that pure snapshot tests can't detect.

4. Coverage reporting in CI

pytest-cov was already in [tool.poetry.group.dev.dependencies] but wasn't invoked by the workflow. Wired it in:

-                    poetry run pytest tests/
+                    poetry run pytest tests/ --cov=financedatabase --cov-report=term-missing

Coverage now appears in the CI log; gaps become visible at PR review time without anyone having to run pytest locally. No Codecov upload — the terminal report is enough for review-time signal; uploading to a third party would add a token + service dependency and can be a focused follow-up if you want a coverage badge / trend chart.

Coverage breakdown (before → after)

File	Before this PR	After this PR
helpers.py	45%	86%
ETFs.py	73%	90%
Equities.py	88%	94%
Funds.py	83%	92%
Indices.py	89%	93%
Cryptos.py / Currencies.py / Moneymarkets.py	90%	97%
Total	78%	92%

5. What this PR explicitly does NOT do (still deferred)

• pytest-xdist parallel execution. Tried it (-n auto); the conftest's compression-regen at import time runs once per worker, racing on the same files. For a ~25s suite the per-worker import overhead actually slows the wall time. The fix is non-trivial (coordinate regen across workers via a lock file, or convert the regen to a pytest-xdist-compatible session fixture) and not worth doing for a suite this small.
• @pytest.mark.parametrize / class-based grouping. Would change pytest node IDs and break the snapshot positional naming convention (test_select.csv, test_select_1.csv, …, test_select_12.csv). Would require renaming 90+ snapshot files in one go — preferred as a focused refactor PR if you want it.
• Replacing the vendored Record/PathTemplate/Recorder (~220 LOC of test infra in conftest.py) with syrupy. Mentioned in passing because it would delete a lot of code, but again touches every snapshot file. Out of scope here.

Test plan

• pytest tests/ — 49 tests pass (was 32 before Baseline test improvements: snapshots, local data, docstrings, infra cleanup #140 + this PR)
• CI coverage block appears in the run log with per-file %
• black --check tests/ clean
• No snapshot files modified — only test code, pyproject.toml, and .github/workflows/testing.yml