test(scripts): add 56 tests for validate_sc_notebooks.py (cycle 71)

[jsboige]

Summary

Add 56 tests for scripts/smartcontracts/validate_sc_notebooks.py — SC notebook structural validator.

Functions tested

Zone	Tests	Key invariants
source_text	7	String/list extraction, missing source
is_nav_cell	7	Back/forward link detection regex
has_exposed_solution	8	TODO/NotImplementedError exclusion, triple-quote Solidity, test function with assertEq
make_source_list	5	STRING→LIST conversion, trailing newline behavior
get_executable_cells	12	TODO/NotImplementedError/pip skip, anvil profile gating, web3/compile_and_deploy skip
validate_notebook	14	Title/nav/objectives/duration, STRING format detection, consecutive code, anvil web3 requirement, missing notebook, invalid JSON
Dataclasses	2	NotebookValidation defaults, CellInfo creation
Colors	1	disable() clears all ANSI codes

Test count

• 56 new tests, 506 lines
• Suite: 1008 passed (0 regression)
• Total scripts tests across all cycles: 390→1008

Scope

• Single file: scripts/tests/test_validate_sc_notebooks.py
• Lane: smartcontracts (po-2024:CoursIA-2, non-validation-notebook)
• No notebooks modified, no source scripts modified

Co-Authored-By: Claude Opus 4.8 noreply@anthropic.com

[clusterManager-Myia]

[NanoClaw]

Solid test file overall — good coverage of the pure functions in validate_sc_notebooks.py with well-organized class-based grouping and proper use of tmp_path fixtures. A few issues worth addressing:

1. sys.path.insert hack (line 14)

sys.path.insert(0, str(Path(__file__).resolve().parent.parent / "smartcontracts"))

This is fragile and won't work if tests are invoked from a different working directory or by a test runner that copies files. Prefer:

• A conftest.py at scripts/tests/conftest.py that adds the import path once
• Or add scripts/smartcontracts to testpaths / pythonpath in pyproject.toml under [tool.pytest]

2. TestColors.test_disable — fragile global state (lines 498-506)

This test mutates class-level attributes (Colors.GREEN, Colors.RED, etc.) and tries to restore them afterward, but:

• Only Colors.GREEN is saved/restored from the original value; the rest (RED, YELLOW, BLUE, CYAN, BOLD, END) are hard-coded back to their expected ANSI codes
• If any other test runs in parallel and reads Colors during the restore window, it sees inconsistent state
• Tests for ANSI color helpers are generally low-value — they test that string constants exist, not behavior

Suggestion: Either remove this test class entirely (it tests that class attributes can be set to ""), or refactor Colors to be instance-based so tests can isolate state without global mutation.

3. Dataclass tests are tautological (lines 485-496)

test_notebook_validation_defaults and test_cell_info_creation just verify that Python's dataclass __init__ assigns fields correctly. These tests will never break unless Python itself is broken. Consider either:

• Removing them (no real value)
• Or testing meaningful behavior (e.g., serialization, equality, or how defaults interact with validation logic)

4. Minor: test data naming mismatch (line 82-84)

def test_list_with_trailing_newline(self):
    cell = {"source": ["line1\n", "line2\n", ""]}
    assert source_text(cell) == "line1\nline2\n"

Comments/docstrings reference "hello/world" in surrounding tests but this one uses "line1/line2" — cosmetic but slightly confusing when reading the class as a whole.

5. Count discrepancy

Title says 56 tests. I count 59 test methods across 8 classes (7+7+8+5+12+17+2+1). Minor — worth fixing the title to match.

6. Missing edge case coverage

• make_source_list with \r\n (Windows line endings) — no test for this, and the split logic (splitting on \n) may produce \r artifacts
• get_executable_cells with foundry profile — standalone and anvil are tested, but foundry is only tested implicitly in test_profile_stored. A foundry-specific filtering test would close the gap

Verdict

Good structural quality and meaningful assertions for the core validation logic. The main actionable items are the sys.path hack (#1, should fix before merge) and the fragile Colors test (#2, consider dropping). The rest are suggestions.