Feature/update

.github/workflows/ci.yml

@@ -0,0 +1,25 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+
+      - name: Install dependencies
+        run: pip install -r requirements-dev.txt
+
+      - name: Run unit tests
+        run: pytest
+
+      - name: Compile check all Python files
+        run: python -m py_compile $(git ls-files '*.py')

conftest.py

@@ -0,0 +1,4 @@
+import sys
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).parent))

requirements-dev.txt

@@ -0,0 +1,2 @@
+-r requirements.txt
+pytest>=8.0

tests/test_critique.py

@@ -0,0 +1,37 @@
+import pytest
+from pydantic import ValidationError
+
+from shared.critique import Critique
+
+
+def test_valid_critique_round_trips_fields() -> None:
+    critique = Critique(
+        score=85,
+        issues=["missing docstring"],
+        verdict="pass",
+        reasoning="Clean and correct, one minor style nit.",
+    )
+    assert critique.score == 85
+    assert critique.issues == ["missing docstring"]
+    assert critique.verdict == "pass"
+
+
+def test_issues_defaults_to_empty_list() -> None:
+    critique = Critique(score=100, verdict="pass", reasoning="No issues found.")
+    assert critique.issues == []
+
+
+@pytest.mark.parametrize("score", [-1, 101])
+def test_score_out_of_range_rejected(score: int) -> None:
+    with pytest.raises(ValidationError):
+        Critique(score=score, verdict="pass", reasoning="x")
+
+
+def test_invalid_verdict_rejected() -> None:
+    with pytest.raises(ValidationError):
+        Critique(score=50, verdict="maybe", reasoning="x")
+
+
+def test_missing_required_field_rejected() -> None:
+    with pytest.raises(ValidationError):
+        Critique(score=50, verdict="pass")

tests/test_llm_client.py

@@ -0,0 +1,73 @@
+from shared.llm_client import TokenBudget, Usage, extract_code_block
+
+
+class TestUsage:
+    def test_total_tokens_sums_input_and_output(self) -> None:
+        usage = Usage(input_tokens=100, output_tokens=50, model="claude-sonnet-5")
+        assert usage.total_tokens == 150
+
+    def test_cost_usd_known_model(self) -> None:
+        # claude-sonnet-5: $2.00/1M in, $10.00/1M out
+        usage = Usage(input_tokens=1_000_000, output_tokens=1_000_000, model="claude-sonnet-5")
+        assert usage.cost_usd == 12.00
+
+    def test_cost_usd_unknown_model_is_zero(self) -> None:
+        usage = Usage(input_tokens=1_000_000, output_tokens=1_000_000, model="not-a-real-model")
+        assert usage.cost_usd == 0.0
+
+
+class TestTokenBudget:
+    def test_record_accumulates_spend_and_log(self) -> None:
+        budget = TokenBudget(max_total_tokens=1000)
+        budget.record(Usage(100, 50, "claude-sonnet-5"))
+        budget.record(Usage(200, 100, "claude-sonnet-5"))
+
+        assert budget.spent_tokens == 450
+        assert len(budget.usage_log) == 2
+
+    def test_exceeded_false_under_cap(self) -> None:
+        budget = TokenBudget(max_total_tokens=1000)
+        budget.record(Usage(100, 50, "claude-sonnet-5"))
+        assert budget.exceeded() is False
+
+    def test_exceeded_true_at_or_over_cap(self) -> None:
+        budget = TokenBudget(max_total_tokens=100)
+        budget.record(Usage(80, 20, "claude-sonnet-5"))
+        assert budget.exceeded() is True
+
+    def test_remaining_floors_at_zero(self) -> None:
+        budget = TokenBudget(max_total_tokens=100)
+        budget.record(Usage(80, 40, "claude-sonnet-5"))
+        assert budget.remaining() == 0
+
+    def test_total_cost_usd_sums_logged_usage(self) -> None:
+        budget = TokenBudget(max_total_tokens=10_000_000)
+        budget.record(Usage(1_000_000, 0, "claude-sonnet-5"))  # $2.00
+        budget.record(Usage(0, 1_000_000, "claude-haiku-4-5"))  # $5.00
+        assert budget.total_cost_usd == 7.00
+
+
+class TestExtractCodeBlock:
+    def test_fenced_block_with_language_tag(self) -> None:
+        text = "```python\ndef foo():\n    pass\n```"
+        assert extract_code_block(text) == "def foo():\n    pass"
+
+    def test_fenced_block_without_language_tag(self) -> None:
+        text = "```\nhello\n```"
+        assert extract_code_block(text) == "hello"
+
+    def test_fenced_block_surrounded_by_prose(self) -> None:
+        text = "Here's the code:\n```python\nx = 1\n```\nDone."
+        assert extract_code_block(text) == "x = 1"
+
+    def test_falls_back_to_raw_text_when_unfenced(self) -> None:
+        text = "  just some text, no fence  "
+        assert extract_code_block(text) == "just some text, no fence"
+
+    def test_returns_only_first_block_when_multiple(self) -> None:
+        text = "```\nfirst\n```\nsome text\n```\nsecond\n```"
+        assert extract_code_block(text) == "first"
+
+    def test_empty_fenced_block(self) -> None:
+        text = "```\n\n```"
+        assert extract_code_block(text) == ""