claude-dev-env 1.28.1 → 1.29.0

package/scripts/test_groq_bugteam.py

@@ -0,0 +1,391 @@
+"""Tests for groq_bugteam.py pure logic.
+
+Network calls (Groq HTTP) and filesystem/git side effects are out of scope for
+unit tests; they are exercised in the live end-to-end run.
+"""
+
+from __future__ import annotations
+
+import importlib.util
+import pathlib
+import sys
+import urllib.error
+
+import pytest
+
+sys.path.insert(0, str(pathlib.Path(__file__).parent))
+for _cached in list(sys.modules):
+    if _cached == "config" or _cached.startswith("config."):
+        del sys.modules[_cached]
+
+from config import groq_bugteam_config  # noqa: E402
+
+
+def _load_groq_bugteam_module():
+    scripts_directory = pathlib.Path(__file__).parent
+    sys.path.insert(0, str(scripts_directory))
+    for cached_module_name in list(sys.modules):
+        if cached_module_name == "config" or cached_module_name.startswith("config."):
+            del sys.modules[cached_module_name]
+    module_path = scripts_directory / "groq_bugteam.py"
+    module_spec = importlib.util.spec_from_file_location("groq_bugteam", module_path)
+    loaded_module = importlib.util.module_from_spec(module_spec)
+    sys.modules["groq_bugteam"] = loaded_module
+    module_spec.loader.exec_module(loaded_module)
+    return loaded_module
+
+
+groq_bugteam = _load_groq_bugteam_module()
+
+
+class TestConstantsSourcedFromConfig:
+    def test_endpoint_is_imported_from_config(self):
+        assert groq_bugteam.GROQ_API_ENDPOINT == groq_bugteam_config.GROQ_API_ENDPOINT
+
+    def test_primary_model_is_imported_from_config(self):
+        assert groq_bugteam.GROQ_PRIMARY_MODEL == groq_bugteam_config.GROQ_PRIMARY_MODEL
+
+
+class TestClampText:
+    def test_returns_text_unchanged_when_under_limit(self):
+        assert groq_bugteam.clamp_text("hello world", 100) == "hello world"
+
+    def test_truncates_long_text_with_marker(self):
+        long_text = "a" * 1000
+        clamped = groq_bugteam.clamp_text(long_text, 200)
+        assert "truncated" in clamped
+        assert len(clamped) < len(long_text)
+        assert clamped.startswith("a")
+        assert clamped.endswith("a")
+
+    def test_preserves_head_and_tail(self):
+        text = "HEAD" + ("x" * 1000) + "TAIL"
+        clamped = groq_bugteam.clamp_text(text, 100)
+        assert clamped.startswith("HEAD")
+        assert clamped.endswith("TAIL")
+
+
+class TestParseJsonObject:
+    def test_parses_clean_json(self):
+        parsed = groq_bugteam.parse_json_object('{"findings": []}')
+        assert parsed == {"findings": []}
+
+    def test_extracts_json_from_surrounding_prose(self):
+        noisy_response = 'Sure, here is the result:\n\n{"findings": [{"severity": "P1"}]}\n\nLet me know if you need more.'
+        parsed = groq_bugteam.parse_json_object(noisy_response)
+        assert parsed == {"findings": [{"severity": "P1"}]}
+
+    def test_raises_when_no_json_present(self):
+        with pytest.raises(ValueError):
+            groq_bugteam.parse_json_object("no braces here")
+
+
+class TestNormalizeFindings:
+    def test_drops_findings_with_unknown_files(self):
+        raw_findings = [
+            {
+                "severity": "P0",
+                "category": "H",
+                "file": "known.py",
+                "line": 10,
+                "title": "t",
+                "description": "d",
+            },
+            {
+                "severity": "P1",
+                "category": "A",
+                "file": "unknown.py",
+                "line": 5,
+                "title": "t2",
+                "description": "d2",
+            },
+        ]
+        normalized = groq_bugteam.normalize_findings(raw_findings, {"known.py": ""})
+        assert len(normalized) == 1
+        assert normalized[0]["file"] == "known.py"
+
+    def test_coerces_non_string_line_to_int(self):
+        raw_findings = [
+            {
+                "severity": "P0",
+                "category": "H",
+                "file": "a.py",
+                "line": "42",
+                "title": "t",
+                "description": "d",
+            },
+        ]
+        normalized = groq_bugteam.normalize_findings(raw_findings, {"a.py": ""})
+        assert normalized[0]["line"] == 42
+
+    def test_defaults_to_zero_line_on_bad_value(self):
+        raw_findings = [
+            {
+                "severity": "P1",
+                "category": "H",
+                "file": "a.py",
+                "line": "not-a-number",
+                "title": "t",
+                "description": "d",
+            },
+        ]
+        normalized = groq_bugteam.normalize_findings(raw_findings, {"a.py": ""})
+        assert normalized[0]["line"] == 0
+
+    def test_clamps_invalid_severity_to_p2(self):
+        raw_findings = [
+            {
+                "severity": "CRITICAL",
+                "category": "H",
+                "file": "a.py",
+                "line": 1,
+                "title": "t",
+                "description": "d",
+            },
+        ]
+        normalized = groq_bugteam.normalize_findings(raw_findings, {"a.py": ""})
+        assert normalized[0]["severity"] == "P2"
+
+    def test_keeps_single_letter_category(self):
+        raw_findings = [
+            {
+                "severity": "P0",
+                "category": "HIJ",
+                "file": "a.py",
+                "line": 1,
+                "title": "t",
+                "description": "d",
+            },
+        ]
+        normalized = groq_bugteam.normalize_findings(raw_findings, {"a.py": ""})
+        assert normalized[0]["category"] == "H"
+
+    def test_handles_empty_input(self):
+        assert groq_bugteam.normalize_findings([], {"a.py": ""}) == []
+
+
+class TestGroupFindingsByFile:
+    def test_groups_findings_and_preserves_global_indexes(self):
+        findings = [
+            {
+                "file": "a.py",
+                "severity": "P0",
+                "category": "H",
+                "line": 1,
+                "title": "t1",
+                "description": "d1",
+            },
+            {
+                "file": "b.py",
+                "severity": "P1",
+                "category": "A",
+                "line": 2,
+                "title": "t2",
+                "description": "d2",
+            },
+            {
+                "file": "a.py",
+                "severity": "P2",
+                "category": "E",
+                "line": 3,
+                "title": "t3",
+                "description": "d3",
+            },
+        ]
+        grouped = groq_bugteam.group_findings_by_file(findings)
+        assert set(grouped.keys()) == {"a.py", "b.py"}
+        assert [index for index, _ in grouped["a.py"]] == [0, 2]
+        assert [index for index, _ in grouped["b.py"]] == [1]
+
+
+class TestBuildReviewBody:
+    def test_returns_clean_body_when_no_findings(self):
+        body = groq_bugteam.build_review_body([], "llama-3.3-70b-versatile", "", [])
+        assert "clean" in body
+        assert "llama-3.3-70b-versatile" in body
+
+    def test_counts_severities_and_lists_findings(self):
+        findings = [
+            {
+                "severity": "P0",
+                "category": "H",
+                "file": "a.py",
+                "line": 10,
+                "title": "SQL injection",
+                "description": "trace",
+            },
+            {
+                "severity": "P1",
+                "category": "F",
+                "file": "b.py",
+                "line": 5,
+                "title": "silent except",
+                "description": "trace2",
+            },
+        ]
+        fix_outcomes = [
+            {"finding_index": 0, "status": "fixed"},
+            {"finding_index": 1, "status": "skipped", "reason": "too complex"},
+        ]
+        body = groq_bugteam.build_review_body(
+            findings, "llama-3.3-70b-versatile", "abc1234", fix_outcomes
+        )
+        assert "1 P0 / 1 P1 / 0 P2" in body
+        assert "abc1234" in body
+        assert "SQL injection" in body
+        assert "silent except" in body
+        assert "fixed" in body
+        assert "skipped: too complex" in body
+
+    def test_marks_findings_without_outcome_as_not_attempted(self):
+        findings = [
+            {
+                "severity": "P2",
+                "category": "E",
+                "file": "a.py",
+                "line": 1,
+                "title": "dead code",
+                "description": "d",
+            },
+        ]
+        body = groq_bugteam.build_review_body(
+            findings, "llama-3.3-70b-versatile", "", []
+        )
+        assert "not attempted" in body
+
+
+class TestIsRecoverableHttpError:
+    def _make_error(self, status_code: int) -> urllib.error.HTTPError:
+        return urllib.error.HTTPError(
+            url="x", code=status_code, msg="", hdrs=None, fp=None
+        )
+
+    @pytest.mark.parametrize("status", [408, 429, 500, 502, 503, 504])
+    def test_recoverable_statuses(self, status):
+        assert groq_bugteam.is_recoverable_http_error(self._make_error(status)) is True
+
+    @pytest.mark.parametrize("status", [400, 401, 403, 404, 422])
+    def test_non_recoverable_statuses(self, status):
+        assert groq_bugteam.is_recoverable_http_error(self._make_error(status)) is False
+
+    def test_413_triggers_skip_to_next_model(self):
+        assert groq_bugteam.should_skip_to_next_model(self._make_error(413)) is True
+
+    @pytest.mark.parametrize("status", [400, 401, 403, 429, 500, 503])
+    def test_other_statuses_do_not_trigger_model_skip(self, status):
+        assert groq_bugteam.should_skip_to_next_model(self._make_error(status)) is False
+
+
+class TestBuildFixUserMessage:
+    def test_embeds_file_content_byte_for_byte_with_trailing_newline(self):
+        original_content = "line1\nline2\n"
+        message = groq_bugteam.build_fix_user_message("some.py", original_content, findings_block="[]")
+        assert original_content in message
+        assert "line2\n</current_file_contents>" in message
+
+    def test_embeds_file_content_byte_for_byte_without_trailing_newline(self):
+        original_content = "line1\nline2"
+        message = groq_bugteam.build_fix_user_message("some.py", original_content, findings_block="[]")
+        assert f"{original_content}\n</current_file_contents>" in message
+        assert "line2\n\n</current_file_contents>" not in message
+
+
+class TestShouldWriteFixedFile:
+    def test_does_not_write_when_no_finding_applied(self):
+        assert groq_bugteam.should_write_fixed_file(
+            applied_indexes=set(),
+            updated_content="new",
+            current_content="old",
+        ) is False
+
+    def test_does_not_write_when_content_unchanged(self):
+        assert groq_bugteam.should_write_fixed_file(
+            applied_indexes={0},
+            updated_content="same",
+            current_content="same",
+        ) is False
+
+    def test_writes_when_finding_applied_and_content_changed(self):
+        assert groq_bugteam.should_write_fixed_file(
+            applied_indexes={0},
+            updated_content="new",
+            current_content="old",
+        ) is True
+
+
+class TestPreserveTrailingNewline:
+    def test_adds_trailing_newline_when_original_had_one(self):
+        preserved = groq_bugteam.preserve_trailing_newline(
+            original="line1\nline2\n", updated="line1\nfixed2"
+        )
+        assert preserved == "line1\nfixed2\n"
+
+    def test_strips_trailing_newline_when_original_lacked_one(self):
+        preserved = groq_bugteam.preserve_trailing_newline(
+            original="no newline", updated="fixed content\n"
+        )
+        assert preserved == "fixed content"
+
+    def test_keeps_matching_form_unchanged(self):
+        assert (
+            groq_bugteam.preserve_trailing_newline(original="x\n", updated="y\n")
+            == "y\n"
+        )
+        assert (
+            groq_bugteam.preserve_trailing_newline(original="x", updated="y") == "y"
+        )
+
+
+class TestIsSafeRelativePath:
+    def test_rejects_absolute_posix_path(self):
+        assert groq_bugteam.is_safe_relative_path("/etc/passwd") is False
+
+    def test_rejects_parent_directory_escape(self):
+        assert groq_bugteam.is_safe_relative_path("../../etc/passwd") is False
+
+    def test_rejects_embedded_parent_reference(self):
+        assert groq_bugteam.is_safe_relative_path("src/../../etc/passwd") is False
+
+    def test_accepts_simple_relative_path(self):
+        assert groq_bugteam.is_safe_relative_path("src/foo.py") is True
+
+    def test_accepts_nested_relative_path(self):
+        assert groq_bugteam.is_safe_relative_path("packages/mod/scripts/foo.py") is True
+
+
+class TestDecodeSubprocessStderr:
+    def test_decodes_bytes_input(self):
+        decoded = groq_bugteam.decode_subprocess_stderr(b"fatal: broken")
+        assert decoded == "fatal: broken"
+
+    def test_returns_str_input_unchanged(self):
+        assert groq_bugteam.decode_subprocess_stderr("fatal: broken") == "fatal: broken"
+
+    def test_handles_none_input(self):
+        assert groq_bugteam.decode_subprocess_stderr(None) == ""
+
+    def test_replaces_undecodable_bytes(self):
+        decoded = groq_bugteam.decode_subprocess_stderr(b"\xff\xfe broken")
+        assert "broken" in decoded
+
+
+class TestRunPipelineRefusals:
+    def test_rejects_missing_api_key(self, monkeypatch):
+        monkeypatch.delenv("GROQ_API_KEY", raising=False)
+        result = groq_bugteam.run_pipeline({"diff": "anything"})
+        assert "error" in result
+        assert "GROQ_API_KEY" in result["error"]
+
+    def test_rejects_empty_diff(self, monkeypatch):
+        monkeypatch.setenv("GROQ_API_KEY", "gsk_test_placeholder_value")
+        result = groq_bugteam.run_pipeline({"diff": "   ", "files_content": {}})
+        assert "error" in result
+        assert "diff is empty" in result["error"]
+
+    def test_rejects_fixes_without_worktree(self, monkeypatch):
+        monkeypatch.setenv("GROQ_API_KEY", "gsk_test_placeholder_value")
+        result = groq_bugteam.run_pipeline(
+            {"diff": "some diff", "files_content": {"a.py": ""}, "apply_fixes": True}
+        )
+        assert "error" in result
+        assert "worktree_path" in result["error"]