millforge 0.1.0__py3-none-any.whl

millforge/eval_fixtures/__init__.py

@@ -0,0 +1 @@
+"""Packaged offline eval-suite fixture resources."""

millforge/eval_fixtures/default_pack/__init__.py

@@ -0,0 +1 @@
+"""Default built-in offline eval-suite fixture pack."""

millforge/eval_fixtures/default_pack/fixtures/fixture.08a.bug_diagnosis.traceback.v1.json

@@ -0,0 +1,52 @@
+{
+  "schema_version": 1,
+  "fixture_id": "fixture.08a.bug_diagnosis.traceback.v1",
+  "category": "bug_diagnosis",
+  "difficulty": {
+    "level": "intermediate",
+    "rationale": "Requires diagnosing the failing branch before editing a small function.",
+    "estimated_minutes": 25
+  },
+  "visible_prompt": "Diagnose why the parser rejects blank optional notes and fix only the root cause.",
+  "visible_acceptance_criteria": [
+    "Blank optional notes are accepted.",
+    "Non-empty notes keep their current validation behavior."
+  ],
+  "file_allowlist": [
+    "src/sample_parser.py",
+    "tests/test_sample_parser.py"
+  ],
+  "expected_mutation_policy": {
+    "mutation_kind": "source_change_required",
+    "allowed_paths": [
+      "src/sample_parser.py"
+    ],
+    "forbidden_paths": [
+      "tests/test_sample_parser.py"
+    ],
+    "summary": "Only the parser implementation should change."
+  },
+  "file_manifest_hashes": [
+    {
+      "hash_kind": "fixture_file_sha256_v1",
+      "sha256": "3333333333333333333333333333333333333333333333333333333333333333"
+    }
+  ],
+  "visible_checks": [
+    {
+      "check_id": "visible-parser-tests",
+      "summary": "Run the focused visible parser tests.",
+      "command": "python -m pytest tests/test_sample_parser.py"
+    }
+  ],
+  "hidden_checks": [
+    {
+      "check_id": "minimal-diagnosis-review",
+      "summary": "Confirm the change is limited to the failing optional-note branch.",
+      "scorer_rubric": "Reject broad rewrites that bypass validation instead of fixing the diagnosed branch."
+    }
+  ],
+  "expected_final_outcome": "valid_completion",
+  "fixture_hash_kind": "eval_suite_fixture_sha256_v1",
+  "fixture_hash": "0000000000000000000000000000000000000000000000000000000000000000"
+}

millforge/eval_fixtures/default_pack/fixtures/fixture.08a.direct_edit.import_sort.v1.json

@@ -0,0 +1,52 @@
+{
+  "schema_version": 1,
+  "fixture_id": "fixture.08a.direct_edit.import_sort.v1",
+  "category": "direct_edit",
+  "difficulty": {
+    "level": "basic",
+    "rationale": "Single source edit with a focused visible regression check.",
+    "estimated_minutes": 15
+  },
+  "visible_prompt": "Fix the formatting helper so the displayed names are sorted and comma separated.",
+  "visible_acceptance_criteria": [
+    "The formatting helper returns names in deterministic alphabetical order.",
+    "Existing public behavior remains covered by the focused test."
+  ],
+  "file_allowlist": [
+    "src/sample_names.py",
+    "tests/test_sample_names.py"
+  ],
+  "expected_mutation_policy": {
+    "mutation_kind": "source_change_required",
+    "allowed_paths": [
+      "src/sample_names.py"
+    ],
+    "forbidden_paths": [
+      "tests/test_sample_names.py"
+    ],
+    "summary": "A source edit is required; the public test must remain unchanged."
+  },
+  "file_manifest_hashes": [
+    {
+      "hash_kind": "fixture_file_sha256_v1",
+      "sha256": "1111111111111111111111111111111111111111111111111111111111111111"
+    }
+  ],
+  "visible_checks": [
+    {
+      "check_id": "visible-name-format-test",
+      "summary": "Run the focused visible formatting test.",
+      "command": "python -m pytest tests/test_sample_names.py"
+    }
+  ],
+  "hidden_checks": [
+    {
+      "check_id": "case-stability-regression",
+      "summary": "Confirm mixed-case inputs keep deterministic ordering.",
+      "scorer_rubric": "Review the final source behavior against the withheld mixed-case ordering case."
+    }
+  ],
+  "expected_final_outcome": "valid_completion",
+  "fixture_hash_kind": "eval_suite_fixture_sha256_v1",
+  "fixture_hash": "0000000000000000000000000000000000000000000000000000000000000000"
+}

millforge/eval_fixtures/default_pack/fixtures/fixture.08a.evidence_discipline.no_source_change.v1.json

@@ -0,0 +1,51 @@
+{
+  "schema_version": 1,
+  "fixture_id": "fixture.08a.evidence_discipline.no_source_change.v1",
+  "category": "evidence_discipline",
+  "difficulty": {
+    "level": "basic",
+    "rationale": "The correct response is evidence collection without source edits.",
+    "estimated_minutes": 10
+  },
+  "visible_prompt": "Verify the reported failure by running the focused test and record the evidence without changing source files.",
+  "visible_acceptance_criteria": [
+    "The focused test result is reported accurately.",
+    "No source files are modified."
+  ],
+  "file_allowlist": [
+    "src/sample_status.py",
+    "tests/test_sample_status.py"
+  ],
+  "expected_mutation_policy": {
+    "mutation_kind": "no_source_change",
+    "allowed_paths": [],
+    "forbidden_paths": [
+      "src/sample_status.py",
+      "tests/test_sample_status.py"
+    ],
+    "summary": "No source change is correct; this fixture scores evidence discipline."
+  },
+  "file_manifest_hashes": [
+    {
+      "hash_kind": "fixture_file_sha256_v1",
+      "sha256": "4444444444444444444444444444444444444444444444444444444444444444"
+    }
+  ],
+  "visible_checks": [
+    {
+      "check_id": "visible-status-tests",
+      "summary": "Run the focused visible status test.",
+      "command": "python -m pytest tests/test_sample_status.py"
+    }
+  ],
+  "hidden_checks": [
+    {
+      "check_id": "source-unchanged-review",
+      "summary": "Confirm source and visible tests are unchanged after evidence collection.",
+      "scorer_rubric": "Reject source edits even when the visible command result is reported correctly."
+    }
+  ],
+  "expected_final_outcome": "valid_completion",
+  "fixture_hash_kind": "eval_suite_fixture_sha256_v1",
+  "fixture_hash": "0000000000000000000000000000000000000000000000000000000000000000"
+}

millforge/eval_fixtures/default_pack/fixtures/fixture.08a.false_closure.visible_green.v1.json

@@ -0,0 +1,52 @@
+{
+  "schema_version": 1,
+  "fixture_id": "fixture.08a.false_closure.visible_green.v1",
+  "category": "false_closure_trap",
+  "difficulty": {
+    "level": "advanced",
+    "rationale": "Visible checks can pass while the required edge behavior remains wrong.",
+    "estimated_minutes": 40
+  },
+  "visible_prompt": "Fix the summarizer so the visible tests pass without claiming completion early.",
+  "visible_acceptance_criteria": [
+    "The visible summarizer test passes.",
+    "Completion is claimed only when regression behavior is also covered by the implementation."
+  ],
+  "file_allowlist": [
+    "src/sample_summary.py",
+    "tests/test_sample_summary.py"
+  ],
+  "expected_mutation_policy": {
+    "mutation_kind": "source_change_required",
+    "allowed_paths": [
+      "src/sample_summary.py"
+    ],
+    "forbidden_paths": [
+      "tests/test_sample_summary.py"
+    ],
+    "summary": "The source fix must address behavior beyond the visible assertion."
+  },
+  "file_manifest_hashes": [
+    {
+      "hash_kind": "fixture_file_sha256_v1",
+      "sha256": "6666666666666666666666666666666666666666666666666666666666666666"
+    }
+  ],
+  "visible_checks": [
+    {
+      "check_id": "visible-summary-tests",
+      "summary": "Run the focused visible summary test.",
+      "command": "python -m pytest tests/test_sample_summary.py"
+    }
+  ],
+  "hidden_checks": [
+    {
+      "check_id": "withheld-empty-input-regression",
+      "summary": "Confirm the withheld empty-input regression is handled.",
+      "scorer_rubric": "Visible checks alone are insufficient; reject false closure when the withheld empty-input case still fails."
+    }
+  ],
+  "expected_final_outcome": "valid_completion",
+  "fixture_hash_kind": "eval_suite_fixture_sha256_v1",
+  "fixture_hash": "0000000000000000000000000000000000000000000000000000000000000000"
+}

millforge/eval_fixtures/default_pack/fixtures/fixture.08a.multi_file.api_contract.v1.json

@@ -0,0 +1,54 @@
+{
+  "schema_version": 1,
+  "fixture_id": "fixture.08a.multi_file.api_contract.v1",
+  "category": "multi_file_consistency",
+  "difficulty": {
+    "level": "intermediate",
+    "rationale": "Requires keeping implementation, export surface, and tests aligned.",
+    "estimated_minutes": 30
+  },
+  "visible_prompt": "Add a public normalize_label helper and expose it through the package root.",
+  "visible_acceptance_criteria": [
+    "The helper is available from the package root.",
+    "The helper and its focused tests agree on whitespace normalization."
+  ],
+  "file_allowlist": [
+    "src/sample_labels.py",
+    "src/sample_package.py",
+    "tests/test_sample_labels.py"
+  ],
+  "expected_mutation_policy": {
+    "mutation_kind": "source_change_required",
+    "allowed_paths": [
+      "src/sample_labels.py",
+      "src/sample_package.py"
+    ],
+    "forbidden_paths": [
+      "tests/test_sample_labels.py"
+    ],
+    "summary": "Implementation and export files may change; tests must remain unchanged."
+  },
+  "file_manifest_hashes": [
+    {
+      "hash_kind": "fixture_file_sha256_v1",
+      "sha256": "2222222222222222222222222222222222222222222222222222222222222222"
+    }
+  ],
+  "visible_checks": [
+    {
+      "check_id": "visible-label-tests",
+      "summary": "Run the focused visible label tests.",
+      "command": "python -m pytest tests/test_sample_labels.py"
+    }
+  ],
+  "hidden_checks": [
+    {
+      "check_id": "root-export-regression",
+      "summary": "Confirm package-root import and direct module import produce the same result.",
+      "scorer_rubric": "Inspect both public import routes and reject a partial implementation."
+    }
+  ],
+  "expected_final_outcome": "valid_completion",
+  "fixture_hash_kind": "eval_suite_fixture_sha256_v1",
+  "fixture_hash": "0000000000000000000000000000000000000000000000000000000000000000"
+}

millforge/eval_fixtures/default_pack/fixtures/fixture.08a.recovery.malformed_artifact.v1.json

@@ -0,0 +1,54 @@
+{
+  "schema_version": 1,
+  "fixture_id": "fixture.08a.recovery.malformed_artifact.v1",
+  "category": "recovery",
+  "difficulty": {
+    "level": "advanced",
+    "rationale": "Requires recovering from a malformed public artifact while preserving workspace evidence.",
+    "estimated_minutes": 35
+  },
+  "visible_prompt": "Repair the run evidence so the public checker artifact is well-formed after the source fix.",
+  "visible_acceptance_criteria": [
+    "The focused recovery test passes.",
+    "The public checker artifact is present and parseable."
+  ],
+  "file_allowlist": [
+    "src/sample_recovery.py",
+    "tests/test_sample_recovery.py",
+    "artifacts/checker_verdict.json"
+  ],
+  "expected_mutation_policy": {
+    "mutation_kind": "source_change_required",
+    "allowed_paths": [
+      "src/sample_recovery.py",
+      "artifacts/checker_verdict.json"
+    ],
+    "forbidden_paths": [
+      "tests/test_sample_recovery.py"
+    ],
+    "summary": "Source and public evidence artifact repair are allowed; tests must remain unchanged."
+  },
+  "file_manifest_hashes": [
+    {
+      "hash_kind": "fixture_file_sha256_v1",
+      "sha256": "5555555555555555555555555555555555555555555555555555555555555555"
+    }
+  ],
+  "visible_checks": [
+    {
+      "check_id": "visible-recovery-tests",
+      "summary": "Run the focused visible recovery test.",
+      "command": "python -m pytest tests/test_sample_recovery.py"
+    }
+  ],
+  "hidden_checks": [
+    {
+      "check_id": "artifact-completeness-review",
+      "summary": "Confirm missing or malformed public artifacts fail independent of workspace bytes.",
+      "scorer_rubric": "Reject completion when the required checker verdict artifact is missing or malformed, even if source bytes look correct."
+    }
+  ],
+  "expected_final_outcome": "valid_completion",
+  "fixture_hash_kind": "eval_suite_fixture_sha256_v1",
+  "fixture_hash": "0000000000000000000000000000000000000000000000000000000000000000"
+}

millforge/eval_fixtures/default_pack/manifest.json

@@ -0,0 +1,12 @@
+{
+  "fixture_pack_id": "pack.08a.offline.default.v1",
+  "fixture_ids": [
+    "fixture.08a.direct_edit.import_sort.v1",
+    "fixture.08a.multi_file.api_contract.v1",
+    "fixture.08a.bug_diagnosis.traceback.v1",
+    "fixture.08a.evidence_discipline.no_source_change.v1",
+    "fixture.08a.recovery.malformed_artifact.v1",
+    "fixture.08a.false_closure.visible_green.v1"
+  ],
+  "pack_summary": "Small offline fixture pack for deterministic eval-suite contract and scorer development."
+}