fix(tests): configuration management system

changes:
  - file: command_handlers.py
    area: cli
    modified: [_show_general_info, check_command, validate_command, batch_command]
  - file: settings_builders.py
    area: cli
    modified: [build_validate_settings, build_batch_settings]
  - file: config.py
    area: config
    modified: [VallmSettings]
  - file: scoring.py
    area: core
    modified: [_get_default_validators]
  - file: regression.py
    area: core
    added: [_parse_failures, _resolve_test_dir, validate, _interpret, _run_pytest, _timeout_result, +5 more]
  - file: conftest.py
    area: test
    added: [Client]
    modified: [disable_external_calls]
  - file: test_pipeline.py
    area: test
    added: [DummyRegressionValidator, validate]
    new_tests: 1
  - file: test_regression_validator.py
    area: test
    added: [TestInterpret, TestParseFailures, setup_method, _make_proposal, TestErrorHelpers, TestValidate, +2 more]
    new_tests: 24

testing:
  new_tests: 25
  scenarios:
    - pipeline_includes_regression_validator
    - pytest_usage_error
    - tests_failed_with_failed_lines
    - pytest_internal_error
    - validate_passes
    - cmd_includes_test_dir
    - validator_name
    - no_tests_collected
    - tests_failed_no_failed_lines
    - pytest_interrupted
    # +15 more

stats:
  lines: "+613/-18 (net +595)"
  files: 9
  complexity: "Large structural change (normalized)"

Author: tom-sapletta-com

.gitignore

@@ -99,3 +99,4 @@ htmlcov/
 # Backup files
 *.bak
 *.backup
+.aider*

CHANGELOG.md

@@ -124,6 +124,16 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - **Tree-sitter for all** — syntax validation for 165+ languages
 - **Example 07** — comprehensive multi-language demo with 8 languages
 
+## [0.1.52] - 2026-03-24
+
+### Test
+- Update tests/conftest.py
+- Update tests/test_pipeline.py
+- Update tests/test_regression_validator.py
+
+### Other
+- Update .gitignore
+
 ## [0.1.51] - 2026-03-23
 
 ### Docs

VERSION

@@ -1 +1 @@
-0.1.51
+0.1.52

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "vallm"
-version = "0.1.51"
+version = "0.1.52"
 description = "A complete toolkit for validating LLM-generated code"
 readme = "README.md"
 license = "Apache-2.0"

src/vallm/__init__.py

@@ -16,4 +16,4 @@
     "detect_language",
 ]
 
-__version__ = "0.1.51"
+__version__ = "0.1.52"

src/vallm/cli/command_handlers.py

@@ -15,6 +15,7 @@
 from vallm.core.proposal import Proposal
 
 console = Console()
+OUTPUT_FORMAT_HELP = "Output format"
 
 
 def validate_command(
@@ -25,8 +26,9 @@ def validate_command(
     config: Optional[Path] = typer.Option(None, "--config", help="Path to vallm.toml"),
     enable_semantic: bool = typer.Option(False, "--semantic", help="Enable LLM-as-judge"),
     enable_security: bool = typer.Option(False, "--security", help="Enable security checks"),
+    enable_regression: bool = typer.Option(False, "--regression", help="Enable regression tests"),
     model: Optional[str] = typer.Option(None, "--model", "-m", help="LLM model for semantic"),
-    output_format: str = typer.Option("rich", "--output", "-o", help="Output format"),
+    output_format: str = typer.Option("rich", "--output", "-o", help=OUTPUT_FORMAT_HELP),
     verbose: bool = typer.Option(False, "--verbose", "-v", help="Show detailed results"),
     exit_on_verdict: bool = typer.Option(False, "--exit", help="Exit with non-zero on fail/review"),
 ) -> None:
@@ -41,7 +43,14 @@ def validate_command(
     detected_language = _detect_and_log_language(file, language)
 
     # Build settings
-    settings = build_validate_settings(config, enable_semantic, enable_security, model, verbose)
+    settings = build_validate_settings(
+        config,
+        enable_semantic,
+        enable_security,
+        enable_regression,
+        model,
+        verbose,
+    )
 
     # Create proposal
     proposal = _build_proposal(code_str, detected_language, ref_code, file)
@@ -79,7 +88,7 @@ def check_command(
     from vallm.scoring import validate
 
     # Only enable syntax validator
-    settings = build_validate_settings(None, False, False, None, False)
+    settings = build_validate_settings(None, False, False, False, None, False)
     settings.enable_syntax = True
     settings.enable_imports = False
     settings.enable_complexity = False
@@ -100,10 +109,11 @@ def batch_command(
     use_gitignore: bool = typer.Option(True, "--gitignore/--no-gitignore", help="Respect .gitignore"),
     enable_semantic: bool = typer.Option(False, "--semantic", help="Enable LLM-as-judge"),
     enable_security: bool = typer.Option(False, "--security", help="Enable security checks"),
+    enable_regression: bool = typer.Option(False, "--regression", help="Enable regression tests"),
     no_imports: bool = typer.Option(False, "--no-imports", help="Skip import validation (faster)"),
     no_complexity: bool = typer.Option(False, "--no-complexity", help="Skip complexity analysis (faster)"),
     model: Optional[str] = typer.Option(None, "--model", "-m", help="LLM model for semantic"),
-    format: str = typer.Option("rich", "--format", "-f", help="Output format"),
+    format: str = typer.Option("rich", "--format", "-f", help=OUTPUT_FORMAT_HELP),
     output: Optional[Path] = typer.Option(None, "--output", "-o", help="Output directory"),
     fail_fast: bool = typer.Option(False, "--fail-fast", help="Stop on first failure"),
     verbose: bool = typer.Option(False, "--verbose", "-v", help="Show detailed validation results for each file"),
@@ -113,7 +123,15 @@ def batch_command(
     from vallm.cli.batch_processor import BatchProcessor
     processor = BatchProcessor(console)
 
-    settings = build_batch_settings(enable_semantic, enable_security, model, verbose, no_imports, no_complexity)
+    settings = build_batch_settings(
+        enable_semantic,
+        enable_security,
+        enable_regression,
+        model,
+        verbose,
+        no_imports,
+        no_complexity,
+    )
 
     results_by_language, failed_files, passed_count, filtered_files = processor.process_batch(
         paths=paths,
@@ -309,13 +327,14 @@ def _show_general_info() -> None:
     console.print("  2. Import validation - Module resolution checking")
     console.print("  3. Complexity analysis - Cyclomatic complexity metrics")
     console.print("  4. Security analysis - Security pattern detection")
-    console.print("  5. Semantic analysis - LLM-powered code review")
+    console.print("  5. Regression testing - Pytest-based regression checks")
+    console.print("  6. Semantic analysis - LLM-powered code review")
 
     # Show cache stats
     cache_stats = get_semantic_cache().get_cache_stats()
-    console.print(f"\n[bold]Semantic Cache:[/bold]")
+    console.print("\n[bold]Semantic Cache:[/bold]")
     console.print(f"  Cached entries: {cache_stats['total_entries']}")
-    console.print(f"  Use 'vallm info --clear-cache' to clear cache")
+    console.print("  Use 'vallm info --clear-cache' to clear cache")
 
     console.print("\n[bold]Import Validation Languages:[/bold]")
     factory = ImportValidatorFactory()

src/vallm/cli/settings_builders.py

@@ -12,6 +12,7 @@ def build_validate_settings(
     config: Optional[Path],
     enable_semantic: bool,
     enable_security: bool,
+    enable_regression: bool,
     model: Optional[str],
     verbose: bool,
 ) -> VallmSettings:
@@ -27,6 +28,8 @@ def build_validate_settings(
         settings.enable_semantic = enable_semantic
     if enable_security:
         settings.enable_security = enable_security
+    if enable_regression:
+        settings.enable_regression = enable_regression
     if model:
         settings.llm_model = model
     if verbose:
@@ -38,6 +41,7 @@ def build_validate_settings(
 def build_batch_settings(
     enable_semantic: bool,
     enable_security: bool,
+    enable_regression: bool,
     model: Optional[str],
     verbose: bool,
     no_imports: bool,
@@ -49,6 +53,7 @@ def build_batch_settings(
     # Override with command line options
     settings.enable_semantic = enable_semantic
     settings.enable_security = enable_security
+    settings.enable_regression = enable_regression
     if model:
         settings.llm_model = model
     settings.verbose = verbose

src/vallm/config.py

@@ -25,6 +25,7 @@ class VallmSettings(BaseSettings):
     enable_imports: bool = True
     enable_complexity: bool = True
     enable_security: bool = False
+    enable_regression: bool = False
     enable_semantic: bool = False
 
     # Complexity limits

src/vallm/scoring.py

@@ -141,7 +141,7 @@ def validate(
     settings = _initialize_settings(settings)
     context = _initialize_context(context)
     validators = _initialize_validators(validators, settings)
-    
+
     results = _run_validation_pipeline(proposal, validators, context, settings)
     return compute_verdict(results, settings, proposal.filename)
 
@@ -160,14 +160,14 @@ def _initialize_validators(validators: Optional[list], settings: VallmSettings)
     """Initialize validators with defaults if not provided and sort by tier."""
     if validators is None:
         validators = _get_default_validators(settings)
-    
+
     # Sort validators by tier for fail-fast behavior
     validators.sort(key=lambda v: v.tier)
     return validators
 
 
-def _run_validation_pipeline(proposal: Proposal, validators: list, 
-                            context: dict, settings: VallmSettings) -> list:
+def _run_validation_pipeline(proposal: Proposal, validators: list,
+                              context: dict, settings: VallmSettings) -> list:
     """Run the validation pipeline and collect results."""
     results = []
     for validator in validators:
@@ -177,7 +177,7 @@ def _run_validation_pipeline(proposal: Proposal, validators: list,
         # Fail fast on errors in tier 1
         if result.has_errors and validator.tier == 1:
             return results
-    
+
     return results
 
 
@@ -205,6 +205,11 @@ def _get_default_validators(settings: VallmSettings) -> list:
 
         validators.append(SecurityValidator())
 
+    if getattr(settings, "enable_regression", False):
+        from vallm.validators.regression import RegressionValidator
+
+        validators.append(RegressionValidator())
+
     if settings.enable_semantic:
         from vallm.validators.semantic import SemanticValidator