Merge pull request #14 from nandanadileep/fix/quality-hotspots-sarif-errors

fix: positional path arg for quality/hotspots, deduplicate hotspots, suppress sarif import noise

Author: M9nx

semantic_code_intelligence/ci/hotspots.py

@@ -172,7 +172,16 @@ def analyze_hotspots(
         w_fan_out += extra
         w_churn = 0.0
 
-    callable_symbols = [s for s in symbols if s.kind in ("function", "method")]
+    # Deduplicate by (file_path, name) so that re-indexed or multiply-parsed
+    # symbols don't appear as separate hotspot entries with identical scores.
+    _seen: set[tuple[str, str]] = set()
+    callable_symbols = []
+    for s in symbols:
+        if s.kind in ("function", "method"):
+            key = (s.file_path, s.name)
+            if key not in _seen:
+                _seen.add(key)
+                callable_symbols.append(s)
 
     # ── Per-symbol raw metrics ───────────────────────────────────
     # Complexity

semantic_code_intelligence/ci/quality.py

@@ -354,8 +354,62 @@ def detect_duplicates(
 # ── Bandit security linting (optional) ───────────────────────────────
 
 try:
-    from bandit.core import manager as _bandit_manager
-    from bandit.core import config as _bandit_config
+    import logging as _logging
+
+    # Bandit eagerly loads all its formatters at import time, including a
+    # SARIF formatter that requires the optional `sarif_om` package.  When
+    # `sarif_om` is absent bandit logs an ERROR that appears on every run
+    # even when SARIF output was never requested.  Suppress that specific
+    # message during the import so users only see errors that are relevant
+    # to them.
+    class _SuppressSarifFilter(_logging.Filter):
+        def filter(self, record: _logging.LogRecord) -> bool:
+            return "Could not load 'sarif'" not in record.getMessage()
+
+    def _logger_ancestry(logger: _logging.Logger) -> list[_logging.Logger]:
+        ancestry: list[_logging.Logger] = []
+        current: _logging.Logger | None = logger
+        while current is not None:
+            ancestry.append(current)
+            if not current.propagate:
+                break
+            parent = current.parent
+            current = parent if isinstance(parent, _logging.Logger) else None
+        return ancestry
+
+    def _add_filter_to_handlers(
+        logger: _logging.Logger,
+        log_filter: _logging.Filter,
+    ) -> list[_logging.Handler]:
+        filtered_handlers: list[_logging.Handler] = []
+        seen_handlers: set[int] = set()
+        for current_logger in _logger_ancestry(logger):
+            for handler in current_logger.handlers:
+                handler_id = id(handler)
+                if handler_id in seen_handlers:
+                    continue
+                handler.addFilter(log_filter)
+                filtered_handlers.append(handler)
+                seen_handlers.add(handler_id)
+        return filtered_handlers
+
+    def _remove_filter_from_handlers(
+        handlers: list[_logging.Handler],
+        log_filter: _logging.Filter,
+    ) -> None:
+        for handler in handlers:
+            handler.removeFilter(log_filter)
+
+    _bandit_root = _logging.getLogger("bandit")
+    _sarif_filter = _SuppressSarifFilter()
+    _filtered_handlers = _add_filter_to_handlers(_bandit_root, _sarif_filter)
+    _bandit_root.addFilter(_sarif_filter)
+    try:
+        from bandit.core import manager as _bandit_manager
+        from bandit.core import config as _bandit_config
+    finally:
+        _bandit_root.removeFilter(_sarif_filter)
+        _remove_filter_from_handlers(_filtered_handlers, _sarif_filter)
 
     _HAS_BANDIT = True
 except ImportError:  # pragma: no cover

semantic_code_intelligence/cli/commands/hotspots_cmd.py

@@ -18,11 +18,17 @@
 
 
 @click.command("hotspots")
+@click.argument(
+    "directory",
+    default=None,
+    required=False,
+    type=click.Path(exists=True, file_okay=False, resolve_path=True),
+)
 @click.option(
     "--path", "-p",
     default=".",
     type=click.Path(exists=True, file_okay=False, resolve_path=True),
-    help="Project root path.",
+    help="Project root path (alternative to the positional argument).",
 )
 @click.option(
     "--json-output", "--json", "json_mode",
@@ -47,6 +53,7 @@
 @click.pass_context
 def hotspots_cmd(
     ctx: click.Context,
+    directory: str | None,
     path: str,
     json_mode: bool,
     pipe: bool,
@@ -62,14 +69,20 @@ def hotspots_cmd(
 
         codexa hotspots
 
+        codexa hotspots .
+
         codexa hotspots --top-n 10 --json
 
         codexa hotspots --no-git --pipe
     """
     from semantic_code_intelligence.ci.hotspots import analyze_hotspots
     from semantic_code_intelligence.context.engine import CallGraph, ContextBuilder, DependencyMap
 
-    root = Path(path).resolve()
+    if directory is not None and ctx.get_parameter_source("path") == click.core.ParameterSource.COMMANDLINE:
+        raise click.UsageError(
+            "Provide either the positional 'directory' argument or '--path', not both."
+        )
+    root = Path(directory or path).resolve()
     builder = ContextBuilder()
     dep_map = DependencyMap()
 

semantic_code_intelligence/cli/commands/quality_cmd.py

@@ -112,12 +112,18 @@ def _output_report_rich(report: "QualityReport", root: Path) -> None:
 
 
 @click.command("quality")
+@click.argument(
+    "directory",
+    default=None,
+    required=False,
+    type=click.Path(exists=True, file_okay=False, resolve_path=True),
+)
 @click.option(
     "--path",
     "-p",
     default=".",
     type=click.Path(exists=True, file_okay=False, resolve_path=True),
-    help="Project root path.",
+    help="Project root path (alternative to the positional argument).",
 )
 @click.option(
     "--json-output",
@@ -148,6 +154,7 @@ def _output_report_rich(report: "QualityReport", root: Path) -> None:
 @click.pass_context
 def quality_cmd(
     ctx: click.Context,
+    directory: str | None,
     path: str,
     json_mode: bool,
     complexity_threshold: int,
@@ -163,6 +170,8 @@ def quality_cmd(
 
         codexa quality
 
+        codexa quality .
+
         codexa quality --json
 
         codexa quality --safety-only --pipe
@@ -172,8 +181,15 @@ def quality_cmd(
     from semantic_code_intelligence.ci.quality import analyze_project, QualityReport
     from semantic_code_intelligence.llm.safety import SafetyValidator
 
-    root = Path(path).resolve()
+    directory_path = Path(directory).resolve() if directory else None
+    option_path = Path(path).resolve()
+
+    if directory_path is not None and directory_path != option_path:
+        raise click.UsageError(
+            "Cannot use both the positional 'directory' argument and '--path' with different values."
+        )
 
+    root = directory_path or option_path
     if safety_only:
         # Fast path: only safety scan
         from semantic_code_intelligence.parsing.parser import EXTENSION_TO_LANGUAGE