perf(ticket-215): optimize exporters line count reuse, purity calculation, and code2logic exclude forwarding

VERSION

@@ -1 +1 @@
-0.5.182
+0.5.183

code2llm/__init__.py

@@ -8,7 +8,7 @@
 and entity resolution with multilingual support.
 """
 
-__version__ = "0.5.182"
+__version__ = "0.5.183"
 __author__ = "STTS Project"
 
 # Core analysis components (lightweight, always needed)

code2llm/cli_exports/code2logic.py

@@ -65,6 +65,10 @@ def _build_code2logic_cmd(args, source_path: Path, output_dir: Path) -> list[str
         "-o",
         str(output_dir),
     ]
+    excludes = getattr(args, "exclude", None)
+    if excludes:
+        cmd.append("--exclude")
+        cmd.extend(str(x) for x in excludes)
     if not args.verbose:
         cmd.append("-q")
     return cmd

code2llm/exporters/flow_constants.py

@@ -27,6 +27,10 @@
     "egg-info",
     ".tox",
     ".mypy_cache",
+    ".subactor",
+    ".worktrees",
+    ".planfile",
+    ".benchmarks",
     # Backup directories that often contain nested venvs
     ".algitex",
     ".backup",

code2llm/exporters/map/module_list.py

@@ -14,7 +14,7 @@ def render_module_list(result: AnalysisResult, is_excluded_path) -> List[str]:
         if is_excluded_path(mi.file):
             continue
         rel = rel_path(mi.file, result.project_path)
-        lc = file_line_count(mi.file)
+        lc = getattr(mi, "line_count", 0) or file_line_count(mi.file)
         modules.append((rel, lc))
 
     lines = [f"M[{len(modules)}]:"]

code2llm/exporters/map/utils.py

@@ -41,7 +41,7 @@ def count_total_lines(result: AnalysisResult, is_excluded_path) -> int:
     for mi in result.modules.values():
         if mi.file and mi.file not in seen and not is_excluded_path(mi.file):
             seen.add(mi.file)
-            total += file_line_count(mi.file)
+            total += getattr(mi, "line_count", 0) or file_line_count(mi.file)
     return total
 
 

code2llm/exporters/map/yaml_export.py

@@ -58,7 +58,7 @@ def _build_module_entry(
 ) -> Dict[str, Any]:
     """Build the YAML dict for a single module."""
     rel = rel_path(mi.file, result.project_path)
-    lc = file_line_count(mi.file)
+    lc = getattr(mi, "line_count", 0) or file_line_count(mi.file)
     return {
         "path": rel,
         "lines": lc,

code2llm/exporters/toon/_render_section_helpers.py

@@ -6,7 +6,7 @@
 
 from collections import defaultdict
 from pathlib import Path
-from typing import Any, Dict, List
+from typing import Any, Dict, List, Optional
 
 from code2llm.core.models import AnalysisResult
 from code2llm.core.config import LANGUAGE_EXTENSIONS
@@ -147,26 +147,33 @@ def _trace_pipeline(start_func: str, result: AnalysisResult, depth: int) -> List
     return chain
 
 
-def _calculate_purity(chain: List[str], result: AnalysisResult) -> float:
+def _calculate_purity(
+    chain: List[str],
+    result: AnalysisResult,
+    name_to_func: Optional[Dict[str, Any]] = None,
+) -> float:
     """Calculate purity ratio (functions without side effects)."""
     if not chain:
         return 0.0
     pure_count = 0
     for func_name in chain:
-        full_name = next(
-            (qname for qname, fi in result.functions.items() if fi.name == func_name),
-            None,
-        )
-        if full_name:
-            func_info = result.functions.get(full_name)
-            if func_info and not getattr(func_info, "has_side_effects", False):
-                pure_count += 1
+        if name_to_func is not None:
+            func_info = name_to_func.get(func_name)
+        else:
+            full_name = next(
+                (qname for qname, fi in result.functions.items() if fi.name == func_name),
+                None,
+            )
+            func_info = result.functions.get(full_name) if full_name else None
+        if func_info and not getattr(func_info, "has_side_effects", False):
+            pure_count += 1
     return pure_count / len(chain)
 
 
 def render_pipelines_section(ctx: Dict[str, Any]) -> List[str]:
     """Render PIPELINES section - data flow pipelines from entry points."""
     result: AnalysisResult = ctx["result"]
+    name_to_func: Dict[str, Any] = {fi.name: fi for fi in result.functions.values()}
     pipelines = []
     for func_name, func_info in result.functions.items():
         if not func_info.called_by and func_info.calls:
@@ -176,7 +183,9 @@ def render_pipelines_section(ctx: Dict[str, Any]) -> List[str]:
                     {
                         "entry": func_name.split(".")[-1],
                         "chain": chain,
-                        "purity": _calculate_purity(chain, result),
+                        "purity": _calculate_purity(
+                            chain, result, name_to_func=name_to_func
+                        ),
                     }
                 )
     if not pipelines:

code2llm/exporters/toon/helpers.py

@@ -93,7 +93,22 @@ def _hotspot_description(fi: FunctionInfo, fan_out: int) -> str:
     return f"calls {fan_out} functions"
 
 
-_WALK_EXCLUDE = {"venv", ".venv", "node_modules", "__pycache__", ".git", "dist", "build", ".tox", ".mypy_cache", "egg-info"}  # noqa: E501
+_WALK_EXCLUDE = {
+    "venv",
+    ".venv",
+    "node_modules",
+    "__pycache__",
+    ".git",
+    "dist",
+    "build",
+    ".tox",
+    ".mypy_cache",
+    "egg-info",
+    ".subactor",
+    ".worktrees",
+    ".planfile",
+    ".benchmarks",
+}
 
 
 def _fast_line_counts(pp: Path, result) -> Dict[str, int]:
@@ -103,7 +118,11 @@ def _fast_line_counts(pp: Path, result) -> Dict[str, int]:
         if not mi.file:
             continue
         try:
-            lc = len(Path(mi.file).read_text(encoding="utf-8", errors="ignore").splitlines())
+            lc = getattr(mi, "line_count", 0)
+            if not lc:
+                lc = len(
+                    Path(mi.file).read_text(encoding="utf-8", errors="ignore").splitlines()
+                )
             counts[mi.file] = lc
             counts[str(Path(mi.file).relative_to(pp))] = lc
         except Exception:

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "code2llm"
-version = "0.5.182"
+version = "0.5.183"
 description = "High-performance Python code flow analysis with optimized TOON format - CFG, DFG, call graphs, and intelligent code queries"
 readme = "README.md"
 requires-python = ">=3.10"

tests/test_perf_exporters.py

@@ -0,0 +1,74 @@
+"""Tests for performance optimizations in exporters and code2logic dispatch."""
+
+from pathlib import Path
+from types import SimpleNamespace
+from code2llm.core.models import AnalysisResult, ModuleInfo, FunctionInfo
+from code2llm.exporters.toon.helpers import _fast_line_counts, _WALK_EXCLUDE
+from code2llm.exporters.flow_constants import EXCLUDE_PATTERNS, is_excluded_path
+from code2llm.exporters.map.utils import count_total_lines
+from code2llm.exporters.toon._render_section_helpers import _calculate_purity
+from code2llm.cli_exports.code2logic import _build_code2logic_cmd
+
+
+def test_fast_line_counts_reuses_module_line_count():
+    """Verify _fast_line_counts uses mi.line_count without reading non-existent file."""
+    result = AnalysisResult()
+    result.modules = {
+        "fake.mod": ModuleInfo(
+            name="fake.mod",
+            file="/non/existent/path/fake.py",
+            line_count=42,
+        )
+    }
+    counts = _fast_line_counts(Path("/non/existent"), result)
+    assert counts.get("/non/existent/path/fake.py") == 42
+
+
+def test_map_count_total_lines_reuses_line_count():
+    """Verify count_total_lines uses mi.line_count."""
+    result = AnalysisResult()
+    result.modules = {
+        "mod1": ModuleInfo(name="mod1", file="/path/a.py", line_count=100),
+        "mod2": ModuleInfo(name="mod2", file="/path/b.py", line_count=50),
+    }
+    total = count_total_lines(result, lambda p: False)
+    assert total == 150
+
+
+def test_calculate_purity_indexed():
+    """Verify _calculate_purity works with indexed name_to_func."""
+    result = AnalysisResult()
+    f1 = FunctionInfo(name="func1", qualified_name="m.func1", file="a.py", line=1)
+    f2 = FunctionInfo(name="func2", qualified_name="m.func2", file="a.py", line=10)
+    setattr(f2, "has_side_effects", True)
+    result.functions = {"m.func1": f1, "m.func2": f2}
+
+    name_to_func = {fi.name: fi for fi in result.functions.values()}
+    purity = _calculate_purity(["func1", "func2"], result, name_to_func=name_to_func)
+    assert purity == 0.5
+
+
+def test_build_code2logic_cmd_forwards_exclude():
+    """Verify _build_code2logic_cmd forwards args.exclude."""
+    args = SimpleNamespace(
+        verbose=False,
+        exclude=["vendor/**", "node_modules"],
+    )
+    cmd = _build_code2logic_cmd(args, Path("/src"), Path("/out"))
+    assert "--exclude" in cmd
+    idx = cmd.index("--exclude")
+    assert cmd[idx + 1 : idx + 3] == ["vendor/**", "node_modules"]
+
+
+def test_exclude_patterns_contain_subactor_worktrees():
+    """Verify standard exclusions contain .subactor, .worktrees, etc."""
+    assert ".subactor" in EXCLUDE_PATTERNS
+    assert ".worktrees" in EXCLUDE_PATTERNS
+    assert ".planfile" in EXCLUDE_PATTERNS
+    assert ".benchmarks" in EXCLUDE_PATTERNS
+
+    assert ".subactor" in _WALK_EXCLUDE
+    assert ".worktrees" in _WALK_EXCLUDE
+
+    assert is_excluded_path("/project/.subactor/cache/file.py")
+    assert is_excluded_path("/project/.worktrees/ticket-123/main.py")