code2logic 1.0.45__tar.gz → 1.0.47__tar.gz

{code2logic-1.0.45 → code2logic-1.0.47}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code2logic
-Version: 1.0.45
+Version: 1.0.47
 Summary: Code2Logic - Source code to logical representation converter for LLM analysis, featuring Tree-sitter parsing, dependency graph analysis, and multi-language support.
 License: Apache-2.0
 License-File: LICENSE

{code2logic-1.0.45 → code2logic-1.0.47}/code2logic/__init__.py

@@ -18,7 +18,7 @@ Example:
     >>> print(output)
 """
 
-__version__ = "1.0.45"
+__version__ = "1.0.47"
 __author__ = "Softreck"
 __email__ = "info@softreck.dev"
 __license__ = "MIT"

{code2logic-1.0.45 → code2logic-1.0.47}/code2logic/analyzer.py

@@ -5,6 +5,8 @@ Provides the high-level API for analyzing codebases.
 """
 
 import logging
+import os
+import subprocess
 import sys
 import time
 from collections import defaultdict
@@ -108,6 +110,7 @@ class ProjectAnalyzer:
         verbose: bool = False,
         include_private: bool = False,
         enable_similarity: bool = True,
+        respect_gitignore: bool = True,
     ):
         """
         Initialize the project analyzer.
@@ -123,6 +126,7 @@ class ProjectAnalyzer:
         self.verbose = verbose
         self.include_private = include_private
         self.enable_similarity = enable_similarity
+        self.respect_gitignore = respect_gitignore
         self.modules: List[ModuleInfo] = []
         self.languages: Dict[str, int] = defaultdict(int)
 
@@ -218,63 +222,191 @@ class ProjectAnalyzer:
 
     def _scan_files(self):
         """Scan and parse all source files."""
-        for fp in self.root_path.rglob('*'):
-            if not fp.is_file():
-                continue
+        scan_start = time.time()
+        files_seen = 0
+        files_parsed = 0
+        files_matched = 0
+        scan_progress_every = 500
+
+        git_file_list: Optional[List[Path]] = None
+        if self.respect_gitignore:
+            git_file_list = self._get_git_nonignored_files()
+            if git_file_list is not None and self.verbose:
+                log.info("Using git file list (non-ignored): files=%d", len(git_file_list))
+
+        if git_file_list is not None:
+            for fp in git_file_list:
+                files_seen += 1
+                filename = fp.name
+
+                if filename in self.IGNORE_FILES:
+                    continue
 
-            # Skip ignored directories
-            if any(d in fp.parts for d in self.IGNORE_DIRS):
-                continue
+                ext = fp.suffix.lower()
+                language = self.LANGUAGE_EXTENSIONS.get(ext)
+                if language is None and ext == '':
+                    try:
+                        with fp.open('r', encoding='utf-8', errors='ignore') as f:
+                            language = self._language_from_shebang(f.readline())
+                    except Exception:
+                        language = None
 
-            # Skip ignored files
-            if fp.name in self.IGNORE_FILES:
-                continue
+                if language is None:
+                    continue
+
+                files_matched += 1
+                self.languages[language] += 1
+
+                if self.verbose and files_seen > 0 and (files_seen % scan_progress_every) == 0:
+                    log.info(
+                        "Scan progress: seen=%d matched=%d parsed=%d modules=%d time=%.2fs",
+                        files_seen,
+                        files_matched,
+                        files_parsed,
+                        len(self.modules),
+                        time.time() - scan_start,
+                    )
 
-            ext = fp.suffix.lower()
-            language = self.LANGUAGE_EXTENSIONS.get(ext)
-            if language is None and ext == '':
                 try:
-                    with fp.open('r', encoding='utf-8', errors='ignore') as f:
-                        language = self._language_from_shebang(f.readline())
+                    content = fp.read_text(encoding='utf-8', errors='ignore')
                 except Exception:
-                    language = None
+                    continue
 
-            if language is None:
-                continue
+                try:
+                    rel_path = str(fp.relative_to(self.root_path))
+                except Exception:
+                    rel_path = str(fp)
 
-            self.languages[language] += 1
+                module = None
+                try:
+                    if self.ts_parser and self.ts_parser.is_available(language):
+                        module = self.ts_parser.parse(rel_path, content, language)
+                except Exception as e:
+                    if self.verbose:
+                        log.debug("Tree-sitter parser failed for %s: %s", rel_path, e)
+
+                if module is None:
+                    try:
+                        module = self.fallback_parser.parse(rel_path, content, language)
+                    except Exception as e:
+                        if self.verbose:
+                            log.debug("Fallback parser failed for %s: %s", rel_path, e)
+                        continue
+
+                if module:
+                    files_parsed += 1
+                    try:
+                        module.file_bytes = fp.stat().st_size
+                    except Exception:
+                        module.file_bytes = len(content.encode('utf-8', errors='ignore'))
+                    self.modules.append(module)
 
-            # Read file
-            try:
-                content = fp.read_text(encoding='utf-8', errors='ignore')
-            except Exception:
-                continue
+        else:
+            for root, dirnames, filenames in os.walk(self.root_path):
+                dirnames[:] = [d for d in dirnames if d not in self.IGNORE_DIRS]
+                for filename in filenames:
+                    files_seen += 1
+                    fp = Path(root) / filename
+
+                    if filename in self.IGNORE_FILES:
+                        continue
+
+                    ext = fp.suffix.lower()
+                    language = self.LANGUAGE_EXTENSIONS.get(ext)
+                    if language is None and ext == '':
+                        try:
+                            with fp.open('r', encoding='utf-8', errors='ignore') as f:
+                                language = self._language_from_shebang(f.readline())
+                        except Exception:
+                            language = None
+
+                    if language is None:
+                        continue
+
+                    files_matched += 1
+                    self.languages[language] += 1
+
+                    if self.verbose and files_seen > 0 and (files_seen % scan_progress_every) == 0:
+                        log.info(
+                            "Scan progress: seen=%d matched=%d parsed=%d modules=%d time=%.2fs",
+                            files_seen,
+                            files_matched,
+                            files_parsed,
+                            len(self.modules),
+                            time.time() - scan_start,
+                        )
+
+                    try:
+                        content = fp.read_text(encoding='utf-8', errors='ignore')
+                    except Exception:
+                        continue
+
+                    try:
+                        rel_path = str(fp.relative_to(self.root_path))
+                    except Exception:
+                        rel_path = str(fp)
+
+                    module = None
+                    try:
+                        if self.ts_parser and self.ts_parser.is_available(language):
+                            module = self.ts_parser.parse(rel_path, content, language)
+                    except Exception as e:
+                        if self.verbose:
+                            log.debug("Tree-sitter parser failed for %s: %s", rel_path, e)
+
+                    if module is None:
+                        try:
+                            module = self.fallback_parser.parse(rel_path, content, language)
+                        except Exception as e:
+                            if self.verbose:
+                                log.debug("Fallback parser failed for %s: %s", rel_path, e)
+                            continue
+
+                    if module:
+                        files_parsed += 1
+                        try:
+                            module.file_bytes = fp.stat().st_size
+                        except Exception:
+                            module.file_bytes = len(content.encode('utf-8', errors='ignore'))
+                        self.modules.append(module)
 
-            rel_path = str(fp.relative_to(self.root_path))
+        if self.verbose:
+            log.info(
+                "Scan finished: seen=%d matched=%d parsed=%d modules=%d time=%.2fs",
+                files_seen,
+                files_matched,
+                files_parsed,
+                len(self.modules),
+                time.time() - scan_start,
+            )
 
-            # Try Tree-sitter first, then fallback
-            module = None
-            try:
-                if self.ts_parser and self.ts_parser.is_available(language):
-                    module = self.ts_parser.parse(rel_path, content, language)
-            except Exception as e:
-                if self.verbose:
-                    log.debug("Tree-sitter parser failed for %s: %s", rel_path, e)
+    def _get_git_nonignored_files(self) -> Optional[List[Path]]:
+        """Return list of non-ignored files according to git, or None if unavailable."""
+        git_dir = self.root_path / '.git'
+        if not git_dir.exists():
+            return None
 
-            if module is None:
-                try:
-                    module = self.fallback_parser.parse(rel_path, content, language)
-                except Exception as e:
-                    if self.verbose:
-                        log.debug("Fallback parser failed for %s: %s", rel_path, e)
-                    continue
+        try:
+            proc = subprocess.run(
+                ['git', '-C', str(self.root_path), 'ls-files', '-co', '--exclude-standard'],
+                check=False,
+                stdout=subprocess.PIPE,
+                stderr=subprocess.DEVNULL,
+                text=True,
+            )
+        except Exception:
+            return None
 
-            if module:
-                try:
-                    module.file_bytes = fp.stat().st_size
-                except Exception:
-                    module.file_bytes = len(content.encode('utf-8', errors='ignore'))
-                self.modules.append(module)
+        if proc.returncode != 0:
+            return None
+
+        files: List[Path] = []
+        for line in (proc.stdout or '').splitlines():
+            rel = (line or '').strip()
+            if not rel:
+                continue
+            files.append(self.root_path / rel)
+        return files
 
     def _detect_entrypoints(self) -> List[str]:
         """Detect project entry points."""

{code2logic-1.0.45 → code2logic-1.0.47}/code2logic/cli.py

@@ -510,6 +510,7 @@ def _code2logic_llm_cli(argv: list[str]) -> None:
 
 
 def main(argv=None):
+    cli_start = time.time()
     parser = argparse.ArgumentParser(
         description='Analyze source code and generate logical representations',
         formatter_class=argparse.RawDescriptionHelpFormatter
@@ -644,6 +645,11 @@ Detail levels (columns in csv/json/yaml):
         action='store_true',
         help='Disable Tree-sitter (use fallback parser)'
     )
+    parser.add_argument(
+        '--no-gitignore',
+        action='store_true',
+        help='Do not respect .gitignore (scan all files under path)'
+    )
     parser.add_argument(
         '--no-similarity',
         action='store_true',
@@ -828,6 +834,7 @@ Detail levels (columns in csv/json/yaml):
         use_treesitter=not args.no_treesitter,
         verbose=args.verbose or args.debug,
         enable_similarity=not args.no_similarity,
+        respect_gitignore=not args.no_gitignore,
     )
     project = analyzer.analyze()
     analyze_time = time.time() - analyze_start

{code2logic-1.0.45 → code2logic-1.0.47}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "code2logic"
-version = "1.0.45"
+version = "1.0.47"
 description = "Code2Logic - Source code to logical representation converter for LLM analysis, featuring Tree-sitter parsing, dependency graph analysis, and multi-language support."
 readme = "README.md"
 license = "Apache-2.0"