@haaaiawd/anws 2.3.0 → 2.4.0

package/templates_en/.agents/skills/nexus-query/scripts/extract_ast.py

@@ -0,0 +1,706 @@
+#!/usr/bin/env python3
+"""
+extract_ast.py — Multi-language repository AST structure extractor
+
+Purpose: Extract module/class/function structure from a code repository using Tree-sitter and output JSON to stdout
+Supports: Python, JavaScript, TypeScript, TSX, Java, Go, Rust, C#, C/C++, Kotlin, Ruby, Swift, PHP, Lua ...
+Usage: python extract_ast.py <repo_path> [--max-nodes 500]
+"""
+
+import sys
+import json
+import argparse
+from pathlib import Path
+from typing import Any, Optional, cast
+
+
+EXCLUDE_DIRS = {'.git', '__pycache__', '.venv', 'venv', 'node_modules',
+                'dist', 'build', '.mypy_cache', '.pytest_cache', 'site-packages',
+                '.nexus-map', '.tox', '.eggs', 'target', 'cmake-build-debug',
+                '.vs', 'out', '_build', 'vendor', '.ruff_cache', '.godot',
+                '.idea', '.vscode', '.nox'}
+
+EXCLUDE_FILE_SUFFIXES = ('.import', '.vulkan.cache')
+
+# ── Built-in language config: load from languages.json in the same directory ────────────────────
+_LANGUAGES_JSON = Path(__file__).parent / 'languages.json'
+
+
+def _load_builtin_languages() -> tuple[dict[str, str], dict[str, dict[str, str]], dict[str, str]]:
+    """Load built-in extension mappings, Tree-sitter queries, and known unsupported extensions from languages.json."""
+    try:
+        data = json.loads(_LANGUAGES_JSON.read_text(encoding='utf-8'))
+    except (FileNotFoundError, json.JSONDecodeError) as exc:
+        sys.stderr.write(f"[ERROR] Failed to load {_LANGUAGES_JSON}: {exc}\n")
+        sys.exit(1)
+
+    extensions: dict[str, str] = data.get('extensions', {})
+    raw_queries: dict[str, dict[str, str]] = data.get('queries', {})
+    unsupported: dict[str, str] = data.get('unsupported_extensions', {})
+
+    # Normalize queries: ensure each language has struct and imports keys
+    queries: dict[str, dict[str, str]] = {}
+    for lang, parts in raw_queries.items():
+        queries[lang] = {
+            'struct': parts.get('struct', ''),
+            'imports': parts.get('imports', ''),
+        }
+
+    return extensions, queries, unsupported
+
+
+BUILTIN_EXTENSION_MAP, BUILTIN_LANG_QUERIES, BUILTIN_KNOWN_UNSUPPORTED_EXTENSIONS = (
+    _load_builtin_languages()
+)
+
+
+def _should_skip_path(repo_path: Path, path: Path) -> bool:
+    rel_path = path.relative_to(repo_path)
+    if any(part in EXCLUDE_DIRS for part in rel_path.parts):
+        return True
+    if path.is_file() and any(path.name.endswith(suffix) for suffix in EXCLUDE_FILE_SUFFIXES):
+        return True
+    return False
+
+
+def write_filtered_file_tree(repo_path: Path, output_path: Path) -> None:
+    lines: list[str] = []
+    for path in sorted(repo_path.rglob('*')):
+        if _should_skip_path(repo_path, path):
+            continue
+        rel_path = path.relative_to(repo_path).as_posix()
+        suffix = '/' if path.is_dir() else ''
+        lines.append(rel_path + suffix)
+
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+    output_path.write_text('\n'.join(lines) + ('\n' if lines else ''), encoding='utf-8')
+
+def _normalize_extension(ext: str) -> str:
+    normalized = ext.strip().lower()
+    if not normalized:
+        raise ValueError('extension must not be empty')
+    if not normalized.startswith('.'):
+        normalized = f'.{normalized}'
+    return normalized
+
+
+def _copy_lang_queries(source: dict[str, dict[str, str]]) -> dict[str, dict[str, str]]:
+    return {
+        lang: {
+            'struct': query_parts.get('struct', ''),
+            'imports': query_parts.get('imports', ''),
+        }
+        for lang, query_parts in source.items()
+    }
+
+
+def _apply_cli_customizations(
+    cli_extensions: list[str] | None,
+    cli_queries: list[list[str]] | None,
+) -> tuple[
+    dict[str, str],
+    dict[str, dict[str, str]],
+    list[str],
+    dict[str, str],
+]:
+    """
+    Apply language customization from CLI arguments (--add-extension and --add-query).
+    Returns (extension_override, query_override, warnings)
+    """
+    extension_override: dict[str, str] = {}
+    query_override: dict[str, dict[str, str]] = {}
+    warnings: list[str] = []
+    custom_query_languages: dict[str, str] = {}
+
+    if cli_extensions:
+        for item in cli_extensions:
+            if '=' not in item:
+                warnings.append(f'ignored invalid extension mapping {item!r}, expected EXT=LANG')
+                continue
+            ext_part, lang_part = item.split('=', 1)
+            try:
+                ext = _normalize_extension(ext_part)
+                lang = lang_part.strip().lower()
+                if not lang:
+                    warnings.append(f'ignored empty language name for extension {ext_part!r}')
+                    continue
+                extension_override[ext] = lang
+            except ValueError as e:
+                warnings.append(f'ignored invalid extension {ext_part!r}: {e}')
+                continue
+
+    if cli_queries:
+        for query_item in cli_queries:
+            if len(query_item) != 3:
+                warnings.append(f'ignored malformed query: expected 3 parts, got {len(query_item)}')
+                continue
+            lang, query_type, query_str = query_item
+            lang = lang.strip().lower()
+            if not lang:
+                warnings.append('ignored empty language name in query')
+                continue
+            if query_type not in ('struct', 'imports'):
+                warnings.append(f'ignored unknown query type {query_type!r} for language {lang!r}')
+                continue
+
+            if lang not in query_override:
+                query_override[lang] = {'struct': '', 'imports': ''}
+            query_override[lang][query_type] = query_str
+            custom_query_languages[lang] = '<cli>'
+
+    return extension_override, query_override, warnings, custom_query_languages
+
+
+def _load_language_customizations(
+    repo_path: Path,
+    explicit_config_path: Optional[str],
+    cli_extension_override: dict[str, str],
+    cli_query_override: dict[str, dict[str, str]],
+    cli_warnings: list[str],
+    cli_custom_query_languages: dict[str, str],
+) -> tuple[
+    dict[str, str],
+    dict[str, dict[str, str]],
+    dict[str, str],
+    list[str],
+    list[str],
+    dict[str, str],
+]:
+    """
+    Load and merge language customization config.
+    
+    Priority: CLI --language-config > CLI --add-* args > built-in config
+    
+    Returns (extension_map, lang_queries, known_unsupported_extensions, warnings, loaded_config_paths, custom_query_languages)
+    """
+    extension_map = dict(BUILTIN_EXTENSION_MAP)
+    lang_queries = _copy_lang_queries(BUILTIN_LANG_QUERIES)
+    known_unsupported_extensions = dict(BUILTIN_KNOWN_UNSUPPORTED_EXTENSIONS)
+    warnings: list[str] = list(cli_warnings)
+    loaded_config_paths: list[str] = []
+    custom_query_languages: dict[str, str] = dict(cli_custom_query_languages)
+
+    # First merge customizations from CLI arguments
+    extension_map.update(cli_extension_override)
+    for lang, query_parts in cli_query_override.items():
+        if lang in lang_queries:
+            # Only override the provided parts
+            if query_parts.get('struct'):
+                lang_queries[lang]['struct'] = query_parts['struct']
+            if query_parts.get('imports'):
+                lang_queries[lang]['imports'] = query_parts['imports']
+        else:
+            lang_queries[lang] = query_parts
+
+    # Then load --language-config file (if provided), highest priority
+    if explicit_config_path:
+        config_path = Path(explicit_config_path)
+        resolved_path = config_path if config_path.is_absolute() else (repo_path / config_path)
+
+        try:
+            config_data = json.loads(resolved_path.read_text(encoding='utf-8'))
+        except FileNotFoundError:
+            warnings.append(f'language config not found: {resolved_path}')
+            return extension_map, lang_queries, known_unsupported_extensions, warnings, loaded_config_paths, custom_query_languages
+        except json.JSONDecodeError as exc:
+            warnings.append(f'language config parse error in {resolved_path}: {exc}')
+            return extension_map, lang_queries, known_unsupported_extensions, warnings, loaded_config_paths, custom_query_languages
+        except OSError as exc:
+            warnings.append(f'language config read error in {resolved_path}: {exc}')
+            return extension_map, lang_queries, known_unsupported_extensions, warnings, loaded_config_paths, custom_query_languages
+
+        if not isinstance(config_data, dict):
+            warnings.append(f'language config ignored because root value is not an object: {resolved_path}')
+            return extension_map, lang_queries, known_unsupported_extensions, warnings, loaded_config_paths, custom_query_languages
+
+        loaded_config_paths.append(str(resolved_path))
+
+        # Load extension mappings from --language-config
+        extensions = config_data.get('extensions', {})
+        if isinstance(extensions, dict):
+            for raw_ext, raw_lang in extensions.items():
+                if isinstance(raw_ext, str) and isinstance(raw_lang, str) and raw_lang.strip():
+                    try:
+                        ext = _normalize_extension(raw_ext)
+                        lang = raw_lang.strip().lower()
+                        extension_map[ext] = lang
+                        known_unsupported_extensions.pop(ext, None)
+                    except ValueError:
+                        pass
+
+        # Load queries from --language-config
+        queries = config_data.get('queries', {})
+        if isinstance(queries, dict):
+            for raw_lang, raw_query_parts in queries.items():
+                if isinstance(raw_lang, str) and raw_lang.strip() and isinstance(raw_query_parts, dict):
+                    lang = raw_lang.strip().lower()
+                    struct_query = raw_query_parts.get('struct', '')
+                    imports_query = raw_query_parts.get('imports', '')
+                    if isinstance(struct_query, str) and isinstance(imports_query, str):
+                        lang_queries[lang] = {
+                            'struct': struct_query,
+                            'imports': imports_query,
+                        }
+                        custom_query_languages[lang] = str(resolved_path)
+
+        # Load unsupported extensions from --language-config
+        unsupported_extensions = config_data.get('unsupported_extensions', {})
+        if isinstance(unsupported_extensions, dict):
+            for raw_ext, raw_lang in unsupported_extensions.items():
+                if isinstance(raw_ext, str) and isinstance(raw_lang, str) and raw_lang.strip():
+                    try:
+                        ext = _normalize_extension(raw_ext)
+                        lang = raw_lang.strip().lower()
+                        known_unsupported_extensions[ext] = lang
+                        extension_map.pop(ext, None)
+                    except ValueError:
+                        pass
+
+    return extension_map, lang_queries, known_unsupported_extensions, warnings, loaded_config_paths, custom_query_languages
+
+
+
+def _load_languages(
+    extension_map: dict[str, str],
+    lang_queries: dict[str, dict[str, str]],
+    requested: Optional[list[str]] = None,
+) -> dict[str, Any]:
+    """
+    Load Tree-sitter language objects and return a {lang_name: Language} dictionary.
+    Prefer tree-sitter-language-pack (160+ languages); fall back to a single-language package when unavailable.
+    """
+    try:
+        from tree_sitter_language_pack import get_language as _get
+
+        def get_language(name: str) -> Any:
+            return _get(cast(Any, name))
+    except ImportError:
+        # Python-only single-language fallback
+        try:
+            import tree_sitter_python
+            from tree_sitter import Language
+
+            def get_language(name: str) -> Any:
+                if name == 'python':
+                    return Language(tree_sitter_python.language())
+                raise LookupError(name)
+        except ImportError:
+            sys.stderr.write(
+                "[ERROR] Missing tree-sitter language support.\n"
+                "Please run: pip install tree-sitter-language-pack\n"
+            )
+            sys.exit(1)
+
+    targets = requested if requested else sorted(set(extension_map.values()) | set(lang_queries.keys()))
+    languages: dict[str, Any] = {}
+    for name in targets:
+        try:
+            languages[name] = get_language(name)
+        except (LookupError, KeyError):
+            # Language package not installed; skip gracefully
+            pass
+
+    if not languages:
+        sys.stderr.write("[ERROR] No available language parsers. Install tree-sitter-language-pack\n")
+        sys.exit(1)
+    return languages
+
+
+def _file_module_id(repo_path: Path, file_path: Path) -> str:
+    """Convert file path to dot-separated module ID.
+    Example: src/nexus/api/routes.py -> src.nexus.api.routes
+        src/core/parser.hpp   → src.core.parser
+    """
+    rel = file_path.relative_to(repo_path)
+    parts = list(rel.parts)
+    stem = Path(parts[-1]).stem  # Remove extension
+    parts[-1] = stem
+    # Python special case: merge __init__ into package path
+    if stem == '__init__' and len(parts) > 1:
+        parts = parts[:-1]
+    return '.'.join(parts) if parts else stem
+
+
+
+
+def extract_file(
+    repo_path: Path,
+    file_path: Path,
+    lang_name: str,
+    language: Any,
+    lang_queries: dict[str, dict[str, str]],
+) -> tuple[list[dict], list[dict], list[str]]:
+    """Parse a single source file and return (nodes, edges, errors)"""
+    from tree_sitter import Parser as TSParser, Query, QueryCursor
+
+    nodes: list[dict] = []
+    edges: list[dict] = []
+    errors: list[str] = []
+
+    try:
+        source = file_path.read_bytes()
+    except OSError as e:
+        errors.append(f"{file_path}: read error: {e}")
+        return nodes, edges, errors
+
+    try:
+        parser = TSParser(language)
+        tree = parser.parse(source)
+    except Exception as e:
+        errors.append(f"{file_path}: parse error: {e}")
+        return nodes, edges, errors
+
+    rel_path = str(file_path.relative_to(repo_path)).replace('\\', '/')
+    module_id = _file_module_id(repo_path, file_path)
+    line_count = source.count(b'\n') + 1
+
+    # Module node (file-level)
+    nodes.append({
+        'id': module_id,
+        'type': 'Module',
+        'label': module_id.split('.')[-1],
+        'path': rel_path,
+        'lines': line_count,
+        'lang': lang_name,
+    })
+
+    queries = lang_queries.get(lang_name, {})
+
+    # ── Structure: classes / functions ──────────────────────────────────────────
+    struct_q_text = queries.get('struct', '')
+    if struct_q_text.strip():
+        try:
+            struct_query = Query(language, struct_q_text)
+            class_ranges: list[tuple[int, int, str]] = []
+
+            for pattern_idx, captures in QueryCursor(struct_query).matches(tree.root_node):
+                capture_names = list(captures.keys())
+                is_class = any('class' in k for k in capture_names)
+                def_key = 'class.def' if is_class else 'func.def'
+                name_key = 'class.name' if is_class else 'func.name'
+
+                def_nodes = captures.get(def_key, [])
+                name_nodes = captures.get(name_key, [])
+                if not def_nodes or not name_nodes:
+                    continue
+
+                def_node = def_nodes[0]
+                name_node = name_nodes[0]
+                name = source[name_node.start_byte:name_node.end_byte].decode('utf-8', 'replace')
+
+                if is_class:
+                    node_id = f"{module_id}.{name}"
+                    nodes.append({
+                        'id': node_id,
+                        'type': 'Class',
+                        'label': name,
+                        'path': rel_path,
+                        'parent': module_id,
+                        'start_line': def_node.start_point[0] + 1,
+                        'end_line': def_node.end_point[0] + 1,
+                    })
+                    class_ranges.append((def_node.start_byte, def_node.end_byte, node_id))
+                    edges.append({'source': module_id, 'target': node_id, 'type': 'contains'})
+                else:
+                    parent_id = module_id
+                    for cls_start, cls_end, cls_id in class_ranges:
+                        if cls_start <= def_node.start_byte and def_node.end_byte <= cls_end:
+                            parent_id = cls_id
+                            break
+                    node_id = f"{parent_id}.{name}"
+                    nodes.append({
+                        'id': node_id,
+                        'type': 'Function',
+                        'label': name,
+                        'path': rel_path,
+                        'parent': parent_id,
+                        'start_line': def_node.start_point[0] + 1,
+                        'end_line': def_node.end_point[0] + 1,
+                    })
+                    edges.append({'source': parent_id, 'target': node_id, 'type': 'contains'})
+
+        except Exception as e:
+            errors.append(f"{file_path}: struct query error: {e}")
+
+    # ── Imports: imports edges ─────────────────────────────────────────
+    import_q_text = queries.get('imports', '')
+    if import_q_text.strip():
+        try:
+            import_query = Query(language, import_q_text)
+            for _pattern_idx, captures in QueryCursor(import_query).matches(tree.root_node):
+                for mod_node in captures.get('mod', []):
+                    target = source[mod_node.start_byte:mod_node.end_byte].decode('utf-8', 'replace').strip('"\'<> ')
+                    if target:
+                        edges.append({'source': module_id, 'target': target, 'type': 'imports'})
+        except Exception as e:
+            errors.append(f"{file_path}: import query error: {e}")
+
+    return nodes, edges, errors
+
+
+def collect_source_files(
+    repo_path: Path,
+    languages: dict[str, Any],
+    extension_map: dict[str, str],
+    known_unsupported_extensions: dict[str, str],
+) -> tuple[list[tuple[Path, str]], dict[str, int], dict[str, int], dict[str, int]]:
+    """Collect source files of all known languages in repo, skipping excluded directories.
+
+    Returns:
+    - [(file_path, lang_name)]
+    - supported_file_counts: {lang_name: file_count}
+    - known_unsupported_file_counts: {lang_name: file_count}
+    - configured_but_unavailable_file_counts: {lang_name: file_count}
+    """
+    files: list[tuple[Path, str]] = []
+    supported_file_counts: dict[str, int] = {}
+    known_unsupported_file_counts: dict[str, int] = {}
+    configured_but_unavailable_file_counts: dict[str, int] = {}
+
+    for p in repo_path.rglob('*'):
+        if not p.is_file():
+            continue
+        if _should_skip_path(repo_path, p):
+            continue
+
+        suffix = p.suffix.lower()
+        lang = extension_map.get(suffix)
+        if lang:
+            if lang in languages:
+                files.append((p, lang))
+                supported_file_counts[lang] = supported_file_counts.get(lang, 0) + 1
+            else:
+                configured_but_unavailable_file_counts[lang] = (
+                    configured_but_unavailable_file_counts.get(lang, 0) + 1
+                )
+            continue
+
+        unsupported_lang = known_unsupported_extensions.get(suffix)
+        if unsupported_lang:
+            known_unsupported_file_counts[unsupported_lang] = (
+                known_unsupported_file_counts.get(unsupported_lang, 0) + 1
+            )
+
+    return (
+        sorted(files, key=lambda x: x[0]),
+        supported_file_counts,
+        known_unsupported_file_counts,
+        configured_but_unavailable_file_counts,
+    )
+
+
+
+def apply_max_nodes(
+    nodes: list[dict],
+    edges: list[dict],
+    max_nodes: int,
+) -> tuple[list[dict], list[dict], bool, int]:
+    """
+    When node count exceeds max_nodes, keep Module/Class first and truncate Function nodes.
+    Returns (filtered_nodes, filtered_edges, truncated, truncated_count)
+    """
+    if len(nodes) <= max_nodes:
+        return nodes, edges, False, 0
+
+    priority_nodes = [n for n in nodes if n['type'] in ('Module', 'Class')]
+    func_nodes = [n for n in nodes if n['type'] == 'Function']
+
+    remaining_slots = max_nodes - len(priority_nodes)
+    if remaining_slots < 0:
+        kept_nodes = priority_nodes
+        truncated_count = len(func_nodes)
+    else:
+        kept_funcs = func_nodes[:remaining_slots]
+        kept_nodes = priority_nodes + kept_funcs
+        truncated_count = len(func_nodes) - len(kept_funcs)
+
+    kept_ids = {n['id'] for n in kept_nodes}
+    kept_edges = [
+        e for e in edges
+        if e['source'] in kept_ids or e['type'] == 'imports'
+    ]
+    return kept_nodes, kept_edges, True, truncated_count
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(
+        description='Extract AST structure from a multi-language repository'
+    )
+    parser.add_argument('repo_path', help='Target repository path')
+    parser.add_argument('--max-nodes', type=int, default=500,
+                        help='Max nodes in output (default: 500). Truncates Function nodes first.')
+    parser.add_argument(
+        '--add-extension',
+        action='append',
+        dest='add_extensions',
+        metavar='EXT=LANG',
+        help='Add extension mapping, e.g., .templ=templ. Can be used multiple times.',
+    )
+    parser.add_argument(
+        '--add-query',
+        action='append',
+        dest='add_queries',
+        nargs=3,
+        metavar=('LANG', 'TYPE', 'QUERY'),
+        help='Add/override a query for a language. TYPE is "struct" or "imports". Can be used multiple times.',
+    )
+    parser.add_argument(
+        '--language-config',
+        help='Optional JSON file that adds or overrides extension mappings and tree-sitter queries. Useful for complex configurations.',
+    )
+    parser.add_argument(
+        '--file-tree-out',
+        help='Optional output path for a filtered file tree (e.g. .nexus-map/raw/file_tree.txt). Uses the same exclude rules as AST collection.',
+    )
+    args = parser.parse_args()
+
+    repo_path = Path(args.repo_path).resolve()
+    if not repo_path.exists():
+        sys.stderr.write(f"[ERROR] repo_path not found: {repo_path}\n")
+        sys.exit(1)
+    if not (repo_path / '.git').exists():
+        sys.stderr.write(f"[WARNING] .git not found in {repo_path}, may not be a git repo\n")
+
+    if args.file_tree_out:
+        file_tree_path = Path(args.file_tree_out)
+        if not file_tree_path.is_absolute():
+            file_tree_path = repo_path / file_tree_path
+        write_filtered_file_tree(repo_path, file_tree_path.resolve())
+
+    # Process CLI customization arguments
+    cli_ext_override, cli_query_override, cli_warnings, cli_custom_query_languages = _apply_cli_customizations(
+        args.add_extensions,
+        args.add_queries,
+    )
+
+    # Load and merge config
+    (
+        extension_map,
+        lang_queries,
+        known_unsupported_extensions,
+        config_warnings,
+        loaded_config_paths,
+        custom_query_languages,
+    ) = _load_language_customizations(
+        repo_path,
+        args.language_config,
+        cli_ext_override,
+        cli_query_override,
+        cli_warnings,
+        cli_custom_query_languages,
+    )
+
+    languages = _load_languages(extension_map, lang_queries)
+    (
+        source_files,
+        supported_file_counts,
+        known_unsupported_file_counts,
+        configured_but_unavailable_file_counts,
+    ) = collect_source_files(
+        repo_path,
+        languages,
+        extension_map,
+        known_unsupported_extensions,
+    )
+
+    if not source_files:
+        sys.stderr.write(f"[WARNING] No supported source files found in {repo_path}\n")
+
+    all_nodes: list[dict] = []
+    all_edges: list[dict] = []
+    all_errors: list[str] = []
+    detected_langs: set[str] = set()
+    total_lines = 0
+    warnings: list[str] = list(config_warnings)
+    module_only_file_counts: dict[str, int] = {}
+    languages_with_structural_queries = sorted(
+        lang for lang, query_parts in lang_queries.items()
+        if query_parts.get('struct', '').strip()
+    )
+
+    for file_path, lang_name in source_files:
+        nodes, edges, errors = extract_file(
+            repo_path,
+            file_path,
+            lang_name,
+            languages[lang_name],
+            lang_queries,
+        )
+        all_nodes.extend(nodes)
+        all_edges.extend(edges)
+        all_errors.extend(errors)
+        if lang_name not in languages_with_structural_queries:
+            module_only_file_counts[lang_name] = module_only_file_counts.get(lang_name, 0) + 1
+        if nodes:
+            detected_langs.add(lang_name)
+            total_lines += nodes[0].get('lines', 0)
+
+    final_nodes, final_edges, truncated, truncated_count = apply_max_nodes(
+        all_nodes, all_edges, args.max_nodes
+    )
+
+    if known_unsupported_file_counts:
+        unsupported_summary = ', '.join(
+            f"{lang} ({count} files)"
+            for lang, count in sorted(known_unsupported_file_counts.items())
+        )
+        warnings.append(
+            "known unsupported languages present; downstream outputs must mark inferred sections explicitly: "
+            f"{unsupported_summary}"
+        )
+
+    if configured_but_unavailable_file_counts:
+        unavailable_summary = ', '.join(
+            f"{lang} ({count} files)"
+            for lang, count in sorted(configured_but_unavailable_file_counts.items())
+        )
+        warnings.append(
+            'some configured languages were detected in source files but no parser could be loaded: '
+            f'{unavailable_summary}'
+        )
+
+    if module_only_file_counts:
+        module_only_summary = ', '.join(
+            f"{lang} ({count} files)"
+            for lang, count in sorted(module_only_file_counts.items())
+        )
+        warnings.append(
+            "some languages were parsed with module-only coverage because no structural query template is bundled: "
+            f"{module_only_summary}"
+        )
+
+    if loaded_config_paths:
+        config_summary = ', '.join(loaded_config_paths)
+        warnings.append(f'custom language configuration loaded: {config_summary}')
+
+    result = {
+        'languages': sorted(detected_langs),
+        'stats': {
+            'total_files': len(source_files),
+            'total_lines': total_lines,
+            'parse_errors': len(all_errors),
+            'truncated': truncated,
+            'truncated_nodes': truncated_count,
+            'supported_file_counts': supported_file_counts,
+            'languages_with_structural_queries': languages_with_structural_queries,
+            'languages_with_custom_queries': sorted(custom_query_languages.keys()),
+            'module_only_file_counts': module_only_file_counts,
+            'known_unsupported_file_counts': known_unsupported_file_counts,
+            'configured_but_unavailable_file_counts': configured_but_unavailable_file_counts,
+            'custom_language_config_paths': loaded_config_paths,
+        },
+        'nodes': final_nodes,
+        'edges': final_edges,
+    }
+
+    if all_errors:
+        result['_errors'] = all_errors[:20]
+    if warnings:
+        result['warnings'] = warnings
+
+    print(json.dumps(result, ensure_ascii=False, indent=2))
+
+
+if __name__ == '__main__':
+    main()
+