codeecho 1.0.1__tar.gz → 1.1.0__tar.gz

{codeecho-1.0.1 → codeecho-1.1.0}/CHANGELOG.md

@@ -1,5 +1,13 @@
 # Changelog
 
+## 1.1.0 - 2026-08-15
+
+### Added
+
+- Gosu (`.gs`, `.gsx`) is now parsed with a dedicated [tree-sitter-gosu](https://github.com/rcw3bb/tree-sitter-gosu)
+  grammar instead of the Java fallback, enabling accurate fragment extraction using Gosu-native node types
+  (`compilation_unit`, `function_declaration`, `constructor_declaration`, `class_declaration`).
+
 ## 1.0.1 - 2026-07-10
 
 ### Fixed

{codeecho-1.0.1 → codeecho-1.1.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: codeecho
-Version: 1.0.1
+Version: 1.1.0
 Summary: A developer tool that scans your codebase to detect and highlight ECHOES of duplicated or near-duplicated code so you can refactor toward cleaner, more maintainable designs.
 License: MIT License
          
@@ -37,18 +37,19 @@ Requires-Dist: logenrich (>=1.0.1,<2.0.0)
 Requires-Dist: rich (>=15.0.0,<16.0.0)
 Requires-Dist: tree-sitter (>=0.26.0,<0.27.0)
 Requires-Dist: tree-sitter-go (>=0.25.0,<0.26.0)
+Requires-Dist: tree-sitter-gosu (>=0.2.0,<0.3.0)
 Requires-Dist: tree-sitter-java (>=0.23.5,<0.24.0)
 Requires-Dist: tree-sitter-javascript (>=0.25.0,<0.26.0)
 Requires-Dist: tree-sitter-python (>=0.25.0,<0.26.0)
 Requires-Dist: tree-sitter-typescript (>=0.23.2,<0.24.0)
 Description-Content-Type: text/markdown
 
-# codeecho 1.0.1
+# codeecho 1.1.0
 
 > A developer tool that scans your codebase to detect and highlight **echoes** of duplicated or near-duplicated code, so you can refactor toward cleaner, more maintainable designs.
 
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
-[![Version](https://img.shields.io/badge/version-1.0.1-blue.svg)](CHANGELOG.md)
+[![Version](https://img.shields.io/badge/version-1.1.0-blue.svg)](CHANGELOG.md)
 
 ## Prerequisites
 
@@ -76,14 +77,14 @@ python -m codeecho <path> [options]
 
 ### Supported languages
 
-| Language | Extensions |
-|----------|-----------|
-| Python | `.py` |
-| JavaScript | `.js`, `.mjs`, `.cjs` |
-| TypeScript | `.ts`, `.tsx` |
-| Java | `.java` |
-| Go | `.go` |
-| Gosu | `.gs`, `.gsx` |
+| Language | Extensions | Grammar |
+|----------|-----------|--------|
+| Python | `.py` | [tree-sitter-python](https://github.com/tree-sitter/tree-sitter-python) |
+| JavaScript | `.js`, `.mjs`, `.cjs` | [tree-sitter-javascript](https://github.com/tree-sitter/tree-sitter-javascript) |
+| TypeScript | `.ts`, `.tsx` | [tree-sitter-typescript](https://github.com/tree-sitter/tree-sitter-typescript) |
+| Java | `.java` | [tree-sitter-java](https://github.com/tree-sitter/tree-sitter-java) |
+| Go | `.go` | [tree-sitter-go](https://github.com/tree-sitter/tree-sitter-go) |
+| Gosu | `.gs`, `.gsx` | [tree-sitter-gosu](https://github.com/rcw3bb/tree-sitter-gosu) |
 
 ### Arguments
 

{codeecho-1.0.1 → codeecho-1.1.0}/README.md

@@ -1,9 +1,9 @@
-# codeecho 1.0.1
+# codeecho 1.1.0
 
 > A developer tool that scans your codebase to detect and highlight **echoes** of duplicated or near-duplicated code, so you can refactor toward cleaner, more maintainable designs.
 
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
-[![Version](https://img.shields.io/badge/version-1.0.1-blue.svg)](CHANGELOG.md)
+[![Version](https://img.shields.io/badge/version-1.1.0-blue.svg)](CHANGELOG.md)
 
 ## Prerequisites
 
@@ -31,14 +31,14 @@ python -m codeecho <path> [options]
 
 ### Supported languages
 
-| Language | Extensions |
-|----------|-----------|
-| Python | `.py` |
-| JavaScript | `.js`, `.mjs`, `.cjs` |
-| TypeScript | `.ts`, `.tsx` |
-| Java | `.java` |
-| Go | `.go` |
-| Gosu | `.gs`, `.gsx` |
+| Language | Extensions | Grammar |
+|----------|-----------|--------|
+| Python | `.py` | [tree-sitter-python](https://github.com/tree-sitter/tree-sitter-python) |
+| JavaScript | `.js`, `.mjs`, `.cjs` | [tree-sitter-javascript](https://github.com/tree-sitter/tree-sitter-javascript) |
+| TypeScript | `.ts`, `.tsx` | [tree-sitter-typescript](https://github.com/tree-sitter/tree-sitter-typescript) |
+| Java | `.java` | [tree-sitter-java](https://github.com/tree-sitter/tree-sitter-java) |
+| Go | `.go` | [tree-sitter-go](https://github.com/tree-sitter/tree-sitter-go) |
+| Gosu | `.gs`, `.gsx` | [tree-sitter-gosu](https://github.com/rcw3bb/tree-sitter-gosu) |
 
 ### Arguments
 

{codeecho-1.0.1 → codeecho-1.1.0}/codeecho/__init__.py

@@ -7,7 +7,7 @@ maintainable designs.
 from env_dir_bootstrap import EnvDirBootstrap
 from logenrich import setup_logger
 
-__version__ = "1.0.1"
+__version__ = "1.1.0"
 
 _bootstrapper = EnvDirBootstrap(
     env_var="CODEECHO_CONFIG_DIR",

{codeecho-1.0.1 → codeecho-1.1.0}/codeecho/__main__.py

@@ -6,6 +6,8 @@ Invoked via::
     poetry run python -m codeecho [OPTIONS] PATH
 """
 
+import json
+import os
 import uuid
 from pathlib import Path
 
@@ -57,7 +59,9 @@ def _parse_types(types_str: str) -> set[int]:
     version=__version__, prog_name="codeecho", message="%(prog)s v%(version)s"
 )
 @click.argument(
-    "path", type=click.Path(exists=True, file_okay=False, dir_okay=True, path_type=Path)
+    "paths",
+    nargs=-1,
+    type=click.Path(exists=True, file_okay=True, dir_okay=True, path_type=Path),
 )
 @click.option(
     "--types",
@@ -117,7 +121,7 @@ def _parse_types(types_str: str) -> set[int]:
     help="Glob pattern(s) to exclude from scanning (repeatable).",
 )
 def main(  # pylint: disable=too-many-arguments,too-many-positional-arguments,too-many-locals,invalid-name
-    path: Path,
+    paths: tuple[Path, ...],
     types: str,
     threshold: float,
     output: str,
@@ -127,10 +131,10 @@ def main(  # pylint: disable=too-many-arguments,too-many-positional-arguments,to
     min_tokens: int,
     exclude: tuple[str, ...],
 ) -> None:
-    """Scan PATH for duplicate and near-duplicate code.
+    """Scan one or more PATH(s) for duplicate and near-duplicate code.
 
-    Generates JSON and/or HTML reports, then removes the intermediate session data
-    from the embedded database.
+    Each PATH may be a file or a directory.  Generates JSON and/or HTML reports,
+    then removes the intermediate session data from the embedded database.
     """
     _console.print(
         Panel(
@@ -139,6 +143,9 @@ def main(  # pylint: disable=too-many-arguments,too-many-positional-arguments,to
         )
     )
 
+    if not paths:
+        raise click.UsageError("At least one PATH is required.")
+
     detect_types = _parse_types(types)
     if output_dir is None:
         output_dir = Path.cwd() / "reports"
@@ -152,14 +159,27 @@ def main(  # pylint: disable=too-many-arguments,too-many-positional-arguments,to
 
     output_dir.mkdir(parents=True, exist_ok=True)
 
+    resolved = [p.resolve() for p in paths]
+
     db_path = Path(get_db_path(str(db_dir) if db_dir else None))
     with SessionDB(db_path=db_path) as session_db:
-        session_db.create_session(session_id, str(path.resolve()), config)
+        session_db.create_session(
+            session_id, json.dumps([str(p) for p in resolved]), config
+        )
 
         # ── Phase 1: File discovery ─────────────────────────────────────────
-        _console.print(f"[dim]Scanning:[/dim] [bold]{path.resolve()}[/bold]")
-        _ignore_file = IgnoreFile(Path(CONF_DIR) / ".ignore", base_dir=path.resolve())
-        files = scanner.scan(path, exclude_patterns=exclude, ignore_file=_ignore_file)
+        for p in resolved:
+            _console.print(f"[dim]Scanning:[/dim] [bold]{p}[/bold]")
+        try:
+            base_dir = Path(os.path.commonpath(resolved))
+            if not base_dir.is_dir():
+                base_dir = base_dir.parent
+        except ValueError:
+            base_dir = Path.cwd()
+        _ignore_file = IgnoreFile(Path(CONF_DIR) / ".ignore", base_dir=base_dir)
+        files = scanner.scan(
+            tuple(resolved), exclude_patterns=exclude, ignore_file=_ignore_file
+        )
         total_fragments = 0
 
         if not files:
@@ -177,7 +197,8 @@ def main(  # pylint: disable=too-many-arguments,too-many-positional-arguments,to
 
         result = ScanResult(
             session_id=session_id,
-            scan_path=str(path.resolve()),
+            version=__version__,
+            scan_path=[str(p) for p in resolved],
             files_scanned=len(files),
             fragments_extracted=total_fragments,
             type1_groups=cnt1,

{codeecho-1.0.1 → codeecho-1.1.0}/codeecho/extractor.py

@@ -43,9 +43,12 @@ _QUERIES: dict[tuple[str, str], str] = {
     ("Java", "function"): "[(method_declaration) (constructor_declaration)] @fragment",
     ("Java", "class"): "(class_declaration) @fragment",
     ("Java", "file"): "(program) @fragment",
-    ("Gosu", "function"): "[(method_declaration) (constructor_declaration)] @fragment",
+    (
+        "Gosu",
+        "function",
+    ): "[(function_declaration) (constructor_declaration)] @fragment",
     ("Gosu", "class"): "(class_declaration) @fragment",
-    ("Gosu", "file"): "(program) @fragment",
+    ("Gosu", "file"): "(compilation_unit) @fragment",
     ("Go", "function"): "[(function_declaration) (method_declaration)] @fragment",
     ("Go", "class"): "(type_declaration) @fragment",
     ("Go", "file"): "(source_file) @fragment",

{codeecho-1.0.1 → codeecho-1.1.0}/codeecho/logging.ini

@@ -1,5 +1,5 @@
 [loggers]
-keys=root
+keys=root,tree_sitter_gosu
 
 [handlers]
 keys=consoleHandler,fileHandler
@@ -11,6 +11,12 @@ keys=logFormatter,consoleFormatter
 level=INFO
 handlers=consoleHandler,fileHandler
 
+[logger_tree_sitter_gosu]
+level=CRITICAL
+handlers=
+qualname=tree_sitter_gosu
+propagate=0
+
 [handler_consoleHandler]
 class=StreamHandler
 formatter=consoleFormatter

{codeecho-1.0.1 → codeecho-1.1.0}/codeecho/models.py

@@ -40,11 +40,12 @@ class CloneGroup:
 
 
 @dataclass
-class ScanResult:
+class ScanResult:  # pylint: disable=too-many-instance-attributes
     """Summary statistics for a completed duplicate detection scan."""
 
     session_id: str
-    scan_path: str
+    version: str
+    scan_path: list[str]
     files_scanned: int
     fragments_extracted: int
     type1_groups: int

{codeecho-1.0.1 → codeecho-1.1.0}/codeecho/parser.py

@@ -1,9 +1,6 @@
 """
 Tree-sitter parser wrapper providing one cached :class:`tree_sitter.Parser` per language.
 
-Gosu (``.gs``) files are parsed using the Java grammar because Gosu is syntactically
-Java-like (methods, classes, blocks).
-
 :author: Ron Webb
 :since: 1.0.0
 """
@@ -16,9 +13,9 @@ from tree_sitter import Language, Parser, Tree
 _logger = logging.getLogger("codeecho.parser")
 
 
-def _build_language(
+def _build_language(  # pylint: disable=too-many-return-statements
     language_name: str,
-) -> Language | None:  # pylint: disable=too-many-return-statements
+) -> Language | None:
     """Instantiate and return the tree-sitter :class:`Language` for *language_name*."""
     match language_name:
         case "Python":
@@ -33,9 +30,13 @@ def _build_language(
             import tree_sitter_typescript as m  # pylint: disable=import-outside-toplevel
 
             return Language(m.language_typescript())
-        case "Java" | "Gosu":
+        case "Java":
             import tree_sitter_java as m  # pylint: disable=import-outside-toplevel
 
+            return Language(m.language())
+        case "Gosu":
+            import tree_sitter_gosu as m  # pylint: disable=import-outside-toplevel
+
             return Language(m.language())
         case "Go":
             import tree_sitter_go as m  # pylint: disable=import-outside-toplevel

{codeecho-1.0.1 → codeecho-1.1.0}/codeecho/reporter/html_reporter.py

@@ -79,7 +79,7 @@ _TEMPLATE: str = """<!DOCTYPE html>
 </style>
 </head>
 <body>
-<h1>CodeEcho</h1>
+<h1>CodeEcho <span style="color:var(--muted);font-size:1rem;font-weight:400">v{{ result.version }}</span></h1>
 <div class="subtitle">Clone Detection Report &mdash; {{ generated_at }}</div>
 
 <div class="toolbar">
@@ -149,7 +149,7 @@ _TEMPLATE: str = """<!DOCTYPE html>
 </details>
 {% endfor %}
 
-<footer>Scan path: {{ result.scan_path }} &bull; Session: {{ result.session_id }}</footer>
+<footer>Scan path{{ 's' if result.scan_path | length > 1 else '' }}: {{ result.scan_path | join(', ') }} &bull; Session: {{ result.session_id }}</footer>
 
 <script>
 function toggleAll(open) {

{codeecho-1.0.1 → codeecho-1.1.0}/codeecho/reporter/json_reporter.py

@@ -62,6 +62,7 @@ def write(
 
     payload = {
         "generated_at": datetime.now(timezone.utc).isoformat(),
+        "version": result.version,
         "session_id": result.session_id,
         "scan_path": result.scan_path,
         "summary": {

{codeecho-1.0.1 → codeecho-1.1.0}/codeecho/scanner.py

@@ -51,27 +51,44 @@ _DEFAULT_EXCLUDE_DIRS: frozenset[str] = frozenset(
 
 
 def scan(
-    root: Path,
+    paths: tuple[Path, ...],
     exclude_patterns: tuple[str, ...] = (),
     ignore_file: IgnoreFile | None = None,
 ) -> list[tuple[Path, str]]:
-    """Walk *root* recursively and return ``(path, language)`` pairs for supported files.
+    """Walk *paths* and return ``(path, language)`` pairs for supported source files.
+
+    Each entry in *paths* may be a file or a directory.  Directories are walked
+    recursively; files are checked for a supported extension and included directly.
+    Duplicate resolved paths across multiple entries are silently deduplicated.
 
-    :param root: Root directory to scan.
+    :param paths: One or more file or directory paths to scan.
     :param exclude_patterns: Glob patterns (fnmatch-style) whose matching paths are skipped.
     :param ignore_file: Optional gitignore-style :class:`~braincraft.IgnoreFile`; matched paths are skipped.
     :returns: List of ``(absolute_path, language_name)`` tuples.
     """
     results: list[tuple[Path, str]] = []
-    root = root.resolve()
-    _logger.debug("Scanning root: %s", root)
+    seen: set[Path] = set()
 
-    for entry in _walk(root, exclude_patterns, ignore_file):
-        suffix = entry.suffix.lower()
-        if language := EXTENSION_TO_LANGUAGE.get(suffix):
-            results.append((entry, language))
+    for path in paths:
+        path = path.resolve()
+        _logger.debug("Scanning: %s", path)
+        if path.is_file():
+            if path in seen or _matches_any(path, exclude_patterns):
+                continue
+            seen.add(path)
+            if ignore_file is not None and ignore_file.is_ignored(path):
+                _logger.debug("Ignored (ignore file): %s", path)
+                continue
+            if language := EXTENSION_TO_LANGUAGE.get(path.suffix.lower()):
+                results.append((path, language))
+        elif path.is_dir():
+            for entry in _walk(path, exclude_patterns, ignore_file):
+                if entry not in seen:
+                    seen.add(entry)
+                    if language := EXTENSION_TO_LANGUAGE.get(entry.suffix.lower()):
+                        results.append((entry, language))
 
-    _logger.debug("Found %d source files under %s", len(results), root)
+    _logger.debug("Found %d source files across %d path(s)", len(results), len(paths))
     return results
 
 

{codeecho-1.0.1 → codeecho-1.1.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "codeecho"
-version = "1.0.1"
+version = "1.1.0"
 description = "A developer tool that scans your codebase to detect and highlight ECHOES of duplicated or near-duplicated code so you can refactor toward cleaner, more maintainable designs."
 authors = [
     {name = "Ron Webb"}
@@ -20,7 +20,8 @@ dependencies = [
     "click (>=8.4.2,<9.0.0)",
     "rich (>=15.0.0,<16.0.0)",
     "jinja2 (>=3.1)",
-    "braincraft (>=1.2.0,<2.0.0)"
+    "braincraft (>=1.2.0,<2.0.0)",
+    "tree-sitter-gosu (>=0.2.0,<0.3.0)"
 ]