FIX - Tree-sitter pkg version conflicts

.gitignore

@@ -19,5 +19,6 @@ __pycache__/
 .mypy_cache/
 .ruff_cache/
 
+**/node_modules/**
 # Claude Differ sentinel
 .claude-writing

README.md

@@ -62,7 +62,7 @@ the index is fast, free to build, and requires zero maintenance.
 
 ```bash
 # 1. Install
-pip install "git+https://github.com/darshil3011/codedrift[mcp]"
+pip install "codedrift[mcp] @ git+https://github.com/darshil3011/codedrift.git"
 
 # 2. Index your project
 cd /path/to/your/project

codedrift/languages/base.py

@@ -29,6 +29,88 @@ class ImportRef:
     import_line: str    # "from auth.jwt import validate_token"
 
 
+# ── tree-sitter API compatibility shim (0.24 ↔ 0.25+) ────────────────────────
+#
+# tree-sitter 0.25 changed every Node/Tree attribute from a property to a
+# method call (root_node(), kind(), start_position(), …) and dropped node.text.
+# _TSNode/_TSTree wrap the new API to expose the old property-based interface
+# so all language adapters continue to work without modification.
+
+class _TSNode:
+    """Wraps a tree-sitter ≥0.25 Node to expose the ≤0.24 property API."""
+
+    def __init__(self, node, source: bytes):
+        self._n = node
+        self._src = source
+
+    @property
+    def type(self) -> str:
+        return self._n.kind()
+
+    @property
+    def start_point(self):
+        p = self._n.start_position()
+        return (p.row, p.column)
+
+    @property
+    def end_point(self):
+        p = self._n.end_position()
+        return (p.row, p.column)
+
+    @property
+    def text(self) -> bytes:
+        br = self._n.byte_range()
+        return self._src[br.start:br.end]
+
+    @property
+    def children(self) -> list:
+        count = self._n.child_count()
+        return [_TSNode(self._n.child(i), self._src) for i in range(count)]
+
+    @property
+    def parent(self):
+        p = self._n.parent()
+        return _TSNode(p, self._src) if p is not None else None
+
+    def child_by_field_name(self, name: str):
+        n = self._n.child_by_field_name(name)
+        return _TSNode(n, self._src) if n is not None else None
+
+    def child(self, i: int):
+        n = self._n.child(i)
+        return _TSNode(n, self._src) if n is not None else None
+
+    def __eq__(self, other):
+        if isinstance(other, _TSNode):
+            return (self._n.start_byte() == other._n.start_byte() and
+                    self._n.end_byte() == other._n.end_byte())
+        return NotImplemented
+
+    def __hash__(self):
+        return hash((self._n.start_byte(), self._n.end_byte()))
+
+
+class _TSTree:
+    """Wraps a tree-sitter ≥0.25 Tree to expose the ≤0.24 property API."""
+
+    def __init__(self, tree, source: bytes):
+        self._t = tree
+        self._src = source
+
+    @property
+    def root_node(self) -> _TSNode:
+        return _TSNode(self._t.root_node(), self._src)
+
+
+def _wrap_tree(raw_tree, source: bytes):
+    """Wrap raw_tree in _TSTree when the ≥0.25 API is detected."""
+    if callable(getattr(raw_tree, "root_node", None)):
+        return _TSTree(raw_tree, source)
+    return raw_tree  # old API — pass through unchanged
+
+
+# ── Language adapter base class ───────────────────────────────────────────────
+
 class LanguageAdapter:
     """Base class for language-specific AST extraction."""
 
@@ -44,7 +126,10 @@ def _get_parser(self):
         return get_parser(self.language_name)
 
     def parse(self, source: bytes):
-        return self._get_parser().parse(source)
+        parser = self._get_parser()
+        # tree-sitter >= 0.25 renamed parse(bytes) → parse_bytes(bytes)
+        parse_fn = getattr(parser, "parse_bytes", parser.parse)
+        return _wrap_tree(parse_fn(source), source)
 
     def _node_text(self, node, source_lines: List[str]) -> str:
         return source_lines[node.start_point[0]] if source_lines else ""

codedrift/languages/javascript_lang.py

@@ -39,7 +39,10 @@ def _parser_for(self, filepath: str):
         return get_parser(lang)
 
     def parse_file(self, source: bytes, filepath: str):
-        return self._parser_for(filepath).parse(source)
+        parser = self._parser_for(filepath)
+        parse_fn = getattr(parser, "parse_bytes", parser.parse)
+        from .base import _wrap_tree
+        return _wrap_tree(parse_fn(source), source)
 
     def extract_functions(self, tree, source_lines: List[str], filepath: str) -> List[Symbol]:
         symbols = []