Method parse_bytes

code_review_graph/parser.py:946–1045

Parse pre-read bytes and return extracted nodes and edges. This avoids re-reading the file from disk, eliminating TOCTOU gaps when the caller has already read the bytes (e.g. for hashing).

(self, path: Path, source: bytes)

        return self.parse_bytes(path, source)

    def parse_bytes(self, path: Path, source: bytes) -> tuple[list[NodeInfo], list[EdgeInfo]]:
        """Parse pre-read bytes and return extracted nodes and edges.

        This avoids re-reading the file from disk, eliminating TOCTOU gaps
        when the caller has already read the bytes (e.g. for hashing).
        """
        language = self.detect_language(path)
        if not language:
            return [], []

        # Vue SFCs: parse with vue parser, then delegate script blocks to JS/TS
        if language == "vue":
            return self._parse_vue(path, source)

        # Svelte SFCs: same approach as Vue — extract <script> blocks
        if language == "svelte":
            return self._parse_svelte(path, source)

        # Jupyter notebooks: extract code cells and parse as Python
        if language == "notebook":
            return self._parse_notebook(path, source)

        # Databricks .py notebook exports.  The header is ALWAYS the very
        # first line, but the file may have CRLF line endings on Windows
        # (git's core.autocrlf=true default).  Match the first line robustly
        # after stripping any trailing ``\r`` so the detection works on both
        # platforms.  See issue #239.
        if language == "python":
            first_newline = source.find(b"\n")
            first_line = (
                source[:first_newline].rstrip(b"\r")
                if first_newline != -1
                else source.rstrip(b"\r")
            )
            if first_line == b"# Databricks notebook source":
                return self._parse_databricks_py_notebook(path, source)

        # ReScript: regex-based parser (no tree-sitter grammar bundled).
        if language == "rescript":
            return self._parse_rescript(path, source)

        # SQL: dedicated parser — tree-sitter for tables/views/functions +
        # regex fallback for CREATE PROCEDURE (unsupported by the grammar).
        if language == "sql":
            return self._parse_sql(path, source)

        parser = self._get_parser(language)
        if not parser:
            return [], []

        tree = parser.parse(source)
        nodes: list[NodeInfo] = []
        edges: list[EdgeInfo] = []
        file_path_str = str(path)

        # File node
        test_file = _is_test_file(file_path_str)
        nodes.append(NodeInfo(