Function generate_function_charts

scripts_python/profiling_statistics.py:463–575

(conn: sqlite3.Connection, output_dir: Path, pbar: tqdm)

def generate_function_charts(conn: sqlite3.Connection, output_dir: Path, pbar: tqdm):
    cursor = conn.cursor()

    cursor.execute("""
        SELECT DISTINCT f.function_name, f.module_name, f.id, f.is_builtin, f.filename
        FROM functions f
        JOIN function_stats fs ON f.id = fs.function_id
        WHERE f.is_builtin = 0
    """)

    functions = cursor.fetchall()

    project_functions = []
    for func_name, module_name, func_id, is_builtin, filename in functions:
        if is_builtin or not filename or filename == "~":
            continue
        if "site-packages" in filename or "/lib/" in filename or "/lib64/" in filename:
            continue
        project_functions.append((func_name, module_name, func_id))

    for func_name, module_name, func_id in project_functions:
        pbar.set_postfix_str(f"Function: {module_name or 'unknown'}_{func_name[:30]}")
        safe_func_name = (
            func_name.replace("/", "_").replace("<", "").replace(">", "").replace(":", "_")
        )
        safe_module_name = (
            (module_name or "unknown")
            .replace("/", "_")
            .replace("<", "")
            .replace(">", "")
            .replace(":", "_")
        )

        func_dir = output_dir / "functions" / f"{safe_module_name}_{safe_func_name}"
        func_dir.mkdir(parents=True, exist_ok=True)

        cursor.execute("SELECT call_count FROM function_stats WHERE function_id = ?", (func_id,))
        call_counts = np.array([row[0] for row in cursor.fetchall()])
        if len(call_counts) > 0:
            create_histogram(
                call_counts,
                f"Call Count Distribution - {func_name}",
                "Call Count",
                func_dir / "hist_call_count.png",
            )

        cursor.execute(
            "SELECT primitive_call_count FROM function_stats WHERE function_id = ?", (func_id,)
        )
        prim_counts = np.array([row[0] for row in cursor.fetchall()])
        if len(prim_counts) > 0:
            create_histogram(
                prim_counts,
                f"Primitive Call Count Distribution - {func_name}",
                "Primitive Call Count",
                func_dir / "hist_primitive_call_count.png",
            )