dedent before ast.parse for non-toplevel functions

Reviewed-on: #2

src/mdevaluate/__init__.py

@@ -16,7 +16,7 @@ from . import reader
 from . import system
 from . import utils
 from . import extra
-from .logging import logger
+from .logging_util import logger
 
 
 def open(

src/mdevaluate/autosave.py

@@ -5,7 +5,7 @@ from typing import Optional, Callable, Iterable
 
 import numpy as np
 from .checksum import checksum
-from .logging import logger
+from .logging_util import logger
 
 autosave_directory: Optional[str] = None
 load_autosave_data = False

src/mdevaluate/checksum.py

@@ -1,9 +1,14 @@
 import functools
 import hashlib
-from .logging import logger
+from .logging_util import logger
 from types import ModuleType, FunctionType
 import inspect
 from typing import Iterable
+import ast
+import io
+import tokenize
+import re
+import textwrap
 
 import numpy as np
 
@@ -28,16 +33,43 @@ def version(version_nr: int, calls: Iterable = ()):
     return decorator
 
 
-def strip_comments(s: str):
+def strip_comments(source: str) -> str:
-    """Strips comment lines and docstring from Python source string."""
+    """Removes docstrings, comments, and irrelevant whitespace from Python source code."""
-    o = ""
+
-    in_docstring = False
+    # Step 1: Remove docstrings using AST
-    for l in s.split("\n"):
+    def remove_docstrings(node):
-        if l.strip().startswith(("#", '"', "'")) or in_docstring:
+        if isinstance(node, (ast.FunctionDef, ast.AsyncFunctionDef, ast.ClassDef, ast.Module)):
-            in_docstring = l.strip().startswith(('"""', "'''")) + in_docstring == 1
+            if (doc := ast.get_docstring(node, clean=False)):
+                first_stmt = node.body[0]
+                if isinstance(first_stmt, ast.Expr) and isinstance(first_stmt.value, ast.Constant):
+                    node.body.pop(0)  # Remove the docstring entirely
+        for child in ast.iter_child_nodes(node):
+            remove_docstrings(child)
+
+    tree = ast.parse(textwrap.dedent(source))
+    remove_docstrings(tree)
+    code_without_docstrings = ast.unparse(tree)
+
+    # Step 2: Remove comments using tokenize
+    tokens = tokenize.generate_tokens(io.StringIO(code_without_docstrings).readline)
+    result = []
+    last_lineno = -1
+    last_col = 0
+
+    for toknum, tokval, (srow, scol), (erow, ecol), line in tokens:
+        if toknum == tokenize.COMMENT:
             continue
-        o += l + "\n"
+        if srow > last_lineno:
-    return o
+            last_col = 0
+        if scol > last_col:
+            result.append(" " * (scol - last_col))
+        result.append(tokval)
+        last_lineno, last_col = erow, ecol
+
+    code_no_comments = ''.join(result)
+
+    # Step 3: Remove empty lines (whitespace-only or truly blank)
+    return "\n".join([line for line in code_no_comments.splitlines() if line.strip() != ""])
 
 
 def checksum(*args, csum=None):

src/mdevaluate/coordinates.py

@@ -1,6 +1,6 @@
 from functools import partial, wraps
 from copy import copy
-from .logging import logger
+from .logging_util import logger
 from typing import Optional, Callable, List, Tuple
 
 import numpy as np

src/mdevaluate/correlation.py

@@ -431,9 +431,9 @@ def non_gaussian_parameter(
     trajectory: Coordinates = None,
     axis: str = "all",
 ) -> float:
-    """
+    r"""
     Calculate the non-Gaussian parameter.
-    ..math:
+    .. math:
       \alpha_2 (t) =
         \frac{3}{5}\frac{\langle r_i^4(t)\rangle}{\langle r_i^2(t)\rangle^2} - 1
     """

src/mdevaluate/pbc.py

@@ -7,7 +7,7 @@ from numpy.typing import ArrayLike, NDArray
 
 from itertools import product
 
-from .logging import logger
+from .logging_util import logger
 
 if TYPE_CHECKING:
     from mdevaluate.coordinates import CoordinateFrame

src/mdevaluate/reader.py

@@ -19,13 +19,13 @@ import MDAnalysis
 from scipy import sparse
 
 from .checksum import checksum
-from .logging import logger
+from .logging_util import logger
 from . import atoms
 from .coordinates import Coordinates
 
 CSR_ATTRS = ("data", "indices", "indptr")
 NOJUMP_MAGIC = 2016
-Group_RE = re.compile("\[ ([-+\w]+) \]")
+Group_RE = re.compile(r"\[ ([-+\w]+) \]")
 
 
 class NojumpError(Exception):

src/mdevaluate/utils.py

@@ -14,7 +14,7 @@ from scipy.ndimage import uniform_filter1d
 from scipy.interpolate import interp1d
 from scipy.optimize import curve_fit
 
-from .logging import logger
+from .logging_util import logger
 from .functions import kww, kww_1e