pytex-preprocessor 0.4.0__tar.gz → 0.4.1__tar.gz

{pytex_preprocessor-0.4.0 → pytex_preprocessor-0.4.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pytex-preprocessor
-Version: 0.4.0
+Version: 0.4.1
 Summary: Type-safe LaTeX document generation with Python
 Author-email: Frederik Beimgraben <frederik@beimgraben.net>
 Requires-Python: >=3.13

{pytex_preprocessor-0.4.0 → pytex_preprocessor-0.4.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pytex-preprocessor"
-version = "0.4.0"
+version = "0.4.1"
 authors = [
   { name="Frederik Beimgraben", email="frederik@beimgraben.net" },
 ]

{pytex_preprocessor-0.4.0 → pytex_preprocessor-0.4.1}/src/pytex/model/math.py

@@ -67,11 +67,46 @@ __all__ = [
 ]
 
 
+def _restrip(raw: Raw, how: str) -> Raw:
+    content = raw.content.lstrip() if how == "l" else raw.content.rstrip()
+    return Raw(content, raw.namespace, raw.allow_replacements)
+
+
+def _is_blank(node: TeX | str) -> bool:
+    return isinstance(node, Raw) and node.content.strip() == ""
+
+
+def _trim(body: TeX | str) -> TeX | str:
+    """Strip whitespace just inside the math delimiters.
+
+    TeX math mode ignores leading/trailing spaces and newlines, so this is
+    render-equivalent; it also avoids a blank line (`\\par`) sneaking inside the
+    delimiters, which is a hard LaTeX error. Only the outer boundary is touched
+    (interior whitespace can be significant, e.g. in `\\text{}` or alignments).
+    """
+    if isinstance(body, str):
+        return body.strip()
+    if isinstance(body, Raw):
+        return Raw(body.content.strip(), body.namespace, body.allow_replacements)
+    if isinstance(body, Concat):
+        elements = list(body.elements)
+        while elements and _is_blank(elements[0]):
+            elements.pop(0)
+        while elements and _is_blank(elements[-1]):
+            elements.pop()
+        if elements and isinstance(elements[0], Raw):
+            elements[0] = _restrip(elements[0], "l")
+        if elements and isinstance(elements[-1], Raw):
+            elements[-1] = _restrip(elements[-1], "r")
+        return Concat(*elements)
+    return body
+
+
 @Registry.add
 def Math(body: TeX | str) -> TeX:
     return Concat(
         ControlSequence("(", ()),
-        body,
+        _trim(body),
         ControlSequence(")", ()),
     )
 
@@ -79,14 +114,14 @@ def Math(body: TeX | str) -> TeX:
 @Registry.add
 def InlineMath(body: TeX | str) -> TeX:
     """Dollar-delimited inline math: ``$body$``."""
-    return Concat(Raw("$"), body, Raw("$"))
+    return Concat(Raw("$"), _trim(body), Raw("$"))
 
 
 @Registry.add
 def DisplayMath(body: TeX | str) -> TeX:
     return Concat(
         ControlSequence("[", ()),
-        body,
+        _trim(body),
         ControlSequence("]", ()),
     )
 

{pytex_preprocessor-0.4.0 → pytex_preprocessor-0.4.1}/src/pytex_analyze/optimize.py

@@ -130,10 +130,24 @@ def _native(raw: Raw) -> TeX:
     return raw
 
 
+_MATH_OPEN = re.compile(r"(\\[\[(])\s+")
+_MATH_CLOSE = re.compile(r"\s+(\\[\])])")
+
+
+def _strip_math_ws(text: str) -> str:
+    r"""Drop whitespace just inside ``\[..\]`` and ``\(..\)``.
+
+    TeX math mode ignores it, so two strings that differ only there render the
+    same — this lets the tokenizer accept a `DisplayMath`/`Math` candidate
+    (which trims that whitespace) as equivalent to the original.
+    """
+    return _MATH_CLOSE.sub(r"\1", _MATH_OPEN.sub(r"\1", text))
+
+
 def _tokenize(raw: Raw) -> TeX | None:
     """Split a `Raw` into literal text and the native nodes for the constructs
-    in `_TOKEN`. Returns `None` when nothing matches or the result would not
-    render identically."""
+    in `_TOKEN`. Returns `None` when nothing matches or the result would change
+    the rendered output (math-delimiter whitespace aside — TeX ignores that)."""
     content = raw.content
     namespace = pytex_namespace(raw.namespace or {})
     parts: list[TeX] = []
@@ -148,7 +162,11 @@ def _tokenize(raw: Raw) -> TeX | None:
     if cursor < len(content):
         parts.append(Raw(content[cursor:], allow_replacements=False))
     candidate = Concat(*(_optimize(part) for part in parts))
-    return candidate if candidate.rendered == raw.rendered else None
+    rendered = candidate.rendered
+    target = raw.rendered
+    if rendered == target or _strip_math_ws(rendered) == _strip_math_ws(target):
+        return candidate
+    return None
 
 
 def _token_node(

{pytex_preprocessor-0.4.0 → pytex_preprocessor-0.4.1}/src/pytex_preprocessor.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pytex-preprocessor
-Version: 0.4.0
+Version: 0.4.1
 Summary: Type-safe LaTeX document generation with Python
 Author-email: Frederik Beimgraben <frederik@beimgraben.net>
 Requires-Python: >=3.13