Merge pull request #459 from VikParuchuri/dev

Fix error with delimiters

marker/schema/blocks/equation.py

@@ -11,7 +11,17 @@ class Equation(Block):
     def assemble_html(self, child_blocks, parent_structure=None):
         if self.latex:
             html_out = f"<p block-type='{self.block_type}'>"
-            for el in self.parse_latex(html.escape(self.latex)):
+
+            try:
+                latex = self.parse_latex(html.escape(self.latex))
+            except ValueError as e:
+                # If we have mismatched delimiters, we'll treat it as a single block
+                # Strip the $'s from the latex
+                latex = [
+                    {"class": "block", "content": self.latex.replace("$", "")}
+                ]
+
+            for el in latex:
                 if el["class"] == "block":
                     html_out += f'<math display="block">{el["content"]}</math>'
                 elif el["class"] == "inline":
@@ -26,6 +36,9 @@ def assemble_html(self, child_blocks, parent_structure=None):
 
     @staticmethod
     def parse_latex(text: str):
+        if text.count("$") % 2 != 0:
+            raise ValueError("Mismatched delimiters in LaTeX")
+
         DELIMITERS = [
             ("$$", "block"),
             ("$", "inline")

marker/settings.py

@@ -23,7 +23,6 @@ class Settings(BaseSettings):
 
     # General models
     TORCH_DEVICE: Optional[str] = None  # Note: MPS device does not work for text detection, and will default to CPU
-    GOOGLE_API_KEY: Optional[str] = None
 
     @computed_field
     @property

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "marker-pdf"
-version = "1.2.2"
+version = "1.2.3"
 description = "Convert PDF to markdown with high speed and accuracy."
 authors = ["Vik Paruchuri <[email protected]>"]
 readme = "README.md"