Update examples, fix table cells with lines

marker/processors/llm/llm_table.py

@@ -233,7 +233,7 @@ def parse_html_table(self, html_text: str, block: Block, page: PageGroup) -> Lis
                 cell_polygon = PolygonBox.from_bbox(cell_bbox)
 
                 cell_obj = TableCell(
-                    text=cell_text,
+                    text_lines=[cell_text],
                     row_id=i,
                     col_id=cur_col,
                     rowspan=rowspan,

marker/processors/table.py

@@ -112,7 +112,7 @@ def __call__(self, document: Document):
 
                     cell_block = TableCell(
                         polygon=cell_polygon,
-                        text=self.finalize_cell_text(cell),
+                        text_lines=self.finalize_cell_text(cell),
                         rowspan=cell.rowspan,
                         colspan=cell.colspan,
                         row_id=cell.row_id,
@@ -135,10 +135,16 @@ def __call__(self, document: Document):
                         page.structure.remove(child.id)
 
     def finalize_cell_text(self, cell: SuryaTableCell):
-        text = "\n".join([t["text"].strip() for t in cell.text_lines]) if cell.text_lines else ""
-        text = re.sub(r"(\s\.){2,}", "", text)  # Replace . . .
-        text = re.sub(r"\.{2,}", "", text)  # Replace ..., like in table of contents
-        return self.normalize_spaces(fix_text(text))
+        fixed_text = []
+        text_lines = cell.text_lines if cell.text_lines else []
+        for line in text_lines:
+            text = line["text"].strip()
+            if not text or text == ".":
+                continue
+            text = re.sub(r"(\s\.){2,}", "", text)  # Replace . . .
+            text = re.sub(r"\.{2,}", "", text)  # Replace ..., like in table of contents
+            fixed_text.append(self.normalize_spaces(fix_text(text)))
+        return fixed_text
 
     @staticmethod
     def normalize_spaces(text):

marker/renderers/markdown.py

@@ -16,6 +16,9 @@ def cleanup_text(full_text):
     full_text = re.sub(r'(\n\s){3,}', '\n\n', full_text)
     return full_text.strip()
 
+def get_text_with_br(element):
+    return ''.join(str(content) if content.name == 'br' else content.strip() for content in element.contents)
+
 
 class Markdownify(MarkdownConverter):
     def __init__(self, paginate_output, page_separator, inline_math_delimiters, block_math_delimiters, **kwargs):
@@ -78,7 +81,7 @@ def convert_table(self, el, text, convert_as_inline):
                     col_idx += 1
 
                 # Fill in grid
-                value = cell.get_text(strip=True).replace("\n", " ").replace("|", " ")
+                value = get_text_with_br(cell).replace("\n", " ").replace("|", " ")
                 rowspan = int(cell.get('rowspan', 1))
                 colspan = int(cell.get('colspan', 1))
 

marker/schema/blocks/tablecell.py

@@ -1,3 +1,5 @@
+from typing import List
+
 from marker.schema import BlockTypes
 from marker.schema.blocks import Block
 
@@ -9,14 +11,21 @@ class TableCell(Block):
     row_id: int
     col_id: int
     is_header: bool
-    text: str = ""
+    text_lines: List[str] | None = None
     block_description: str = "A cell in a table."
 
+    @property
+    def text(self):
+        return "\n".join(self.text_lines)
+
     def assemble_html(self, document, child_blocks, parent_structure=None):
         tag_cls = "th" if self.is_header else "td"
         tag = f"<{tag_cls}"
         if self.rowspan > 1:
             tag += f" rowspan={self.rowspan}"
         if self.colspan > 1:
             tag += f" colspan={self.colspan}"
-        return f"{tag}>{self.text}</{tag_cls}>"
+        if self.text_lines is None:
+            self.text_lines = []
+        text = "<br>".join(self.text_lines)
+        return f"{tag}>{text}</{tag_cls}>"