switch to skip_existing and fix html format outputs

convert.py

@@ -42,7 +42,7 @@ def process_single_pdf(args):
 
     out_folder = config_parser.get_output_folder(fpath)
     base_name = config_parser.get_base_filename(fpath)
-    if output_exists(out_folder, base_name):
+    if cli_options.get('skip_existing') and output_exists(out_folder, base_name):
         return
 
     try:

marker/config/parser.py

@@ -32,7 +32,7 @@ def common_options(fn):
                           help="Path to JSON file with additional configuration.")(fn)
         fn = click.option("--languages", type=str, default=None, help="Comma separated list of languages to use for OCR.")(fn)
         fn = click.option("--disable_multiprocessing", is_flag=True, default=False, help="Disable multiprocessing.")(fn)
-        fn = click.option('-l', is_flag=True, help="List available builders, processors and converters")(fn)
+        fn = click.option("--skip_existing", is_flag=True, default=False, help="Skip existing files.")(fn)
         return fn
 
     def generate_config_dict(self) -> Dict[str, any]:

marker/schema/blocks/form.py

@@ -12,5 +12,4 @@ class Form(Block):
     cells: List[SpanTableCell] | None = None
 
     def assemble_html(self, child_blocks, parent_structure=None):
-        return html_format(self.cells)
-
+        return str(html_format(self.cells))

marker/schema/blocks/table.py

@@ -13,7 +13,7 @@ class Table(Block):
 
     def assemble_html(self, child_blocks, parent_structure=None):
         if self.cells:
-            return html_format(self.cells)
+            return str(html_format(self.cells))
         else:
             template = super().assemble_html(child_blocks, parent_structure)
             return f"<p>{template}</p>"

marker/schema/blocks/toc.py

@@ -12,4 +12,4 @@ class TableOfContents(Block):
     cells: List[SpanTableCell] | None = None
 
     def assemble_html(self, child_blocks, parent_structure=None):
-        return html_format(self.cells)
+        return str(html_format(self.cells))