toggleable behavior for empty lines

OCR-D · Apr 6, 2021 · cd0bb8d · cd0bb8d
1 parent eb256b7
commit cd0bb8d
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 9 deletions.
diff --git a/ocrd_page_to_alto/cli.py b/ocrd_page_to_alto/cli.py
@@ -5,13 +5,19 @@
 @click.command()
 @click.option('--check-words/--no-check-words', default=True, help='Check whether PAGE-XML contains any Words and fail if not')
 @click.option('--check-border/--no-check-border', default=True, help='Check whether PAGE-XML contains Border or PrintSpace')
+@click.option('--skip-empty-lines/--no-skip-empty-lines', default=False, help='Whether to omit or keep empty lines in PAGE-XML')
 @click.argument('filename')
-def main(check_words, check_border, filename):
+def main(check_words, check_border, skip_empty_lines, filename):
     """
     Convert PAGE to ALTO
     """
     initLogging()
-    converter = OcrdPageAltoConverter(page_filename=filename, check_words=check_words, check_border=check_border)
+    converter = OcrdPageAltoConverter(
+        page_filename=filename,
+        check_words=check_words,
+        check_border=check_border,
+        skip_empty_lines=skip_empty_lines,
+    )
     converter.convert()
     print(converter)
 

diff --git a/ocrd_page_to_alto/convert.py b/ocrd_page_to_alto/convert.py
@@ -36,14 +36,16 @@
 
 class OcrdPageAltoConverter():
 
-    def __init__(self, *, check_words=True, check_border=True, page_filename=None, page_etree=None, pcgts=None, logger=None):
+    def __init__(self, *, check_words=True, check_border=True, skip_empty_lines=False, page_filename=None, page_etree=None, pcgts=None, logger=None):
         """
         Keyword Args:
             check_words (boolean): Whether to check if PAGE-XML contains any words before conversion and fail if not
             check_border (boolean): Whether to abort if neither Border nor PrintSpace is defined
+            skip_empty_lines (boolean): Whether to omit empty lines completely (True) or create a placeholder empty String in ALTO (False)
         """
         if not (page_filename or page_etree or pcgts):
             raise ValueError("Must pass either pcgts, page_etree or page_filename to constructor")
+        self.skip_empty_lines = skip_empty_lines
         self.logger = logger if logger else getLogger('page-to-alto')
         if pcgts:
             self.page_pcgts = pcgts
@@ -161,17 +163,17 @@ def convert_metadata(self):
 
     def _convert_textlines(self, reg_alto, reg_page):
         for line_page in reg_page.get_TextLine():
+            is_empty_line = not(line_page.get_TextEquiv() and line_page.get_TextEquiv()[0].get_Unicode())
+            if is_empty_line and self.skip_empty_lines:
+                return
             line_alto = ET.SubElement(reg_alto, 'TextLine')
+            if is_empty_line:
+                word_alto_empty = ET.SubElement(line_alto, 'String')
+                word_alto_empty.set('CONTENT', '')
             set_alto_id_from_page_id(line_alto, line_page)
             set_alto_xywh_from_coords(line_alto, line_page)
             set_alto_shape_from_coords(line_alto, line_page)
-            is_empty_line = not(line_page.get_TextEquiv() and line_page.get_TextEquiv()[0].get_Unicode())
-            if not line_page.get_Word() and not is_empty_line:
-                raise ValueError("pc:TextLine '%s' has no pc:Word" % line_page.id)
             # XXX ALTO does not allow TextLine without at least one String
-            if is_empty_line:
-                word_alto_empty = ET.SubElement(line_alto, 'String')
-                word_alto_empty.set('CONTENT', '')
             for word_page in line_page.get_Word():
                 word_alto = ET.SubElement(line_alto, 'String')
                 set_alto_id_from_page_id(word_alto, word_page)