Merge pull request #13 from bertsky/fixes

Various improvements

ocrd_page_to_alto/cli.py

@@ -1,8 +1,12 @@
 import click
 from .convert import OcrdPageAltoConverter
 from ocrd_utils import initLogging
+from ocrd.decorators import ocrd_loglevel
 
-@click.command()
+CONTEXT_SETTINGS = dict(help_option_names=['-h', '--help'])
+
+@click.command(context_settings=CONTEXT_SETTINGS)
+@ocrd_loglevel
 @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')
@@ -11,8 +15,10 @@
 @click.option('--dummy-word/--no-dummy-word', default=True, help='Whether to create a Word for TextLine that have TextEquiv/Unicode but no Word')
 @click.option('--textequiv-index', default=0, help='If multiple textequiv, use the n-th TextEquiv by @index')
 @click.option('--textequiv-fallback-strategy', default='last', type=click.Choice(['raise', 'first', 'last']), help="What to do if nth textequiv isn't available. 'raise' will lead to a runtime error, 'first' will use the first TextEquiv, 'last' will use the last TextEquiv on the element")
-@click.argument('filename')
-def main(check_words, check_border, skip_empty_lines, trailing_dash_to_hyp, dummy_textline, dummy_word, textequiv_index, textequiv_fallback_strategy, filename):
+@click.option('-O', '--output-file', default='-', help='Output filename (or "-" for standard output, the default)',
+              type=click.Path(dir_okay=False, writable=True, exists=False, allow_dash=True))
+@click.argument('filename',  type=click.Path(dir_okay=False, exists=True))
+def main(log_level, check_words, check_border, skip_empty_lines, trailing_dash_to_hyp, dummy_textline, dummy_word, textequiv_index, textequiv_fallback_strategy, output_file, filename):
     """
     Convert PAGE to ALTO
     """
@@ -29,7 +35,8 @@ def main(check_words, check_border, skip_empty_lines, trailing_dash_to_hyp, dumm
         textequiv_fallback_strategy=textequiv_fallback_strategy
     )
     converter.convert()
-    print(converter)
+    with open(1 if output_file == '-' else output_file, 'w') as output:
+        output.write(str(converter))
 
 if __name__ == '__main__':
     main() # pylint: disable=no-value-for-parameter

ocrd_page_to_alto/convert.py

@@ -283,7 +283,7 @@ def _convert_table(self, parent_alto, parent_page, level=0):
                 self._convert_textlines(textblock_alto, parent_page)
 
     def convert_text(self):
-        for reg_page in self.page_page.get_AllRegions(depth=1):
+        for reg_page in self.page_page.get_AllRegions(depth=0):
             reg_page_type = reg_page.__class__.__name__[0:-10] # len('RegionType') == 10
             reg_alto_type = REGION_PAGE_TO_ALTO[reg_page_type]
             if not reg_alto_type:

ocrd_page_to_alto/styles.py

@@ -122,6 +122,7 @@ def __init__(self):
     def set_alto_tag_from_type(self, reg_alto, reg_page):
         typ = reg_page.get_type() if hasattr(reg_page, 'get_type') else None
         if typ:
-            reg_alto.set('TYPE', typ)
+            if hasattr(reg_alto, 'TYPE'):
+                reg_alto.set('TYPE', typ)
             reg_alto.set('TAGREFS', self.get_id(label=typ))