Processor helpers

ocrd_models/ocrd_models/ocrd_file.py

@@ -57,7 +57,7 @@ def __str__(self):
         #  return 'OcrdFile[' + '\n\t' + props + '\n\t]'
         props = ', '.join([
             '='.join([k, getattr(self, k) if getattr(self, k) else '---'])
-            for k in ['mimetype', 'ID', 'url', 'local_filename']
+            for k in ['fileGrp', 'ID', 'mimetype', 'url', 'local_filename']
         ])
         return '<OcrdFile ' + props + ']/> '
 

ocrd_utils/ocrd_utils/__init__.py

@@ -51,7 +51,7 @@
 
     FS-related utilities
 
-* ``is_string``, ``membername``, ``concat_padded``, ``nth_url_segment``, ``remove_non_path_from_url``, ``parse_json_string_or_file``
+* ``is_string``, ``membername``, ``concat_padded``, ``nth_url_segment``, ``remove_non_path_from_url``, ``parse_json_string_or_file``, ``assert_file_grp_cardinality``, ``make_file_id``
 
     String and OOP utilities
 
@@ -72,6 +72,7 @@
     'abspath',
     'adjust_canvas_to_rotation',
     'adjust_canvas_to_transposition',
+    'assert_file_grp_cardinality',
     'bbox_from_points',
     'bbox_from_xywh',
     'bbox_from_polygon',
@@ -88,6 +89,7 @@
     'nth_url_segment',
     'remove_non_path_from_url',
     'logging',
+    'make_file_id',
     'membername',
     'image_from_polygon',
     'parse_json_string_or_file',
@@ -809,3 +811,39 @@ def parse_json_string_or_file(value='{}'):    # pylint: disable=unused-argument
     if err:
         raise err       # pylint: disable=raising-bad-type
     return ret
+
+def assert_file_grp_cardinality(grps, n):
+    """
+    Assert that a string of comma-separated fileGrps contains exactly ``n`` entries.
+    """
+    if isinstance(grps, str):
+        grps = grps.split(',')
+    assert len(grps) == n, \
+            "Expected exactly %d output file group%s, but '%s' has %d" % (
+                n, '' if n == 1 else 's', grps, len(grps))
+
+
+def make_file_id(ocrd_file, output_file_grp):
+    """
+    Derive a new file ID for an output file from an existing input file ``ocrd_file``
+    and the name of the output file's ``fileGrp/@USE``, ``output_file_grp``.
+    If ``ocrd_file``'s ID contains the input file's fileGrp name, then replace it by ``output_file_grp``.
+    Otherwise use ``output_file_grp`` together with the position of ``ocrd_file`` within the input fileGrp
+    (as a fallback counter). Increment counter until there is no more ID conflict.
+    """
+    ret = ocrd_file.ID.replace(ocrd_file.fileGrp, output_file_grp)
+    if ret == ocrd_file.ID:
+        m = re.match(r'.*?(\d{3,}).*', ocrd_file.pageId or '')
+        if m:
+            n = m.group(1)
+        else:
+            ids = [f.ID for f in ocrd_file.mets.find_files(fileGrp=ocrd_file.fileGrp, mimetype=ocrd_file.mimetype)]
+            try:
+                n = ids.index(ocrd_file.ID)
+            except ValueError:
+                n = len(ids)
+        ret = concat_padded(output_file_grp, n)
+        while ocrd_file.mets.find_files(ID=ret):
+            n += 1
+            ret = concat_padded(output_file_grp, n)
+    return ret

tests/test_utils.py

@@ -8,6 +8,9 @@
 from ocrd_utils import (
     abspath,
 
+    assert_file_grp_cardinality,
+    make_file_id,
+
     bbox_from_points,
     bbox_from_xywh,
 
@@ -38,6 +41,22 @@
     MIME_TO_PIL, PIL_TO_MIME,
 )
 from ocrd_models.utils import xmllint_format
+from ocrd_models import OcrdFile, OcrdMets
+
+class MockOcrdFile(OcrdFile):
+    """
+    OcrdFile with mocked fileGrp access
+    """
+    @property
+    def fileGrp(self):
+        return self.__filegrp
+    @fileGrp.setter
+    def fileGrp(self, fileGrp):
+        self.__filegrp = fileGrp
+    def __init__(self, *args, fileGrp=None, ocrd_mets=None, **kwargs):
+        super(MockOcrdFile, self).__init__(*args, **kwargs)
+        self.fileGrp = fileGrp if fileGrp else None
+        self.ocrd_mets = ocrd_mets if ocrd_mets else None
 
 class TestUtils(TestCase):
 
@@ -212,6 +231,34 @@ def test_mime_ext(self):
         self.assertEqual(MIME_TO_PIL['image/jp2'], 'JP2')
         self.assertEqual(PIL_TO_MIME['JP2'], 'image/jp2')
 
+    def test_assert_file_grp_cardinality(self):
+        with self.assertRaisesRegex(AssertionError, "Expected exactly 5 output file groups, but '.'FOO', 'BAR'.' has 2"):
+            assert_file_grp_cardinality('FOO,BAR', 5)
+        with self.assertRaisesRegex(AssertionError, "Expected exactly 1 output file group, but '.'FOO', 'BAR'.' has 2"):
+            assert_file_grp_cardinality('FOO,BAR', 1)
+        assert_file_grp_cardinality('FOO,BAR', 2)
+
+    def test_mock_file(self):
+        f = MockOcrdFile(None, ID="MAX_0012", fileGrp='MAX')
+        self.assertEqual(f.fileGrp, 'MAX')
+
+    def test_make_file_id_simple(self):
+        self.assertEqual(make_file_id(MockOcrdFile(None, ID="MAX_0012", fileGrp='MAX'), 'FOO'), 'FOO_0012')
+
+    def test_make_file_id_mets(self):
+        mets = OcrdMets.empty_mets()
+        for i in range(1, 10):
+            mets.add_file('FOO', ID="FOO_%04d" % (i), mimetype="image/tiff")
+            mets.add_file('BAR', ID="BAR_%04d" % (i), mimetype="image/tiff")
+        self.assertEqual(make_file_id(mets.find_files(ID='BAR_0007')[0], 'FOO'), 'FOO_0007')
+        f = mets.add_file('ABC', ID="BAR_7", mimetype="image/tiff")
+        self.assertEqual(make_file_id(f, 'FOO'), 'FOO_0010')
+        mets.remove_file(fileGrp='FOO')
+        self.assertEqual(make_file_id(f, 'FOO'), 'FOO_0001')
+        mets.add_file('FOO', ID="FOO_0001", mimetype="image/tiff")
+        # print('\n'.join(['%s' % of for of in mets.find_files()]))
+        self.assertEqual(make_file_id(f, 'FOO'), 'FOO_0002')
+
 
 if __name__ == '__main__':
-    main()
+    main(__file__)