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

@@ -829,15 +829,21 @@ def make_file_id(ocrd_file, output_file_grp):
     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).
+    (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:
-            files = ocrd_file.mets.find_files(fileGrp=ocrd_file.fileGrp, mimetype=ocrd_file.mimetype)
-            n = files.index(ocrd_file)
+            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

@@ -41,7 +41,7 @@
     MIME_TO_PIL, PIL_TO_MIME,
 )
 from ocrd_models.utils import xmllint_format
-from ocrd_models import OcrdFile
+from ocrd_models import OcrdFile, OcrdMets
 
 class MockOcrdFile(OcrdFile):
     """
@@ -53,10 +53,10 @@ def fileGrp(self):
     @fileGrp.setter
     def fileGrp(self, fileGrp):
         self.__filegrp = fileGrp
-    def __init__(self, *args, fileGrp=None, **kwargs):
+    def __init__(self, *args, fileGrp=None, ocrd_mets=None, **kwargs):
         super(MockOcrdFile, self).__init__(*args, **kwargs)
-        if fileGrp:
-            self.fileGrp = fileGrp
+        self.fileGrp = fileGrp if fileGrp else None
+        self.ocrd_mets = ocrd_mets if ocrd_mets else None
 
 class TestUtils(TestCase):
 
@@ -242,10 +242,22 @@ def test_mock_file(self):
         f = MockOcrdFile(None, ID="MAX_0012", fileGrp='MAX')
         self.assertEqual(f.fileGrp, 'MAX')
 
-    def test_make_file_id(self):
-        self.assertEqual(make_file_id(MockOcrdFile(None, ID="MAX_0012", fileGrp='MAX'), 'FOO', 0), 'FOO_0012')
-        self.assertEqual(make_file_id(MockOcrdFile(None, ID="MAX_0012", fileGrp='BAR'), 'FOO', 0), 'FOO_0012')
-        self.assertEqual(make_file_id(MockOcrdFile(None, ID="MAXMAXMAX", fileGrp='BAR'), 'FOO', 11), 'FOO_0012')
+    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__':