Merge branch 'rename-concat-padded'

Conflicts:
	CHANGELOG.md

CHANGELOG.md

@@ -15,6 +15,10 @@ Fixed:
 
   * CLI: validate of ocrd-tool.json didn't parse JSON, #115
 
+Changed:
+
+  * `mets_file_id` in utils is now `concat_padded` and more flexible
+
 ## [0.3.0] - 2018-06-18
 
 Added

ocrd/utils.py

@@ -15,6 +15,7 @@
 import subprocess
 import logging
 import re
+import sys
 
 logging.basicConfig(level=logging.DEBUG)
 #  logging.getLogger('ocrd.resolver').setLevel(logging.INFO)
@@ -81,12 +82,20 @@ def xmllint_format(xml):
     output, _ = proc.communicate(xml)
     return output
 
-# TODO better name
-def mets_file_id(grp, n):
+def is_string(val):
+    return isinstance(val, (str, unicode)) if sys.version_info < (3, 0) else isinstance(val, str)
+
+def concat_padded(base, *args):
     """
     Concatenate string and zero-padded 4 digit number
     """
-    return "%s_%04i"  % (grp, n + 1)
+    ret = base
+    for n in args:
+        if is_string(n):
+            ret = "%s_%s" % (ret, n)
+        else:
+            ret = "%s_%04i"  % (ret, n + 1)
+    return ret
 
 def safe_filename(url):
     ret = re.sub('[^A-Za-z0-9]+', '.', url)

test/test_utils.py

@@ -1,6 +1,7 @@
 from test.base import TestCase, main
 from ocrd.utils import (
     points_from_xywh,
+    concat_padded,
     points_from_x0y0x1y1,
     xywh_from_points,
     polygon_from_points
@@ -33,5 +34,14 @@ def test_polygon_from_points(self):
             polygon_from_points('100,100 200,100 200,200 100,200'),
             [[100, 100], [200, 100], [200, 200], [100, 200]])
 
+    def test_concat_padded(self):
+        self.assertEqual(concat_padded('x', 0), 'x_0001')
+        self.assertEqual(concat_padded('x', 0, 1, 2), 'x_0001_0002_0003')
+        self.assertEqual(concat_padded('x', 0, '1', 2), 'x_0001_1_0003')
+
+    def test_is_string(self):
+        self.assertTrue('x')
+        self.assertTrue(u'x')
+
 if __name__ == '__main__':
     main()