🚧 dinglehopper: WIP data structure for extracted text

extracted_text.py

@@ -1,4 +1,5 @@
 import attr
+import unicodedata
 
 
 @attr.s(frozen=True)
@@ -23,10 +24,25 @@ class ExtractedText:
             i += len(self.joiner)
 
 
+NORM_NFC = 0
+
+
+def normalize(text, normalization):
+    if normalization == NORM_NFC:
+        return unicodedata.normalize('NFC', text)
+    else:
+        raise ValueError()
+
+
 @attr.s(frozen=True)
 class ExtractedTextSegment:
     id = attr.ib(type=str)
     text = attr.ib(type=str)
+    @text.validator
+    def check(self, attribute, value):
+        if normalize(value, self.normalization) != value:
+            raise ValueError('String "{}" is not normalized.'.format(value))
+    normalization = attr.ib(default=NORM_NFC)
 
 
 test1 = ExtractedText([
@@ -40,3 +56,6 @@ assert test1.text == 'foo bar bazinga'
 assert test1.segment_id_for_pos(0) == 's0'
 assert test1.segment_id_for_pos(3) == None
 assert test1.segment_id_for_pos(10) == 's2'
+
+# ExtractedTextSegment('foo', unicodedata.normalize('NFD', 'Schlyñ'))
+ExtractedTextSegment('foo', unicodedata.normalize('NFC', 'Schlyñ'))