Add basic document snippets.

sourcing/model.py

@@ -230,6 +230,9 @@ class XanaDoc(Item):
 
     id = Column(Integer, ForeignKey(Item.id), primary_key=True)
 
+    def snippet(self):
+        return self.text
+
 class XanaLink(Item):
     __tablename__ = 'xanalink'
     __mapper_args__ = {'polymorphic_identity': 'xanalink'}
@@ -274,6 +277,9 @@ class XanaLink(Item):
                 titles[item] = source_of_title.text[start:length + start]
         return titles
 
+    def snippet(self):
+        return self.text
+
 class SourceDoc(Item):
     __tablename__ = 'sourcedoc'
     __mapper_args__ = {'polymorphic_identity': 'sourcedoc'}
@@ -290,6 +296,16 @@ class SourceDoc(Item):
     def price_per_character(self):
         return self.db_price_per_character or self.db_document_price / len(self.text)
 
+    def snippet(self, length=255, killwords=False, end='...', leeway=5):
+        s = self.text
+        assert length >= len(end), 'expected length >= %s, got %s' % (len(end), length)
+        assert leeway >= 0, 'expected leeway >= 0, got %s' % leeway
+        if len(s) <= length + leeway:
+            return s
+        if killwords:
+            return s[:length - len(end)] + end
+        result = s[:length - len(end)].rsplit(' ', 1)[0]
+        return result + end
 
 
 configure_mappers()