Rename shorten_selector_data to shorten and improve its implementation

parsel/selector.py

@@ -7,7 +7,7 @@
 import six
 from lxml import etree, html
 
-from .utils import flatten, iflatten, extract_regex, shorten_selector_data
+from .utils import flatten, iflatten, extract_regex, shorten
 from .csstranslator import HTMLTranslator, GenericTranslator
 
 
@@ -358,6 +358,6 @@ def __bool__(self):
     __nonzero__ = __bool__
 
     def __str__(self):
-        data = repr(shorten_selector_data(self.get(), width=40))
+        data = repr(shorten(self.get(), width=40))
         return "<%s xpath=%r data=%s>" % (type(self).__name__, self._expr, data)
     __repr__ = __str__

parsel/utils.py

@@ -82,7 +82,13 @@ def extract_regex(regex, text, replace_entities=True):
         return strings
     return [w3lib_replace_entities(s, keep=['lt', 'amp']) for s in strings]
 
-def shorten_selector_data(data, width):
-    """Shortens the preview of extracted data by adding placeholder '...'
-    """
-    return data[:width - 3] + "..." if (len(data) > width - 3) else data
+
+def shorten(text, width, suffix='...'):
+    """Truncate the given text to fit in the given width."""
+    if len(text) <= width:
+        return text
+    if width > len(suffix):
+        return text[:width-len(suffix)] + suffix
+    if width >= 0:
+        return suffix[len(suffix)-width:]
+    raise ValueError('width must be equal or greater than 0')

tests/test_utils.py

@@ -0,0 +1,26 @@
+from parsel.utils import shorten
+
+from pytest import mark, raises
+import six
+
+
+@mark.parametrize(
+    'width,expected',
+    (
+        (-1, ValueError),
+        (0, u''),
+        (1, u'.'),
+        (2, u'..'),
+        (3, u'...'),
+        (4, u'f...'),
+        (5, u'fo...'),
+        (6, u'foobar'),
+        (7, u'foobar'),
+    )
+)
+def test_shorten(width, expected):
+    if isinstance(expected, six.string_types):
+        assert shorten(u'foobar', width) == expected
+    else:
+        with raises(expected):
+            shorten(u'foobar', width)