MCPcopy
hub / github.com/rspeer/python-ftfy / test_entities

Function test_entities

tests/test_entities.py:7–37  ·  view source on GitHub ↗
()

Source from the content-addressed store, hash-verified

5
6
7def test_entities():
8 example = "&amp;\n<html>\n&amp;"
9 assert fix_text(example) == "&\n<html>\n&amp;"
10 assert fix_text_segment(example) == "&amp;\n<html>\n&amp;"
11
12 assert fix_text(example, unescape_html=True) == "&\n<html>\n&"
13 assert fix_text_segment(example, unescape_html=True) == "&\n<html>\n&"
14
15 assert fix_text(example, unescape_html=False) == "&amp;\n<html>\n&amp;"
16 assert fix_text_segment(example, unescape_html=False) == "&amp;\n<html>\n&amp;"
17
18 assert fix_text_segment("&lt;&gt;", unescape_html=False) == "&lt;&gt;"
19 assert fix_text_segment("&lt;&gt;", unescape_html=True) == "<>"
20 assert fix_text_segment("&lt;&gt;") == "<>"
21 assert fix_text_segment("jednocze&sacute;nie") == "jednocześnie"
22 assert fix_text_segment("JEDNOCZE&Sacute;NIE") == "JEDNOCZEŚNIE"
23 assert fix_text_segment("ellipsis&#133;", normalization="NFKC") == "ellipsis..."
24 assert fix_text_segment("ellipsis&#x85;", normalization="NFKC") == "ellipsis..."
25 assert fix_text_segment("broken&#x81;") == "broken\x81"
26 assert fix_text_segment("&amp;amp;amp;") == "&"
27 assert unescape_html("euro &#x80;") == "euro €"
28 assert unescape_html("EURO &EURO;") == "EURO €"
29 assert unescape_html("not an entity &#20x6;") == "not an entity &#20x6;"
30 assert unescape_html("JEDNOCZE&SACUTE;NIE") == "JEDNOCZEŚNIE"
31 assert unescape_html("V&SCARON;ICHNI") == "VŠICHNI"
32 assert unescape_html("&#xffff;") == ""
33 assert unescape_html("&#xffffffff;") == "\ufffd"
34 assert (
35 fix_text_segment("this is just informal english &not html")
36 == "this is just informal english &not html"
37 )
38
39
40def test_old_parameter_name():

Callers

nothing calls this directly

Calls 3

fix_textFunction · 0.90
fix_text_segmentFunction · 0.90
unescape_htmlFunction · 0.90

Tested by

no test coverage detected