Attempt different character encodings when previewing a URL. (#11077)

This follows similar logic to BeautifulSoup where we attempt different
character encodings until we find one which works.
This commit is contained in:
Patrick Cloke 2021-10-14 10:17:20 -04:00 committed by GitHub
parent 1609ccf8fe
commit e2f0b49b3f
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 80 additions and 67 deletions

View file

@ -15,7 +15,7 @@
from synapse.rest.media.v1.preview_url_resource import (
_calc_og,
decode_body,
get_html_media_encoding,
get_html_media_encodings,
summarize_paragraphs,
)
@ -159,7 +159,7 @@ class CalcOgTestCase(unittest.TestCase):
</html>
"""
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(og, {"og:title": "Foo", "og:description": "Some text."})
@ -175,7 +175,7 @@ class CalcOgTestCase(unittest.TestCase):
</html>
"""
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(og, {"og:title": "Foo", "og:description": "Some text."})
@ -194,7 +194,7 @@ class CalcOgTestCase(unittest.TestCase):
</html>
"""
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(
@ -216,7 +216,7 @@ class CalcOgTestCase(unittest.TestCase):
</html>
"""
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(og, {"og:title": "Foo", "og:description": "Some text."})
@ -230,7 +230,7 @@ class CalcOgTestCase(unittest.TestCase):
</html>
"""
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(og, {"og:title": None, "og:description": "Some text."})
@ -245,7 +245,7 @@ class CalcOgTestCase(unittest.TestCase):
</html>
"""
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(og, {"og:title": "Title", "og:description": "Some text."})
@ -260,7 +260,7 @@ class CalcOgTestCase(unittest.TestCase):
</html>
"""
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(og, {"og:title": None, "og:description": "Some text."})
@ -268,13 +268,13 @@ class CalcOgTestCase(unittest.TestCase):
def test_empty(self):
"""Test a body with no data in it."""
html = b""
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
self.assertIsNone(tree)
def test_no_tree(self):
"""A valid body with no tree in it."""
html = b"\x00"
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
self.assertIsNone(tree)
def test_invalid_encoding(self):
@ -287,7 +287,7 @@ class CalcOgTestCase(unittest.TestCase):
</body>
</html>
"""
tree = decode_body(html, "invalid-encoding")
tree = decode_body(html, "http://example.com/test.html", "invalid-encoding")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(og, {"og:title": "Foo", "og:description": "Some text."})
@ -302,15 +302,29 @@ class CalcOgTestCase(unittest.TestCase):
</body>
</html>
"""
tree = decode_body(html)
tree = decode_body(html, "http://example.com/test.html")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(og, {"og:title": "ÿÿ Foo", "og:description": "Some text."})
def test_windows_1252(self):
"""A body which uses windows-1252, but doesn't declare that."""
html = b"""
<html>
<head><title>\xf3</title></head>
<body>
Some text.
</body>
</html>
"""
tree = decode_body(html, "http://example.com/test.html")
og = _calc_og(tree, "http://example.com/test.html")
self.assertEqual(og, {"og:title": "ó", "og:description": "Some text."})
class MediaEncodingTestCase(unittest.TestCase):
def test_meta_charset(self):
"""A character encoding is found via the meta tag."""
encoding = get_html_media_encoding(
encodings = get_html_media_encodings(
b"""
<html>
<head><meta charset="ascii">
@ -319,10 +333,10 @@ class MediaEncodingTestCase(unittest.TestCase):
""",
"text/html",
)
self.assertEqual(encoding, "ascii")
self.assertEqual(list(encodings), ["ascii", "utf-8", "windows-1252"])
# A less well-formed version.
encoding = get_html_media_encoding(
encodings = get_html_media_encodings(
b"""
<html>
<head>< meta charset = ascii>
@ -331,11 +345,11 @@ class MediaEncodingTestCase(unittest.TestCase):
""",
"text/html",
)
self.assertEqual(encoding, "ascii")
self.assertEqual(list(encodings), ["ascii", "utf-8", "windows-1252"])
def test_meta_charset_underscores(self):
"""A character encoding contains underscore."""
encoding = get_html_media_encoding(
encodings = get_html_media_encodings(
b"""
<html>
<head><meta charset="Shift_JIS">
@ -344,11 +358,11 @@ class MediaEncodingTestCase(unittest.TestCase):
""",
"text/html",
)
self.assertEqual(encoding, "Shift_JIS")
self.assertEqual(list(encodings), ["Shift_JIS", "utf-8", "windows-1252"])
def test_xml_encoding(self):
"""A character encoding is found via the meta tag."""
encoding = get_html_media_encoding(
encodings = get_html_media_encodings(
b"""
<?xml version="1.0" encoding="ascii"?>
<html>
@ -356,11 +370,11 @@ class MediaEncodingTestCase(unittest.TestCase):
""",
"text/html",
)
self.assertEqual(encoding, "ascii")
self.assertEqual(list(encodings), ["ascii", "utf-8", "windows-1252"])
def test_meta_xml_encoding(self):
"""Meta tags take precedence over XML encoding."""
encoding = get_html_media_encoding(
encodings = get_html_media_encodings(
b"""
<?xml version="1.0" encoding="ascii"?>
<html>
@ -370,7 +384,7 @@ class MediaEncodingTestCase(unittest.TestCase):
""",
"text/html",
)
self.assertEqual(encoding, "UTF-16")
self.assertEqual(list(encodings), ["UTF-16", "ascii", "utf-8", "windows-1252"])
def test_content_type(self):
"""A character encoding is found via the Content-Type header."""
@ -384,10 +398,10 @@ class MediaEncodingTestCase(unittest.TestCase):
'text/html; charset=ascii";',
)
for header in headers:
encoding = get_html_media_encoding(b"", header)
self.assertEqual(encoding, "ascii")
encodings = get_html_media_encodings(b"", header)
self.assertEqual(list(encodings), ["ascii", "utf-8", "windows-1252"])
def test_fallback(self):
"""A character encoding cannot be found in the body or header."""
encoding = get_html_media_encoding(b"", "text/html")
self.assertEqual(encoding, "utf-8")
encodings = get_html_media_encodings(b"", "text/html")
self.assertEqual(list(encodings), ["utf-8", "windows-1252"])