diff options
| author | Alexandre Flament <alex@al-f.net> | 2022-04-16 11:54:49 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2022-04-16 11:54:49 +0200 |
| commit | c474616642890af97ce79700baff7dd47b2406b1 (patch) | |
| tree | 62885acb36461db216caa40560900fdaf0919f49 /searx/utils.py | |
| parent | ab9a4b732815cce32a442ca474d83e4f34f36a1e (diff) | |
| parent | 4224607c62d14a457ef3f46bf83dca6f271a592d (diff) | |
Merge pull request #1071 from return42/fix-lang-dailymotion
[fix] dailymotion engine: filter by language & country
Diffstat (limited to 'searx/utils.py')
| -rw-r--r-- | searx/utils.py | 4 |
1 files changed, 3 insertions, 1 deletions
diff --git a/searx/utils.py b/searx/utils.py index 43a7578d7..ffc9a39d6 100644 --- a/searx/utils.py +++ b/searx/utils.py @@ -88,6 +88,8 @@ class _HTMLTextExtractor(HTMLParser): # pylint: disable=W0223 # (see https://b def handle_starttag(self, tag, attrs): self.tags.append(tag) + if tag == 'br': + self.result.append(' ') def handle_endtag(self, tag): if not self.tags: @@ -142,7 +144,7 @@ def html_to_text(html_str: str) -> str: >>> html_to_text('<style>.span { color: red; }</style><span>Example</span>') 'Example' """ - html_str = html_str.replace('\n', ' ') + html_str = html_str.replace('\n', ' ').replace('\r', ' ') html_str = ' '.join(html_str.split()) s = _HTMLTextExtractor() try: |