diff options
| author | Noémi Ványi <kvch@users.noreply.github.com> | 2019-01-04 22:27:05 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2019-01-04 22:27:05 +0100 |
| commit | 1ea56576dc4f2fb65b9028c900c2e4fb063a8b9a (patch) | |
| tree | e4d0e50c0494ca49a41351edea532c00289386a3 /searx/engines/startpage.py | |
| parent | 0e493db2fb06fd78ac05f82abd8b1cc86089684f (diff) | |
| parent | 899ba5d6dee82faacb572b4d9bc4c58570628531 (diff) | |
Merge branch 'master' into devel_google_videos
Diffstat (limited to 'searx/engines/startpage.py')
| -rw-r--r-- | searx/engines/startpage.py | 11 |
1 files changed, 4 insertions, 7 deletions
diff --git a/searx/engines/startpage.py b/searx/engines/startpage.py index 3e067597e..55efdc884 100644 --- a/searx/engines/startpage.py +++ b/searx/engines/startpage.py @@ -32,8 +32,9 @@ search_url = base_url + 'do/search' # specific xpath variables # ads xpath //div[@id="results"]/div[@id="sponsored"]//div[@class="result"] # not ads: div[@class="result"] are the direct childs of div[@id="results"] -results_xpath = '//div[@class="result"]' +results_xpath = '//li[contains(@class, "search-result") and contains(@class, "search-item")]' link_xpath = './/h3/a' +content_xpath = './p[@class="search-item__body"]' # do search-request @@ -73,14 +74,10 @@ def response(resp): if re.match(r"^http(s|)://(www\.)?startpage\.com/do/search\?.*$", url): continue - # block ixquick search url's - if re.match(r"^http(s|)://(www\.)?ixquick\.com/do/search\?.*$", url): - continue - title = extract_text(link) - if result.xpath('./p[@class="desc clk"]'): - content = extract_text(result.xpath('./p[@class="desc clk"]')) + if result.xpath(content_xpath): + content = extract_text(result.xpath(content_xpath)) else: content = '' |