diff options
| author | Nicolas Lœuillet <nicolas@loeuillet.org> | 2014-07-13 10:15:40 +0200 |
|---|---|---|
| committer | Nicolas Lœuillet <nicolas@loeuillet.org> | 2014-07-13 10:15:40 +0200 |
| commit | 4e067ceabd705201a16b4c92cf4b23f3b990326c (patch) | |
| tree | 939f3a8e5ff3ab9ee414a57a895d3e78e1d46ce3 /inc/3rdparty/site_config/standard/wyborcza.pl.txt | |
| parent | 58dbe103889148def78b0fc8744d3f94c56a1561 (diff) | |
| download | wallabag-4e067ceabd705201a16b4c92cf4b23f3b990326c.tar.gz wallabag-4e067ceabd705201a16b4c92cf4b23f3b990326c.tar.zst wallabag-4e067ceabd705201a16b4c92cf4b23f3b990326c.zip | |
updated specific configuration for parsing
Diffstat (limited to 'inc/3rdparty/site_config/standard/wyborcza.pl.txt')
| -rwxr-xr-x[-rw-r--r--] | inc/3rdparty/site_config/standard/wyborcza.pl.txt | 20 |
1 files changed, 9 insertions, 11 deletions
diff --git a/inc/3rdparty/site_config/standard/wyborcza.pl.txt b/inc/3rdparty/site_config/standard/wyborcza.pl.txt index f99467c2..638583dc 100644..100755 --- a/inc/3rdparty/site_config/standard/wyborcza.pl.txt +++ b/inc/3rdparty/site_config/standard/wyborcza.pl.txt | |||
| @@ -1,11 +1,9 @@ | |||
| 1 | title:h1 | 1 | body: //div[@id='article'] |
| 2 | author: //*[@class = 'author'] | 2 | strip: //div[@class='head'] |
| 3 | date: //*[@class = 'date'] | 3 | |
| 4 | body: //*[@id = 'art'] | 4 | strip_id_or_class: txt_upl |
| 5 | next_page_link: //*[@id='Str']/a[contains(text(), 'nastepne')] | 5 | |
| 6 | strip: //*[@class = 'rel_zdjTOP'] | 6 | single_page_link: //div[@id='gazeta_article_tools']//a[contains(@class, 'print')] |
| 7 | strip: //*[@id = 'rel'] | 7 | |
| 8 | strip: //*[@class = 'txt_upl'] | 8 | test_url: http://wyborcza.pl/1,123455,11536088,Gdy_peknie_fejs__obryzga_wszystko.html?as=1&startsz=x |
| 9 | strip: //*[@id='Str'] | 9 | test_url: http://wyborcza.pl/1,75478,14880255,Biskup_Dydycz_o_pedofilii_i_tajemnicy_spowiedzi__Zamiast.html \ No newline at end of file |
| 10 | strip: //*[@id='source'] | ||
| 11 | test_url: http://wyborcza.pl/1,123455,11536088,Gdy_peknie_fejs__obryzga_wszystko.html?as=1&startsz=x \ No newline at end of file | ||