s434695/twilight-library
1
0
Fork 0
forked from filipg/twilight-library
Code Issues Pull Requests Releases Wiki Activity

nawet cos dziala

Browse Source
This commit is contained in:
s434695 2021-04-13 20:35:20 +02:00
parent f14d53337e
commit fbd7f7e448
1 changed files with 4 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
app/pbsociety.hs
View File

@ -13,8 +13,9 @@ import Text.Printf
extractRecords = extractLinksWithText "//a[@class='image-link']" -- pary adres-tytuł extractRecords = extractLinksWithText "//a[@class='image-link']" -- pary adres-tytuł
>>> second (arr $ replace "\n" "") >>> second (arr $ replace "\n" " ")
>>> first (extractLinksWithText "//a[@class='image-link']") >>> first (extractLinksWithText "//div/a[contains(@href,'.pdf') and not(@class)]") -- pobieramy stronę z adresu URL i wyciągamy linki z tej strony pasujące do wyrażenia XPathowego
-- ostatecznie wyjdą trójki ((adres URL, tytuł artykułu), tytuł rocznika)
-- ... a tutaj te trójki przerabiamy do docelowej struktury ShadowItem -- ... a tutaj te trójki przerabiamy do docelowej struktury ShadowItem
toShadowItem :: ((String, String), String) -> ShadowItem toShadowItem :: ((String, String), String) -> ShadowItem
@ -26,6 +27,7 @@ toShadowItem ((url, articleTitle), yearlyTitle) =
finalUrl = url finalUrl = url
} }
where title = "Pbsociety " ++ yearlyTitle where title = "Pbsociety " ++ yearlyTitle
date = getDate url
getDate url = getDate url =
case url =~~ "/(19[0-9][0-9]|20[0-9][0-9])/" :: Maybe [[String]] of case url =~~ "/(19[0-9][0-9]|20[0-9][0-9])/" :: Maybe [[String]] of