s440054/twilight-library
1
0
Fork 0
forked from filipg/twilight-library
Code Issues Pull Requests Releases Wiki Activity

Fetching text

Browse Source
This commit is contained in:
Norbert Litkowski 2021-04-07 03:54:56 +02:00
parent 10c06c503c
commit cb5948925d
1 changed files with 4 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
app/ZborBielawa.hs
View File

@ -4,21 +4,20 @@ import ShadowLibrary.Core
import Text.XML.HXT.Core
import Text.XML.HXT.XPath
-- import Text.XML.HXT.Curl
import Data.List
import Data.List.Utils (replace)
import Text.Regex.Posix
import Text.Printf
import Debug.Trace
getLinkAndText xpathCondition = proc doc -> do
xpathTrees <- getXPathTrees xpathCondition -< doc
name <- getElemName -< xpathTrees
txt <- (getXPathTrees "../text()" >>> getText) -< xpathTrees
txt <- (listA (deep isText >>> getText) >>> arr (intercalate " ")) -< xpathTrees
href <- (getXPathTrees "//a" >>> getAttrValue "href") -< xpathTrees
returnA -< trace ("AAAAAAAAAAAAAAAA " ++ show name) $ traceShowId (href, txt)
returnA -< href, txt
extractNestedLinksWithText xpathCondition = proc url -> do
@ -31,7 +30,7 @@ extractNestedLinksWithText xpathCondition = proc url -> do
extractRecords = proc x -> do
(a, b) <- extractLinksWithText "(//aside[@class='widget widget_maxmegamenu']//a[@class='mega-menu-link'])[1]" -< x -- pary adres-tytuł podstrony
(a', b') <- (extractLinksWithText "(//aside[@class='widget widget_maxmegamenu']//a[@class='mega-menu-link'])[1]") -< a -- pobieramy podstronę i kolejne podstrony z menu
a'' <- (extractNestedLinksWithText "//big/a[contains(@href,'.pdf')][img]") -< a' -- pobieramy stronę z adresu URL i wyciągamy linki z tej strony pasujące do wyrażenia XPathowego
a'' <- (extractNestedLinksWithText "//big[a[contains(@href,'.pdf')][img]]") -< a' -- pobieramy stronę z adresu URL i wyciągamy linki z tej strony pasujące do wyrażenia XPathowego
returnA -< ((a'', b'), b)
-- ostatecznie wyjdą krotki (((adres URL, tytuł nr-u), tytuł podstrony 2), tytuł podstrony 1)