s426134/isi-2021-twilight-library
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

change file

Browse Source
This commit is contained in:
Łukasz Góreczny 2021-04-04 00:46:19 +02:00
parent 199c98fb27
commit 8eedac11d7
1 changed files with 47 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

55
app/maszynista.hs
View File

@ -6,7 +6,8 @@ import Text.XML.HXT.Core
import Text.XML.HXT.XPath import Text.XML.HXT.XPath
-- import Text.XML.HXT.Curl -- import Text.XML.HXT.Curl
import Data.List import Data.List
import Data.List.Utils (replace)
import Data.List.Utils (replace, split, join)
import Text.Regex.Posix import Text.Regex.Posix
import Text.Printf import Text.Printf
@ -25,7 +26,39 @@ extractRecords = extractLinksWithText "//section[@class='widget widget_zzm_gm_ar
>>> first (arr $ first (arr $ second(arr $ replace "ś" "s"))) >>> first (arr $ first (arr $ second(arr $ replace "ś" "s")))
>>> first (arr $ first (arr $ second(arr $ replace "ź" "z"))) >>> first (arr $ first (arr $ second(arr $ replace "ź" "z")))
>>> first (arr $ first (arr $ second(arr $ replace "ł" "l"))) >>> first (arr $ first (arr $ second(arr $ replace "ł" "l")))
>>> first (arr $ second (arr $ replace "Styczen" "01"))
>>> first (arr $ second (arr $ replace "styczen" "01"))
>>> first (arr $ second (arr $ replace "Luty" "02"))
>>> first (arr $ second (arr $ replace "luty" "02"))
>>> first (arr $ second (arr $ replace "Marzec" "03"))
>>> first (arr $ second (arr $ replace "marzec" "03"))
>>> first (arr $ second (arr $ replace "Kwiecien" "04"))
>>> first (arr $ second (arr $ replace "kwiecien" "04"))
>>> first (arr $ second (arr $ replace "Maj" "05"))
>>> first (arr $ second (arr $ replace "maj" "05"))
>>> first (arr $ second (arr $ replace "Czerwiec" "06"))
>>> first (arr $ second (arr $ replace "czerwiec" "06"))
>>> first (arr $ second (arr $ replace "Lipiec" "07"))
>>> first (arr $ second (arr $ replace "lipiec" "07"))
>>> first (arr $ second (arr $ replace "Sierpien" "08"))
>>> first (arr $ second (arr $ replace "sierpien" "08"))
>>> first (arr $ second (arr $ replace "sierpnia" "08"))
>>> first (arr $ second (arr $ replace "Wrzesien" "09"))
>>> first (arr $ second (arr $ replace "wrzesien" "09" ))
>>> first (arr $ second (arr $ replace "wrzesnia" "09" ))
>>> first (arr $ second (arr $ replace "pazdziernika" "10" ))
>>> first (arr $ second (arr $ replace "Pazdziernik" "10" ))
>>> first (arr $ second (arr $ replace "pazdziernik" "10" ))
>>> first (arr $ second (arr $ replace "listopada" "11" ))
>>> first (arr $ second (arr $ replace "Listopad" "11" ))
>>> first (arr $ second (arr $ replace "listopad" "11" ))
>>> first (arr $ second (arr $ replace "Grudzien" "12"))
>>> first (arr $ second (arr $ replace "grudzien" "12"))
>>> first (arr $ second (arr $ replace "grudnia" "12"))
>>> first (arr $ second (arr $ replace " " "-"))
-- ... a tutaj te trójki przerabiamy do docelowej struktury ShadowItem -- ... a tutaj te trójki przerabiamy do docelowej struktury ShadowItem
@ -37,13 +70,19 @@ toShadowItem (((url, articleTitle), articleDatePart), year) =
format = Just "pdf", format = Just "pdf",
finalUrl = url finalUrl = url
} }
where title = articleTitle where title = fixTitle articleTitle
fixTitle :: String -> String
fixTitle title
| title =~ "Glos Maszynisty .*" = title
| otherwise = "Glos Maszynisty " ++ title
getDate :: String -> String
getDate url =
case url =~~ ("[0-9]*-?[0-9]{2}-[0-9]{4}" :: String) of
Just date -> concat([join "-" (reverse (split "-" date))]) :: String
otherwise -> ""
getDate :: String -> Maybe String
getDate url =
case url =~~ ("[0-9]* ?[a-zA-Z]+ [0-9]{4}" :: String) of
Just year -> Just year
otherwise -> Nothing
main = do main = do