diff --git a/ShadowLibrary/Core.hs b/ShadowLibrary/Core.hs index 678df33..e0b36eb 100644 --- a/ShadowLibrary/Core.hs +++ b/ShadowLibrary/Core.hs @@ -33,7 +33,7 @@ import Data.Tree.NTree.TypeDefs import Data.Maybe import Control.Monad.Trans import Text.XML.HXT.XPath --- import Text.XML.HXT.Curl +import Text.XML.HXT.Curl import Text.XML.HXT.HTTP import Text.Regex.TDFA @@ -64,8 +64,8 @@ downloadDocument = readFromDocument [withParseHTML yes, withEncodingErrors no, withPreserveComment yes, withStrictInput yes, - withHTTP [] --- withCurl [("curl--user-agent","AMU Digital Libraries Indexing Agent")] +-- withHTTP [] + withCurl [("curl--user-agent","AMU Digital Libraries Indexing Agent")] ] downloadDocumentWithEncoding enc = readFromDocument [withParseHTML yes, @@ -73,13 +73,13 @@ downloadDocumentWithEncoding enc = readFromDocument [withParseHTML yes, withEncodingErrors no, withPreserveComment yes, withInputEncoding enc, - withHTTP []] --- withCurl []] +-- withHTTP []] + withCurl []] downloadXmlDocument = readFromDocument [withWarnings no, withEncodingErrors no, - withHTTP []] --- withCurl [] ] +-- withHTTP []] + withCurl [] ] data ShadowLibrary = ShadowLibrary { logoUrl :: Maybe String, diff --git a/app/instytutTechnikiGorniczej.hs b/app/instytutTechnikiGorniczej.hs new file mode 100644 index 0000000..58347b4 --- /dev/null +++ b/app/instytutTechnikiGorniczej.hs @@ -0,0 +1,48 @@ + +{-# LANGUAGE Arrows, NoMonomorphismRestriction #-} +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 + +extractRecords = extractLinksWithText "//li/a[contains(@href,'/lata')]" + >>> second (arr $ replace "\r\n " " ") + >>> first (extractLinksWithText "//p/a[contains(@href, '/archiwum/images/gazety')]") + +toShadowItem :: ((String, String), String) -> ShadowItem +toShadowItem ((url, articleTitle), yearlyTitle) = + (defaultShadowItem url title) { + originalDate = Just date, + itype = "periodical", + format = Just form, +-- format = Just "pdf", + finalUrl = url + } + where title = "Instytut Techniki Gorniczej " ++ yearlyTitle ++ " " ++ (replace "\r\n" "" (replace "\r\n " "" articleTitle)) + date = getDate $ replace "%20" "" url + form = unwords $ map stringToLower [getFormat url] + +getDate url = + case url =~~ "(19[0-9][0-9]|20[0-9][0-9])" :: Maybe [[String]] of + Just [[_, year]] -> year + otherwise -> error $ "unexpected url: " ++ url + +getFormat url = + case url =~~ "([a-zA-Z]+$)" :: Maybe [[String]] of + Just [[_, format]] -> format + otherwise -> error $ "unexpected url: " ++ url + +main = do + let start = "http://komag.gliwice.pl/archiwum/historia-komag/" + let shadowLibrary = ShadowLibrary {logoUrl=Nothing, + lname="Instytut Techniki Gorniczej", + abbrev="Komag", + lLevel=0, + webpage=start} + extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem) diff --git a/shadow-library.cabal b/shadow-library.cabal index cd77195..fb65eb6 100644 --- a/shadow-library.cabal +++ b/shadow-library.cabal @@ -20,6 +20,7 @@ library , HTTP , hxt , hxt-http + , hxt-curl , hxt-xpath , MissingH , monad-logger @@ -59,6 +60,17 @@ executable almanachmuszyny , shadow-library default-language: Haskell2010 +executable instytutTechnikiGorniczej + hs-source-dirs: app + main-is: instytutTechnikiGorniczej.hs + ghc-options: -threaded -rtsopts -with-rtsopts=-N + build-depends: base + , hxt + , hxt-xpath + , MissingH + , regex-posix + , shadow-library + default-language: Haskell2010 source-repository head type: git diff --git a/stack.yaml.lock b/stack.yaml.lock new file mode 100644 index 0000000..cc80ef3 --- /dev/null +++ b/stack.yaml.lock @@ -0,0 +1,19 @@ +# This file was autogenerated by Stack. +# You should not edit this file by hand. +# For more information, please see the documentation at: +# https://docs.haskellstack.org/en/stable/lock_files + +packages: +- completed: + hackage: hxt-xpath-9.1.2.2@sha256:9cd590ae93a04573db8f90fa4094625ebd97dded45da7667c577ce6b38a42900,1999 + pantry-tree: + size: 2225 + sha256: aee2f75974e868ff429b8ff349a29667536c60397098f5dfedc968d1951511bb + original: + hackage: hxt-xpath-9.1.2.2 +snapshots: +- completed: + size: 507596 + url: https://raw.githubusercontent.com/commercialhaskell/stackage-snapshots/master/lts/11/9.yaml + sha256: 42f472dbf06482da1b3319241f3e3b3593a45bd7d4f537d2789f21386b9b2ad3 + original: lts-11.9