From 0a4d3c8f12dc5797a919a00b6bcaf759947687cc Mon Sep 17 00:00:00 2001 From: Joris Date: Sun, 17 Jun 2018 23:24:47 +0200 Subject: Add ouest france parser --- src/parser/haskell/Parser/OuestFranceParser.hs | 27 ++++++++++++++++++++++++++ 1 file changed, 27 insertions(+) create mode 100644 src/parser/haskell/Parser/OuestFranceParser.hs (limited to 'src/parser/haskell/Parser/OuestFranceParser.hs') diff --git a/src/parser/haskell/Parser/OuestFranceParser.hs b/src/parser/haskell/Parser/OuestFranceParser.hs new file mode 100644 index 0000000..a7b6360 --- /dev/null +++ b/src/parser/haskell/Parser/OuestFranceParser.hs @@ -0,0 +1,27 @@ +module Parser.OuestFranceParser + ( parse + ) where + +import Data.Maybe (catMaybes) +import Data.Text (Text) +import qualified Data.Text as T + +import Text.HTML.TagSoup + +import Model.Ad (Ad (Ad)) + +import Parser.Utils + +parse :: Text -> [Ad] +parse page = + catMaybes . fmap parseAd $ partitions (~== "") tags + where tags = getTagsBetween "
" "
" (parseTags page) + +parseAd :: [Tag Text] -> Maybe Ad +parseAd tags = do + name <- getTagTextAfter "" tags + location <- getTagTextAfter "" tags + let price = getTagTextAfter "" tags + let startUrl = T.pack "https://www.ouestfrance-immo.com/" + url <- getTagAttribute "" (T.pack "href") tags + return (Ad name location price (T.concat [startUrl, url])) -- cgit v1.2.3