Add 'gtf2bed/' from commit '1059f83a'

git-subtree-dir: gtf2bed
git-subtree-mainline: 0f82a98e
git-subtree-split: 1059f83a

gtf2bed/ensembl_gtf2bed.hs

+-- import System.Environment (getArgs)
+-- https://mail.haskell.org/pipermail/beginners/2011-October/008787.html
+-- import Data.List.Utils (join, split)
+-- import qualified Data.Text.Lazy as T
+-- faster than Lazy
+-- import qualified Data.Text as T
+import qualified Data.ByteString.Lazy.Char8 as C
+
+-- import Numeric (readDec)
+-- fastRead :: T.Text -> Int
+-- fastRead t = case readDec (T.unpack t) of
+--     [(n, "")] -> n
+
+fastRead :: C.ByteString -> Int
+fastRead t = let Just (n, _) = C.readInt t in n
+
+splitFields :: C.ByteString -> [C.ByteString]
+splitFields = C.split '\t'
+
+splitAnnots :: C.ByteString -> [C.ByteString]
+-- C.init is to avoid the last ";"
+-- C.dropWhile (== ' ') is to skip the spaces that followed the ';'
+splitAnnots = map (C.dropWhile (== ' ')) . C.split ';' . C.init
+
+splitAnnot :: C.ByteString -> (C.ByteString, C.ByteString)
+splitAnnot t = (k, v)
+    where [k, v] = C.split ' ' t
+
+notGeneId :: (C.ByteString, C.ByteString) -> Bool
+notGeneId p = fst p /= C.pack "gene_id"
+
+joinFields :: [C.ByteString] -> C.ByteString
+joinFields = C.intercalate (C.pack "\t")
+
+processLine :: C.ByteString -> C.ByteString
+processLine l = bedline
+    where bedline = joinFields [chrom, bedStart, end, geneId, score, strand]
+          bedStart = C.pack $ show ((fastRead start :: Int) - 1)
+          geneId = C.init . C.tail . snd . head . dropWhile notGeneId $ annots
+          annots = map splitAnnot (splitAnnots annotField)
+          (chrom : _ : _ : start : end : score : strand : _ : annotField : _) = splitFields l
+
+main :: IO ()
+-- Main processing done as String to limit memory usage by strict Text
+main = C.interact (C.unlines . map processLine . C.lines)