Utils.hs revision 90e8679b4bc38316fce1c6f438bf3e9184a65437
{- |
Module :$Header$
Description : utilitary functions used throughout the CMDL interface
Copyright : uni-bremen and DFKI
License : similar to LGPL, see HetCATS/LICENSE.txt or LIZENZ.txt
Maintainer : r.pascanu@jacobs-university.de
Stability : provisional
Portability : portable
PGIP.Utils contains different basic functions that are
used throughout the CMDL interface and could not be found in
Prelude
-}
module PGIP.Utils
( isWhiteSpace
, trim
, trimLeft
, trimRight
, decomposeIntoGoals
, obtainNodeList
, createEdgeNames
, obtainEdgeList
, obtainGoalEdgeList
, unfinishedEdgeName
, stripComments
, lastChar
, lastString
, getInt
, safeTail
, fileFilter
, fileExtend
, prettyPrintList
, prettyPrintErrList
, nodeContainsGoals
, edgeContainsGoals
, checkIntString
, cleanPrompter
)where
import Data.List
import Data.Char
import Data.Graph.Inductive.Graph
import Static.GTheory
import Static.DevGraph
import System.Directory
import Common.AS_Annotation
import qualified Common.OrderedMap as OMap
cleanPrompter :: String -> String
cleanPrompter str = case find (=='.') str of
Nothing -> reverse $ safeTail $ safeTail $
reverse str
Just _ -> reverse $ safeTail $
dropWhile (/= '.') $ reverse str
-- | Checks if a string represents a int or not
checkIntString :: String -> Bool
checkIntString = not . any (not . isDigit)
-- | List of all characters considered white spaces
whiteSpaces ::String
whiteSpaces = " \t\n\r\v"
-- | Predicate that tells if a character is a white space
-- or not
isWhiteSpace ::Char -> Bool
isWhiteSpace x = any (x==) whiteSpaces
-- | trims a string both on left and right hand side
trim :: String -> String
trim = reverse . dropWhile isWhiteSpace . reverse
. dropWhile isWhiteSpace
-- | trims a string only on the left side
trimLeft :: String -> String
trimLeft = dropWhile isWhiteSpace
-- | trims a string only on the right side
trimRight :: String -> String
trimRight = reverse . dropWhile isWhiteSpace . reverse
-- | Given a string inserts spaces before and after an
-- arrow
spacesAroundArrows :: String -> String -> String
spacesAroundArrows s output
= let
--function to tell if in the string follows a arrow
isArrow text = case take 2 $ trimLeft text of
"->" ->True
_ ->False
in case s of
[] -> output
_ ->
case isArrow s of
True -> spacesAroundArrows (drop 2 $ trimLeft s)
(output ++" -> ")
False -> spacesAroundArrows (safeTail s)
(output ++ [head s])
-- | Given a string the function decomposes it into 4 lists,
-- one for node goals, the other for edges, the third for
-- numbered edges and the last for names that could not be
-- processed due to errors
decomposeIntoGoals :: String -> ([String],[String],[String],[String])
decomposeIntoGoals input
= let
-- the new input where words and arrows are separated
-- by exactly one space
nwInput = words $ spacesAroundArrows input []
-- funtion to parse the input and decompose it into
-- the three goal list
parse info nbOfArrows word sw listNode listEdge
listNbEdge listError
= case info of
[] -> case nbOfArrows :: Integer of
0 -> ((word:listNode), listEdge, listNbEdge,listError)
1 -> (listNode, (word:listEdge), listNbEdge,listError)
2 -> (listNode, listEdge, (word:listNbEdge),listError)
_ -> (listNode, listEdge, listNbEdge, (word:listError))
"->":l -> case word of
[] -> (listNode,listEdge,listNbEdge,
(word:listError))
_ -> parse l (nbOfArrows+1) (word++" -> ")
True listNode listEdge listNbEdge listError
x:l -> case sw of
True -> parse l nbOfArrows (word++x) False
listNode listEdge listNbEdge listError
False ->
case nbOfArrows of
0 -> parse l 0 x False
(word:listNode) listEdge listNbEdge listError
1 -> parse l 0 x False
listNode (word:listEdge) listNbEdge listError
2 -> parse l 0 x False
listNode listEdge (word:listNbEdge) listError
_ -> parse l 0 x False
listNode listEdge listNbEdge (word:listError)
in parse nwInput 0 [] True [] [] [] []
-- | mapAndSplit maps a function to a list. If the function can not
-- be applied to an element it is stored in a different list for
-- producing error message later on
mapAndSplit :: (a -> Maybe b) -> [a] -> ([a],[b])
mapAndSplit fn ls
= let mapAndSplit' fn' ls' errs mapped =
case ls' of
[] -> (errs,mapped)
x:l -> case fn' x of
Just y -> mapAndSplit' fn' l errs (y:mapped)
Nothing-> mapAndSplit' fn' l (x:errs) mapped
in mapAndSplit' fn ls [] []
-- | concatMapAndSplit is similar to mapAndSplit, just that it behaves
-- in a similar manner to concatMap (i.e. sums up lists produced by
-- the function
concatMapAndSplit :: (a -> [b]) -> [a] -> ([a],[b])
concatMapAndSplit fn ls
= let concatMapAndSplit' fn' ls' errs mapped =
case ls' of
[] -> (errs, mapped)
x:l -> case fn' x of
[] -> concatMapAndSplit' fn' l (x:errs) mapped
l' -> concatMapAndSplit' fn' l errs (mapped++l')
in concatMapAndSplit' fn ls [] []
-- | Given a list of node names and the list of all nodes
-- the function returns all the nodes that have their name
-- in the name list
obtainNodeList :: [String] ->[LNode DGNodeLab]
->([String],[LNode DGNodeLab])
obtainNodeList lN allNodes
= mapAndSplit
(\x -> find (\(_,label) -> getDGNodeName label == x) allNodes) lN
-- | Given a node decides if it contains goals or not
nodeContainsGoals:: LNode DGNodeLab -> G_theory -> Bool
nodeContainsGoals (_,l) th
= (not (isDGRef l)) &&
((case th of
G_theory _ _ _ sens _ ->
not $ OMap.null $ OMap.filter
(\s-> (not (isAxiom s)) && (not (isProvenSenStatus s))) sens
) || hasOpenConsStatus False l)
-- | Given an edge decides if it contains goals or not
edgeContainsGoals:: LEdge DGLinkLab -> Bool
edgeContainsGoals (_,_,l)
= case thmLinkStatus $ dgl_type l of
Just LeftOpen -> True
_ -> False
getInt :: EdgeId -> Int
getInt val
= case val of
EdgeId v -> v
-- | Given a list of edges and the complete list of all
-- edges computes not only the names of edges but also the
-- numbered name of edges
createEdgeNames:: [LNode DGNodeLab] -> [LEdge DGLinkLab] -> [String]
createEdgeNames lsN lsE
=
let
-- function that returns the name of a node given its number
nameOf x ls = case find(\(nb,_) -> nb == x) ls of
Nothing -> "Unknown node"
Just (_, nlab) -> showName $ dgn_name nlab
ordFn x y = let (x1,x2,_) = x
(y1,y2,_) = y
in if (x1,x2) > (y1,y2) then GT
else if (x1,x2) < (y1,y2) then LT
else EQ
-- sorted and grouped list of edges
edgs = groupBy ( \(x1,x2,_) (y1,y2,_)-> (x1,x2)==(y1,y2)) $
sortBy ordFn lsE
allEds= concatMap (\l -> case l of
[(x,y,_)]->[((nameOf x lsN) ++ " -> " ++
(nameOf y lsN))]
_ -> map (\(x,y,edgLab) ->
(nameOf x lsN) ++ " -> " ++
(show $ getInt $ dgl_id edgLab)
++ " -> " ++ (nameOf y lsN)) l) edgs
in allEds
-- | Given a list of edge names and numbered edge names
-- and the list of all nodes and edges the function
-- identifies the edges that appear in the name lists
obtainEdgeList :: [String] ->[String] ->[LNode DGNodeLab]
-> [LEdge DGLinkLab]-> ([String],[LEdge DGLinkLab])
obtainEdgeList lsEdge lsNbEdge allNodes allEdges
=
-- function that searches through a list of nodes to
-- find the node number for a given name.
let getNodeNb s ls =
case find ( \(_,label) ->
getDGNodeName label == s) ls of
Nothing -> Nothing
Just (nb,_) -> Just nb
-- converts a string to a number (for some reason I
-- could not find such a function already implemented
-- in the Prelude )
l1=concatMapAndSplit
(\nme ->
let allx = words nme
node1 = getNodeNb (allx!!0) allNodes
node2 = getNodeNb (allx!!2) allNodes
in
case node1 of
Nothing -> []
Just x ->
case node2 of
Nothing -> []
Just y ->
filter(\(x1,y1,_)->(x==x1)&&(y==y1)) allEdges
) lsEdge
-- compute the list of all numbered edges
l2=mapAndSplit
(\nme ->
let allx = words nme
node1= getNodeNb (allx!!0) allNodes
node2= getNodeNb (allx!!5) allNodes
nb = read (allx!!3)
in
case node1 of
Nothing -> Nothing
Just x ->
case node2 of
Nothing -> Nothing
Just y ->
let ls = filter(\(x1,y1,elab)->(x==x1)&&(y==y1)&&
(dgl_id elab==EdgeId nb)) allEdges
in case ls of
[] -> Nothing
els:_ -> Just els ) lsNbEdge
in ((fst l1)++(fst l2),(snd l1)++(snd l2))
-- | Giben a listof edgenamesand numbered edge names and
-- the list of all nodes and edges the function identifies
-- the edges that appearin the name list and are also goals
obtainGoalEdgeList :: [String] -> [String] ->
[LNode DGNodeLab] -> [LEdge DGLinkLab]
-> ([String],[LEdge DGLinkLab])
obtainGoalEdgeList ls1 ls2 ls3 ls4
= let (l1,l2) = obtainEdgeList ls1 ls2 ls3 ls4
l2' = filter edgeContainsGoals l2
in (l1,l2')
-- | Function that given a string removes comments contained
-- in the string
stripComments::String -> String
stripComments input
= let fn ls = case ls of
'#':_ -> []
'%':ll->
case ll of
'%':_ ->[]
'{':_ ->[]
_ -> '%':(fn ll)
[] -> []
l:ll -> l:(fn ll)
in trim $ fn input
-- | The function obtain the unfinished edge name from the
-- last position of the input or list of possible unfinished
-- edge names
unfinishedEdgeName :: String -> String
unfinishedEdgeName input
=
-- we need a penultimum (the one before the last) function
let prevLast s = lastString $ reverse $ safeTail
$ reverse s
-- and the one before the penultimum
prevPrevLast s = lastString $ reverse $ safeTail $
safeTail $ reverse s
prev2PrevLast s = lastString $ reverse $ safeTail $
safeTail $ safeTail $ reverse s
prev3PrevLast s = lastString $ reverse $ safeTail $
safeTail$ safeTail$ safeTail $
reverse s
in
-- is the last character an empty space?
case isWhiteSpace $ lastChar input of
True ->
-- if so, then either the last word is an arrow, and
-- then we have the consider last two words, or it
-- is not an arrow and then we need to consider just
-- the last word
case lastString $ words input of
"->" -> case prevPrevLast $ words input of
"->"->(prev2PrevLast $ words input) ++
" -> " ++ (prevLast $ words input)
++ " -> "
_ ->(prevLast $ words input) ++ " -> "
--anyhting else
_ -> case prevLast $ words input of
-- an entire edge name was just inserted
-- before and now we need a fresh new start
"->" -> []
-- if just the first word of an edge was
-- inserted then return that
_ -> case lastString $ words input of
[]-> []
_ -> (lastString $ words input)++" "
False ->
-- then we could be in the middle of the first node
-- name, arrow or the second node name
case prevLast $ words input of
-- in the middle of the last word
"->" -> case prev2PrevLast $ words input of
"->"->(prev3PrevLast $ words input) ++
" -> "++(prevPrevLast $ words
input)++" -> "++(lastString $
words input)
_->(prevPrevLast $ words input) ++
" -> " ++ (lastString $ words input)
_ -> case prevPrevLast $ words input of
-- in the middle of the first word
"->" -> lastString $ words input
-- in the middle of the arrow
_ -> case prevLast $ words input of
[] -> lastString $ words input
_ ->( (prevLast $ words input) ++
" " ++ (lastString $
words input) )
-- | Given a list of files and folders the function filters
-- only directory names and files ending in extenstion
-- .casl
fileFilter::String ->[String]->[String] -> IO [String]
fileFilter lPath ls cons
= case ls of
[] -> return cons
x:l ->
do
-- check if current element is a directory
b <- doesDirectoryExist (lPath++x)
case b of
-- if it is,then add "/" to indicate is a folder
True -> fileFilter lPath l ((x++"/"):cons)
-- if it is not a folder then it must be a file
-- so check the extension
False-> case isSuffixOf ".casl" x of
True -> fileFilter lPath l (x:cons)
False-> fileFilter lPath l cons
-- | Given a list of files and folders the function expands
-- the list adding the content of all folders in the list
fileExtend::String->[String]->[String]-> IO [String]
fileExtend lPath ls cons
= case ls of
[] -> return cons
x:l ->
do
-- check if current element is a directory
b<- doesDirectoryExist (lPath++x)
case b of
-- if it is not then leave the file alone
False -> fileExtend lPath l (x:cons)
-- if it is a folder add its content
True ->
do
ll <- getDirectoryContents (lPath++x)
nll<- fileFilter (lPath++x) ll []
let nll'=map (\y -> x++y) nll
fileExtend lPath l (nll' ++ cons)
-- | The function behaves exactly as tail just that
-- in the case of empty list returns an empty list
-- instead of an error
safeTail::[a]->[a]
safeTail ls
= case ls of
[] -> []
_:l -> l
-- | The function behaves exactly like last just that
-- in case of an empty list returns the space
-- character (it works only for lists of chars)
lastChar::[Char]->Char
lastChar ls
= case ls of
[] -> ' '
_ -> last ls
-- | The function behaves exactly like last just that
-- in case of an empty list returns the empty string
-- (it is meant only for list of strings)
lastString::[String]->String
lastString ls
= case ls of
[] -> ""
_ -> last ls
-- | The function nicely outputs a list of errors
prettyPrintErrList:: [String]->String
prettyPrintErrList list
= let
tmpPrint ls acc =
case ls of
[] -> []
x:ll -> tmpPrint ll $ ("Input "++x++
" could not be processed\n")++acc
in tmpPrint list []
-- | The function nicely ouputs a list of strings
prettyPrintList ::[String]->String
prettyPrintList ls
= unlines ls