]> Git — Sourcephile - gargantext.git/blob - src/Gargantext/Text/Parsers.hs
[API] New route for new corpus or new documents in existing corpus.
[gargantext.git] / src / Gargantext / Text / Parsers.hs
1 {-|
2 Module : Gargantext.Text.Parsers
3 Description : All parsers of Gargantext in one file.
4 Copyright : (c) CNRS, 2017
5 License : AGPL + CECILL v3
6 Maintainer : team@gargantext.org
7 Stability : experimental
8 Portability : POSIX
9
10 Gargantext enables analyzing semi-structured text that should be parsed
11 in order to be analyzed.
12
13 The parsers suppose we know the format of the Text (TextFormat data
14 type) according to which the right parser is chosen among the list of
15 available parsers.
16
17 This module mainly describe how to add a new parser to Gargantext,
18 please follow the types.
19 -}
20
21 {-# LANGUAGE NoImplicitPrelude #-}
22 {-# LANGUAGE PackageImports #-}
23 {-# LANGUAGE OverloadedStrings #-}
24
25 module Gargantext.Text.Parsers (FileFormat(..), clean, parseFile)
26 where
27
28 --import Data.ByteString (ByteString)
29 import "zip" Codec.Archive.Zip (withArchive, getEntry, getEntries)
30 import Control.Concurrent.Async as CCA (mapConcurrently)
31 import Control.Monad (join)
32 import qualified Data.ByteString.Char8 as DBC
33 import Data.Attoparsec.ByteString (parseOnly, Parser)
34 import Data.Either(Either(..))
35 import Data.Either.Extra (partitionEithers)
36 import Data.List (concat)
37 import Data.List (lookup)
38 import Data.Ord()
39 import Data.String (String())
40 import Data.String()
41 import Data.Text (Text)
42 import Data.Text.Encoding (decodeUtf8)
43 import Data.Tuple.Extra (both, first, second)
44 import System.FilePath (FilePath(), takeExtension)
45 import qualified Data.ByteString as DB
46 import qualified Data.Map as DM
47 import qualified Data.Text as DT
48 import Gargantext.Core (Lang(..))
49 import Gargantext.Prelude
50 import Gargantext.Database.Types.Node (HyperdataDocument(..))
51 import qualified Gargantext.Text.Parsers.WOS as WOS
52 import qualified Gargantext.Text.Parsers.RIS as RIS
53 import Gargantext.Text.Parsers.RIS.Presse (presseEnrich)
54 import qualified Gargantext.Text.Parsers.Date as Date
55 import Gargantext.Text.Parsers.CSV (parseHal)
56 import Gargantext.Text.Terms.Stop (detectLang)
57 ------------------------------------------------------------------------
58
59 type ParseError = String
60 --type Field = Text
61 --type Document = DM.Map Field Text
62 --type FilesParsed = DM.Map FilePath FileParsed
63 --data FileParsed = FileParsed { _fileParsed_errors :: Maybe ParseError
64 -- , _fileParsed_result :: [Document]
65 -- } deriving (Show)
66
67
68 -- | According to the format of Input file,
69 -- different parser are available.
70 data FileFormat = WOS | RIS | RisPresse
71 | CsvGargV3 | CsvHalFormat
72 deriving (Show)
73
74 -- Implemented (ISI Format)
75 -- | DOC -- Not Implemented / import Pandoc
76 -- | ODT -- Not Implemented / import Pandoc
77 -- | PDF -- Not Implemented / pdftotext and import Pandoc ?
78 -- | XML -- Not Implemented / see :
79
80
81 {-
82 parseFormat :: FileFormat -> ByteString -> [HyperdataDocument]
83 parseFormat = undefined
84 -}
85
86 -- | Parse file into documents
87 -- TODO manage errors here
88 -- TODO: to debug maybe add the filepath in error message
89 parseFile :: FileFormat -> FilePath -> IO [HyperdataDocument]
90 parseFile CsvHalFormat p = parseHal p
91 parseFile RisPresse p = join $ mapM (toDoc RIS) <$> snd <$> enrichWith RisPresse <$> readFileWith RIS p
92 parseFile WOS p = join $ mapM (toDoc WOS) <$> snd <$> enrichWith WOS <$> readFileWith WOS p
93 parseFile ff p = join $ mapM (toDoc ff) <$> snd <$> enrichWith ff <$> readFileWith ff p
94
95 toDoc :: FileFormat -> [(Text, Text)] -> IO HyperdataDocument
96 -- TODO use language for RIS
97 toDoc ff d = do
98 let abstract = lookup "abstract" d
99 let lang = maybe EN identity (join $ detectLang <$> (fmap (DT.take 50) abstract))
100
101 let dateToParse = DT.replace "-" " " <$> lookup "PY" d <> Just " " <> lookup "publication_date" d
102
103 (utcTime, (pub_year, pub_month, pub_day)) <- Date.split lang dateToParse
104
105 pure $ HyperdataDocument (Just $ DT.pack $ show ff)
106 (lookup "doi" d)
107 (lookup "URL" d)
108 Nothing
109 Nothing
110 Nothing
111 (lookup "title" d)
112 Nothing
113 (lookup "authors" d)
114 (lookup "source" d)
115 (lookup "abstract" d)
116 (fmap (DT.pack . show) utcTime)
117 (pub_year)
118 (pub_month)
119 (pub_day)
120 Nothing
121 Nothing
122 Nothing
123 (Just $ (DT.pack . show) lang)
124
125 enrichWith :: FileFormat
126 -> (a, [[[(DB.ByteString, DB.ByteString)]]]) -> (a, [[(Text, Text)]])
127 enrichWith RisPresse = enrichWith' presseEnrich
128 enrichWith WOS = enrichWith' (map (first WOS.keys))
129 enrichWith _ = enrichWith' identity
130
131
132 enrichWith' :: ([(DB.ByteString, DB.ByteString)] -> [(DB.ByteString, DB.ByteString)])
133 -> (a, [[[(DB.ByteString, DB.ByteString)]]]) -> (a, [[(Text, Text)]])
134 enrichWith' f = second (map both' . map f . concat)
135 where
136 both' = map (both decodeUtf8)
137
138 readFileWith :: FileFormat -> FilePath
139 -> IO ([ParseError], [[[(DB.ByteString, DB.ByteString)]]])
140 readFileWith format path = do
141 files <- case takeExtension path of
142 ".zip" -> openZip path
143 _ -> pure <$> clean <$> DB.readFile path
144 partitionEithers <$> mapConcurrently (runParser format) files
145
146
147 -- | withParser:
148 -- According to the format of the text, choose the right parser.
149 -- TODO withParser :: FileFormat -> Parser [Document]
150 withParser :: FileFormat -> Parser [[(DB.ByteString, DB.ByteString)]]
151 withParser WOS = WOS.parser
152 withParser RIS = RIS.parser
153 --withParser ODT = odtParser
154 --withParser XML = xmlParser
155 withParser _ = panic "[ERROR] Parser not implemented yet"
156
157 runParser :: FileFormat -> DB.ByteString
158 -> IO (Either String [[(DB.ByteString, DB.ByteString)]])
159 runParser format text = pure $ parseOnly (withParser format) text
160
161 openZip :: FilePath -> IO [DB.ByteString]
162 openZip fp = do
163 entries <- withArchive fp (DM.keys <$> getEntries)
164 bs <- mapConcurrently (\s -> withArchive fp (getEntry s)) entries
165 pure bs
166
167 clean :: DB.ByteString -> DB.ByteString
168 clean txt = DBC.map clean' txt
169 where
170 clean' '’' = '\''
171 clean' '\r' = ' '
172 clean' c = c