src/Twitter/Xml.hs

   1 -- |Application-specific XML functions.
   2 module Twitter.Xml
   3 where
   4
   5 import Data.Char (chr)
   6 import Test.HUnit
   7 import Text.Regex (matchRegex, mkRegex, subRegex)
   8 import Text.XML.HaXml
   9
  10 -- |Returns the 'CharData' contained within the given 'Content', or
  11 -- 'Nothing' if no acceptable CharData was found. It will parse either
  12 -- a 'CString' ('String') or 'CRef' (XML entity reference).
  13 get_char_data :: Content i -> (Maybe CharData)
  14 get_char_data (CString _ cd _) = Just cd
  15 get_char_data (CRef ref _) = Just (verbatim ref) -- Entities.
  16 get_char_data _ = Nothing
  17
  18
  19 -- |A 'CFilter' returning all top-level <status> elements.
  20 -- The name is due to the fact that if we retrieve more than
  21 -- one status, they will be wrapped in a <statuses> tag, and
  22 -- thus not be top-level.
  23 single_status :: CFilter i
  24 single_status = (tag "status")
  25
  26 -- |A 'CFilter' returning all <status> tags within <statuses>.
  27 all_statuses :: CFilter i
  28 all_statuses = (tag "statuses" /> tag "status")
  29
  30 -- |Finds the text of the <id> element contained within some other
  31 -- content. Called unique_id here because status_id is used elsewhere.
  32 unique_id :: CFilter i
  33 unique_id = keep /> (tag "id") /> txt
  34
  35 -- |Finds the text of the <created_at> element contained within some
  36 -- other element.
  37 status_created_at :: CFilter i
  38 status_created_at = keep /> (tag "created_at") /> txt
  39
  40 -- |Finds the text of the <text> element contained within some
  41 -- other element.
  42 status_text :: CFilter i
  43 status_text = keep /> (tag "text") /> txt
  44
  45 -- |Finds the XML of the <user> element contained within some other
  46 -- element.
  47 status_user :: CFilter i
  48 status_user = keep /> (tag "user")
  49
  50 -- | Finds the text of the <retweeted> element contained within some
  51 --   other element.
  52 status_retweeted :: CFilter i
  53 status_retweeted = keep /> (tag "retweeted") /> txt
  54
  55 -- | Finds the text of the <in_reply_to_status_id> element contained
  56 --   within some other element.
  57 status_reply_to_status_id :: CFilter i
  58 status_reply_to_status_id = keep /> (tag "in_reply_to_status_id") /> txt
  59
  60 -- |Finds the text of the <screen_name> element contained within some
  61 -- other element.
  62 user_screen_name :: CFilter i
  63 user_screen_name = keep /> (tag "screen_name") /> txt
  64
  65 -- |A wrapper around the 'read' function which returns either Nothing
  66 -- or (Just <the thing that could be read>).
  67 maybe_read :: (Read a) => String -> Maybe a
  68 maybe_read str =
  69     case (reads str) of
  70       [] -> Nothing
  71       ((y,_):_) -> Just y
  72
  73 -- |Takes a unicode codepoint in decimal and returns it as a
  74 -- one-character string.
  75 entity_from_codepoint :: String -> String
  76 entity_from_codepoint codepoint =
  77     case (maybe_read codepoint) of
  78       Nothing  -> ""
  79       Just num -> [(chr num)]
  80
  81
  82 -- |A list of tuples whose first entry is a regular expression
  83 -- matching XML entities, and whose second entry is the ASCII
  84 -- character represented by that entity.
  85 xml_entities :: [(String, String)]
  86 xml_entities = [("[lr]dquo", "\""),
  87                 ("quot",     "\""),
  88                 ("[mn]dash", "-"),
  89                 ("nbsp",     " "),
  90                 ("amp",      "&"),
  91                 ("lt",       "<"),
  92                 ("gt",       ">"),
  93                 ("hellip",   "…")]
  94
  95 -- |Replace all of the XML entities in target.
  96 replace_entities :: String -> String
  97 replace_entities target =
  98     unescape_numeric (unescape_recursive xml_entities target)
  99
 100 -- |Recursively unescape all numeric entities in the given String.
 101 unescape_numeric :: String -> String
 102 unescape_numeric target =
 103     case match of
 104       Nothing -> target
 105       Just subexprs ->
 106           case subexprs of
 107             []   -> target
 108             s1:_ ->
 109                 let this_entity_regex = mkRegex ("&#" ++ s1 ++ ";") in
 110                 let replacement = entity_from_codepoint s1 in
 111                 let new_target = subRegex this_entity_regex target replacement in
 112                 unescape_numeric new_target
 113     where
 114       from = "&#([0-9]+);"
 115       match = matchRegex (mkRegex from) target
 116
 117
 118
 119 -- |The recursive function which does the real work for
 120 -- 'replace_entities'.
 121 unescape_recursive :: [(String, String)] -> String -> String
 122 unescape_recursive [] target = target
 123 unescape_recursive replacements target =
 124     unescape_recursive (tail replacements) (subRegex (mkRegex from) target to)
 125     where
 126       replacement = (replacements !! 0)
 127       from = "&" ++ (fst replacement) ++ ";"
 128       to = (snd replacement)
 129
 130
 131
 132 xml_tests :: [Test]
 133 xml_tests = [ test_replace_entities ]
 134
 135
 136 test_replace_entities :: Test
 137 test_replace_entities =
 138     TestCase $ assertEqual "All entities are replaced correctly." expected_text actual_text
 139     where
 140       actual_text = (replace_entities "&quot;The moon is gay&#8230;&hellip;&quot; said &lt;insert the current president of the United States of America&gt;. &ldquo;It&#8217;s OK&mdash;&ndash;he&#8217;s not a real doctor.&rdquo;")
 141       expected_text = "\"The moon is gay……\" said <insert the current president of the United States of America>. \"It’s OK--he’s not a real doctor.\""