X-Git-Url: http://gitweb.michael.orlitzky.com/?p=dead%2Fhalcyon.git;a=blobdiff_plain;f=src%2FTwitter%2FXml.hs;h=e284a0da5ae6601c2a8fac8ee0a9730989325773;hp=e4cf5aceffa1bff8a1f9e8b5c824702cb5d067fb;hb=ee2deb5a51e416d607cce45ddd10d4e19c050771;hpb=94484087fbfe98d6735aa82798a9bf506f97fd19

diff --git a/src/Twitter/Xml.hs b/src/Twitter/Xml.hs
index e4cf5ac..e284a0d 100644
--- a/src/Twitter/Xml.hs
+++ b/src/Twitter/Xml.hs
@@ -1,48 +1,113 @@
+-- |Application-specific XML functions.
 module Twitter.Xml
 where
 
-import Data.Maybe
-import Text.Regex (mkRegex, subRegex)
+import Data.Char (chr)
+import Test.HUnit
+import Text.Regex (matchRegex, mkRegex, subRegex)
 import Text.XML.HaXml
 
+-- |Returns the 'CharData' contained within the given 'Content', or
+-- 'Nothing' if no acceptable CharData was found. It will parse either
+-- a 'CString' ('String') or 'CRef' (XML entity reference).
 get_char_data :: Content -> (Maybe CharData)
 get_char_data (CString _ cd) = Just cd
 get_char_data (CRef ref) = Just (verbatim ref) -- Entities.
 get_char_data _ = Nothing
 
 
+-- |A 'CFilter' returning all top-level <status> elements.
+-- The name is due to the fact that if we retrieve more than
+-- one status, they will be wrapped in a <statuses> tag, and
+-- thus not be top-level.
+single_status :: CFilter
+single_status = (tag "status")
+
+-- |A 'CFilter' returning all <status> tags within <statuses>.
 all_statuses :: CFilter
 all_statuses = (tag "statuses" /> tag "status")
 
--- Called unique_id here because status_id is used elsewhere.
+-- |Finds the text of the <id> element contained within some other
+-- content. Called unique_id here because status_id is used elsewhere.
 unique_id :: CFilter
 unique_id = keep /> (tag "id") /> txt
 
+-- |Finds the text of the <created_at> element contained within some
+-- other element.
 status_created_at :: CFilter
 status_created_at = keep /> (tag "created_at") /> txt
 
+-- |Finds the text of the <text> element contained within some
+-- other element.
 status_text :: CFilter
 status_text = keep /> (tag "text") /> txt
 
+-- |Finds the XML of the <user> element contained within some other
+-- element.
 status_user :: CFilter
 status_user = keep /> (tag "user")
 
+-- |Finds the text of the <screen_name> element contained within some
+-- other element.
 user_screen_name :: CFilter
 user_screen_name = keep /> (tag "screen_name") /> txt
 
+-- |A wrapper around the 'read' function which returns either Nothing
+-- or (Just <the thing that could be read>).
+maybe_read :: (Read a) => String -> Maybe a
+maybe_read str =
+    case (reads str) of
+      [] -> Nothing
+      ((y,_):_) -> Just y
+
+-- |Takes a unicode codepoint in decimal and returns it as a
+-- one-character string.
+entity_from_codepoint :: String -> String
+entity_from_codepoint codepoint =
+    case (maybe_read codepoint) of
+      Nothing  -> ""
+      Just num -> [(chr num)]
 
+
+-- |A list of tuples whose first entry is a regular expression
+-- matching XML entities, and whose second entry is the ASCII
+-- character represented by that entity.
 xml_entities :: [(String, String)]
 xml_entities = [("[lr]dquo", "\""),
+                ("quot",     "\""),
                 ("[mn]dash", "-"),
                 ("nbsp",     " "),
-                ("#8217",    "'"),
                 ("amp",      "&"),
                 ("lt",       "<"),
-                ("gt",       ">")]
+                ("gt",       ">"),
+                ("hellip",   "…")]
 
+-- |Replace all of the XML entities in target.
 replace_entities :: String -> String
-replace_entities target = unescape_recursive xml_entities target
+replace_entities target =
+    unescape_numeric (unescape_recursive xml_entities target)
+
+-- |Recursively unescape all numeric entities in the given String.
+unescape_numeric :: String -> String
+unescape_numeric target =
+    case match of
+      Nothing -> target
+      Just subexprs ->
+          case subexprs of
+            []   -> target
+            s1:_ ->
+                let this_entity_regex = mkRegex ("&#" ++ s1 ++ ";") in
+                let replacement = entity_from_codepoint s1 in
+                let new_target = subRegex this_entity_regex target replacement in
+                unescape_numeric new_target
+    where
+      from = "&#([0-9]+);"
+      match = matchRegex (mkRegex from) target
+
+
 
+-- |The recursive function which does the real work for
+-- 'replace_entities'.
 unescape_recursive :: [(String, String)] -> String -> String
 unescape_recursive [] target = target
 unescape_recursive replacements target =
@@ -51,3 +116,16 @@ unescape_recursive replacements target =
       replacement = (replacements !! 0)
       from = "&" ++ (fst replacement) ++ ";"
       to = (snd replacement)
+
+
+
+xml_tests :: [Test]
+xml_tests = [ test_replace_entities ]
+
+
+test_replace_entities :: Test
+test_replace_entities =
+    TestCase $ assertEqual "All entities are replaced correctly." expected_text actual_text
+    where
+      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;")
+      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.\""