X-Git-Url: http://gitweb.fperrin.net/?a=blobdiff_plain;f=src%2Fcom%2Fhughes%2Fandroid%2Fdictionary%2Fparser%2FEnWiktionaryXmlParser.java;h=c6aee3b97d28cea455aef0d80e5483b483d308a3;hb=57d93b56ca2ffa3be718469a9f89f66b4716ad4e;hp=c2e6e7c2726e1f7ea5d8b5c0aa3d33fbd295bc2e;hpb=2f137177af2a16411d31280a22bd5dad6ba03dd2;p=DictionaryPC.git

diff --git a/src/com/hughes/android/dictionary/parser/EnWiktionaryXmlParser.java b/src/com/hughes/android/dictionary/parser/EnWiktionaryXmlParser.java
index c2e6e7c..c6aee3b 100644
--- a/src/com/hughes/android/dictionary/parser/EnWiktionaryXmlParser.java
+++ b/src/com/hughes/android/dictionary/parser/EnWiktionaryXmlParser.java
@@ -38,16 +38,17 @@ import com.hughes.android.dictionary.engine.PairEntry.Pair;
 
 public class EnWiktionaryXmlParser {
   
+  private static final String TRANSLITERATION_FORMAT = " (tr. %s)";
+
   static final Logger LOG = Logger.getLogger(EnWiktionaryXmlParser.class.getName());
   
-  // TODO: look for {{ and [[ and <adf> <!-- in output.
   // TODO: process {{ttbc}} lines
   
   static final Pattern partOfSpeechHeader = Pattern.compile(
       "Noun|Verb|Adjective|Adverb|Pronoun|Conjunction|Interjection|" +
       "Preposition|Proper noun|Article|Prepositional phrase|Acronym|" +
       "Abbreviation|Initialism|Contraction|Prefix|Suffix|Symbol|Letter|" +
-      "Ligature|Idiom|Phrase|" +
+      "Ligature|Idiom|Phrase|\\{\\{acronym\\}\\}|\\{\\{initialism\\}\\}|" +
       // These are @deprecated:
       "Noun form|Verb form|Adjective form|Nominal phrase|Noun phrase|" +
       "Verb phrase|Transitive verb|Intransitive verb|Reflexive verb|" +
@@ -118,8 +119,8 @@ public class EnWiktionaryXmlParser {
     heading = heading.replaceAll("=", "").trim(); 
     if (heading.equals("English")) {
       doEnglishWord(title, text);
-    } else if (langPattern.matcher(heading).matches()){
-      doForeignWord(title, text);
+    } else if (langPattern.matcher(heading).find()){
+      doForeignWord(heading, title, text);
     }
         
   }  // endPage()
@@ -145,6 +146,8 @@ public class EnWiktionaryXmlParser {
         if (partOfSpeechHeader.matcher(headerName).matches()) {
           posDepth = wikiTokenizer.headingDepth();
           pos = wikiTokenizer.headingWikiText();
+          // TODO: if we're inside the POS section, we should handle the first title line...
+          
         } else if (headerName.equals("Translations")) {
           if (pos == null) {
             LOG.warning("Translations without POS: " + title);
@@ -166,9 +169,10 @@ public class EnWiktionaryXmlParser {
   
   private void doTranslations(final String title, final WikiTokenizer wikiTokenizer, final String pos) {
     if (title.equals("absolutely")) {
-      System.out.println();
+      //System.out.println();
     }
     
+    String topLevelLang = null;
     String sense = null;
     boolean done = false;
     while (wikiTokenizer.nextToken() != null) {
@@ -182,8 +186,6 @@ public class EnWiktionaryXmlParser {
       
       // Check whether we care about this line:
       
-      //line = WikiLineReader.removeSquareBrackets(line);
-      
       if (wikiTokenizer.isFunction()) {
         final String functionName = wikiTokenizer.functionName();
         final List<String> positionArgs = wikiTokenizer.functionPositionArgs();
@@ -200,13 +202,14 @@ public class EnWiktionaryXmlParser {
           sense = null;
         } else if (functionName.equals("trans-mid")) {
         } else if (functionName.equals("trans-see")) {
-          // TODO
+          // TODO: would also be nice...
         } else if (functionName.startsWith("picdic")) {
         } else if (functionName.startsWith("checktrans")) {
+          done = true;
         } else if (functionName.startsWith("ttbc")) {
           wikiTokenizer.nextLine();
-          // TODO: would be great to handle
-          //TODO: Check this: done = true;
+          // TODO: would be great to handle ttbc
+          // TODO: Check this: done = true;
         } else {
           LOG.warning("Unexpected translation wikifunction: " + wikiTokenizer.token() + ", title=" + title);
         }
@@ -214,6 +217,10 @@ public class EnWiktionaryXmlParser {
         final String line = wikiTokenizer.listItemWikiText();
         // This line could produce an output...
         
+        if (line.contains("ich hoan dich gear")) {
+          //System.out.println();
+        }
+        
         // First strip the language and check whether it matches.
         // And hold onto it for sub-lines.
         final int colonIndex = line.indexOf(":");
@@ -221,16 +228,28 @@ public class EnWiktionaryXmlParser {
           continue;
         }
         
-        final String lang = line.substring(0, colonIndex);
-        if (!this.langPattern.matcher(lang).find()) {
+        final String lang = trim(WikiTokenizer.toPlainText(line.substring(0, colonIndex)));
+        final boolean appendLang;
+        if (wikiTokenizer.listItemPrefix().length() == 1) {
+          topLevelLang = lang;
+          final boolean thisFind = langPattern.matcher(lang).find();
+          if (!thisFind) {
+            continue;
+          }
+          appendLang = !langPattern.matcher(lang).matches();
+        } else if (topLevelLang == null) {
           continue;
+        } else {
+          // Two-level -- the only way we won't append is if this second level matches exactly.
+          if (!langPattern.matcher(lang).matches() && !langPattern.matcher(topLevelLang).find()) {
+            continue;
+          }
+          appendLang = !langPattern.matcher(lang).matches();
         }
         
         String rest = line.substring(colonIndex + 1).trim();
         if (rest.length() > 0) {
-          doTranslationLine(line, title, pos, sense, rest);
-        } else {
-          // TODO: do lines that are like Greek:
+          doTranslationLine(line, appendLang ? lang : null, title, pos, sense, rest);
         }
         
       } else if (wikiTokenizer.remainderStartsWith("''See''")) {
@@ -259,7 +278,7 @@ public class EnWiktionaryXmlParser {
     return index < list.size() ? list.get(index) : null;
   }
   
-  private void doTranslationLine(final String line, final String title, final String pos, final String sense, final String rest) {
+  private void doTranslationLine(final String line, final String lang, final String title, final String pos, final String sense, final String rest) {
     // Good chance we'll actually file this one...
     final PairEntry pairEntry = new PairEntry();
     final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
@@ -302,17 +321,21 @@ public class EnWiktionaryXmlParser {
               otherText.append(String.format(" {%s}", gender));
             }
             if (transliteration != null) {
-              otherText.append(String.format(" (tr. %s)", transliteration));
+              otherText.append(String.format(TRANSLITERATION_FORMAT, transliteration));
               otherIndexBuilder.addEntryWithString(indexedEntry, transliteration, EntryTypeName.WIKTIONARY_TRANSLITERATION);
             }
           //}
         } else if (functionName.equals("qualifier")) {
-          String qualifier = args.get(0);
-          if (!namedArgs.isEmpty() || args.size() > 1) {
-            LOG.warning("weird qualifier: " + line);
+          if (args.size() == 0) {
+           otherText.append(wikiTokenizer.token()); 
+          } else { 
+            String qualifier = args.get(0);
+            if (!namedArgs.isEmpty() || args.size() > 1) {
+              LOG.warning("weird qualifier: " + line);
+            }
+            // Unindexed!
+            otherText.append("(").append(qualifier).append(")");
           }
-          // Unindexed!
-          otherText.append("(").append(qualifier).append(")");
         } else if (encodings.contains(functionName)) {
           otherText.append("").append(args.get(0));
           otherIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
@@ -360,6 +383,10 @@ public class EnWiktionaryXmlParser {
       return;
     }
     
+    if (lang != null) {
+      otherText.insert(0, String.format("(%s) ", lang));
+    }
+    
     StringBuilder englishText = new StringBuilder();
     
     englishText.append(title);
@@ -403,7 +430,7 @@ public class EnWiktionaryXmlParser {
   
   // -------------------------------------------------------------------------
   
-  private void doForeignWord(final String title, final String text) {
+  private void doForeignWord(final String lang, final String title, final String text) {
     final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
     while (wikiTokenizer.nextToken() != null) {
       if (wikiTokenizer.isHeading()) {
@@ -413,7 +440,7 @@ public class EnWiktionaryXmlParser {
         } else if (headingName.equals("Pronunciation")) {
           //doPronunciation(wikiLineReader);
         } else if (partOfSpeechHeader.matcher(headingName).matches()) {
-          doForeignPartOfSpeech(title, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
+          doForeignPartOfSpeech(lang, title, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
         }
       } else {
       }
@@ -439,9 +466,9 @@ public class EnWiktionaryXmlParser {
 
 
   int foreignCount = 0;
-  private void doForeignPartOfSpeech(String title, final String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
+  private void doForeignPartOfSpeech(final String lang, String title, final String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
     if (++foreignCount % 1000 == 0) {
-      LOG.info("***" + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
+      LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
     }
     if (title.equals("moro")) {
       System.out.println();
@@ -467,7 +494,7 @@ public class EnWiktionaryXmlParser {
       }
       
       if (currentHeadingDepth > posDepth) {
-        // TODO
+        // TODO: deal with other neat info sections
         continue;
       }
       
@@ -500,6 +527,59 @@ public class EnWiktionaryXmlParser {
         } else {
           //foreignBuilder.append(title);
         }
+      } else if (name.equals("attention") || name.equals("zh-attention")) {
+        // See: http://en.wiktionary.org/wiki/Template:attention
+        // Ignore these.
+      } else if (name.equals("infl")) {
+        // See: http://en.wiktionary.org/wiki/Template:infl
+        final String langCode = get(args, 0);
+        namedArgs.remove("sc");
+        final String tr = namedArgs.remove("tr");
+        final String g = namedArgs.remove("g");
+        final String g2 = namedArgs.remove("g2");
+        final String g3 = namedArgs.remove("g3");
+        if (!namedArgs.isEmpty()) {
+          LOG.warning("Didn't parse infl: " + wikiTokenizer.token());
+          foreignBuilder.append(wikiTokenizer.token());
+        } else {
+          String head = namedArgs.get("head");
+          if (head == null) {
+            head = title;
+          } else {
+            head = WikiTokenizer.toPlainText(head);
+          }
+          foreignBuilder.append(head);
+    
+          if (g != null) {
+            foreignBuilder.append(" {").append(g);
+            if (g2 != null) {
+              foreignBuilder.append("|").append(g2);
+            }
+            if (g3 != null) {
+              foreignBuilder.append("|").append(g3);
+            }
+            foreignBuilder.append("}");
+          }
+    
+          if (tr != null) {
+            foreignBuilder.append(String.format(TRANSLITERATION_FORMAT, tr));
+            wordForms.add(tr);
+          }
+    
+          final String pos = get(args, 1);
+          if (pos != null) {
+            foreignBuilder.append(" (").append(pos).append(")");
+          }
+          for (int i = 2; i < args.size(); i += 2) {
+            final String inflName = get(args, i);
+            final String inflValue = get(args, i + 1);
+            foreignBuilder.append(", ").append(WikiTokenizer.toPlainText(inflName));
+            if (inflValue != null && inflValue.length() > 0) {
+              foreignBuilder.append(": ").append(WikiTokenizer.toPlainText(inflValue));
+              wordForms.add(inflValue);
+            }
+          }
+        }
       } else if (name.equals("it-noun")) {
           final String base = get(args, 0);
           final String gender = get(args, 1);
@@ -556,10 +636,13 @@ public class EnWiktionaryXmlParser {
     
     } finally {
       // Here's where we exit.
-      // TODO: Should we make an entry even if there are no foreign list items?
+      // Should we make an entry even if there are no foreign list items?
       String foreign = foreignBuilder.toString().trim();
       if (!foreign.toLowerCase().startsWith(title.toLowerCase())) {
-        foreign = title + " " + foreign;
+        foreign = String.format("%s %s", title, foreign);
+      }
+      if (!langPattern.matcher(lang).matches()) {
+        foreign = String.format("(%s) %s", lang, foreign);
       }
       for (final ListSection listSection : listSections) {
         doForeignListItem(foreign, title, wordForms, listSection);
@@ -579,6 +662,7 @@ public class EnWiktionaryXmlParser {
     
     final String prefix = listSection.firstPrefix;
     if (prefix.length() > 1) {
+      // Could just get looser and say that any prefix longer than first is a sublist.
       LOG.warning("Prefix too long: " + listSection);
       return;
     }
@@ -666,50 +750,66 @@ public class EnWiktionaryXmlParser {
         mdashLen = 3;
       }
       
-      // TODO: index and clean these!!!
-      if (nextPrefix.equals("#:") && dash != -1) {
+      if ((nextPrefix.equals("#:") || nextPrefix.equals("##:")) && dash != -1) {
         final String foreignEx = nextLine.substring(0, dash);
         final String englishEx = nextLine.substring(dash + mdashLen);
-        final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder), formatAndIndexExampleString(foreignEx, otherIndexBuilder), swap);
-        pairEntry.pairs.add(pair);
+        final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder, indexedEntry), formatAndIndexExampleString(foreignEx, otherIndexBuilder, indexedEntry), swap);
+        if (pair.lang1 != "--" && pair.lang1 != "--") {
+          pairEntry.pairs.add(pair);
+        }
         lastForeign = null;
-      } else if (nextPrefix.equals("#:")){
-        final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null), swap);
+      } else if (nextPrefix.equals("#:") || nextPrefix.equals("##:")){
+        final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
         lastForeign = nextLine;
-        pairEntry.pairs.add(pair);
+        if (pair.lang1 != "--" && pair.lang1 != "--") {
+          pairEntry.pairs.add(pair);
+        }
       } else if (nextPrefix.equals("#::") || nextPrefix.equals("#**")) {
-        if (lastForeign != null) {
+        if (lastForeign != null && pairEntry.pairs.size() > 0) {
           pairEntry.pairs.remove(pairEntry.pairs.size() - 1);
-          final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder), formatAndIndexExampleString(lastForeign, otherIndexBuilder), swap);
-          pairEntry.pairs.add(pair);
+          final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder, indexedEntry), formatAndIndexExampleString(lastForeign, otherIndexBuilder, indexedEntry), swap);
+          if (pair.lang1 != "--" && pair.lang1 != "--") {
+            pairEntry.pairs.add(pair);
+          }
+          lastForeign = null;
         } else {
-          LOG.warning("English example with no foreign: " + title + ", " + nextLine);
+          LOG.warning("TODO: English example with no foreign: " + title + ", " + nextLine);
+          // TODO: add something.
         }
       } else if (nextPrefix.equals("#*")) {
         // Can't really index these.
-        final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null), swap);
+        final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
         lastForeign = nextLine;
-        pairEntry.pairs.add(pair);
-      } else if (nextPrefix.equals("#::*")) {
-        final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null), swap);
-        pairEntry.pairs.add(pair);
-      } else {
-        assert false;
+        if (pair.lang1 != "--" && pair.lang1 != "--") {
+          pairEntry.pairs.add(pair);
+        }
+      } else if (nextPrefix.equals("#::*") || nextPrefix.equals("##") || nextPrefix.equals("#*:") || nextPrefix.equals("#:*") || true) {
+        final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
+        if (pair.lang1 != "--" && pair.lang1 != "--") {
+          pairEntry.pairs.add(pair);
+        }
+//      } else {
+//        assert false;
       }
     }
   }
   
-  private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder) {
+  private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder, final IndexedEntry indexedEntry) {
     final WikiTokenizer wikiTokenizer = new WikiTokenizer(example, false);
     final StringBuilder builder = new StringBuilder();
     boolean insideTripleQuotes = false;
     while (wikiTokenizer.nextToken() != null) {
       if (wikiTokenizer.isPlainText()) {
         builder.append(wikiTokenizer.token());
-        
+        if (indexBuilder != null) {
+          indexBuilder.addEntryWithString(indexedEntry, wikiTokenizer.token(), EntryTypeName.WIKTIONARY_EXAMPLE);
+        }
       } else if (wikiTokenizer.isWikiLink()) {
-        builder.append(wikiTokenizer.wikiLinkText());
-        
+        final String text = wikiTokenizer.wikiLinkText().replaceAll("'", ""); 
+        builder.append(text);
+        if (indexBuilder != null) {
+          indexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_EXAMPLE);
+        }
       } else if (wikiTokenizer.isFunction()) {
         builder.append(wikiTokenizer.token());
       } else if (wikiTokenizer.isMarkup()) {
@@ -722,7 +822,8 @@ public class EnWiktionaryXmlParser {
         LOG.warning("unexpected token: " + wikiTokenizer.token());
       }
     }
-    return trim(builder.toString()); 
+    final String result = trim(builder.toString());
+    return result.length() > 0 ? result : "--";
   }