- Timestamp:
- 2019-10-17T23:12:38+13:00 (5 years ago)
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
gs3-extensions/maori-lang-detection/src/org/greenstone/atea/TextDumpPage.java
r33580 r33582 77 77 } 78 78 else { // dealing with the rest of the page dump's metadata 79 // QTODO: nutch's text dump output is problematic 80 // strange characters are in the stream and end up here 81 // and can't detect end of metadata or even end of line. 79 82 endIndex = line.indexOf(":"); 80 83 if(endIndex != -1) { … … 83 86 tuples.put(k.trim(), v.trim()); 84 87 } else { 85 error("No meta key for meta: " + line); 88 if(NutchTextDumpProcessor.DEBUG_MODE) { 89 error("No meta key for meta: " + line); 90 error(unparsedPageDump); 91 } 86 92 } 87 93 } … … 102 108 103 109 // START DEBUG 104 debug("__________________________________________"); 105 for(Map.Entry<String, String> entry : tuples.entrySet()) { 106 String key = entry.getKey(); 107 String value = entry.getValue(); 108 debug(key + " - " + value); 109 } 110 debug("__________________________________________"); 110 debugTuples(); 111 111 // END DEBUG 112 112 113 113 } 114 114 115 public void debugTuples() { 116 if(NutchTextDumpProcessor.DEBUG_MODE) { 117 debug("__________________________________________"); 118 for(Map.Entry<String, String> entry : tuples.entrySet()) { 119 String key = entry.getKey(); 120 String value = entry.getValue(); 121 debug(key + " - " + value); 122 } 123 debug("__________________________________________"); 124 } 125 } 126 115 127 116 128 public String getPageURL() { 117 return tuples.get(" url");129 return tuples.get("pageURL"); 118 130 } 119 131 … … 126 138 } 127 139 140 public void add(String key, String value) { 141 tuples.put(key, value); 142 } 143 144 public void addMRILanguageStatus(boolean status) { 145 if(status) { 146 tuples.put("isMRI", "true"); 147 } else { 148 tuples.put("isMRI", "false"); 149 } 150 } 151 152 public boolean getMRILanguageStatus() { 153 String value = tuples.get("isMRI"); 154 if(value == null) { 155 return false; 156 } 157 if(value.equals("true")) { 158 return true; 159 } 160 else { 161 return false; 162 } 163 164 } 165 128 166 // --------------- STATIC METHODS AND INNER CLASSED USED BY MAIN -------------- // 129 167 public static void info(String msg) {
Note:
See TracChangeset
for help on using the changeset viewer.