private static void loadPairs(
HMapIV<ArrayListOfIntsWritable> pwsimMapping, int langID, JobConf job, Reporter reporter) {
try {
Path[] localFiles = DistributedCache.getLocalCacheFiles(job);
String pwsimFile = job.get("PwsimPairs");
for (Path localFile : localFiles) {
if (localFile.toString().contains(getFilename(pwsimFile))) {
SequenceFile.Reader reader =
new SequenceFile.Reader(FileSystem.getLocal(job), localFile, job);
PairOfInts key = (PairOfInts) reader.getKeyClass().newInstance();
IntWritable value = (IntWritable) reader.getValueClass().newInstance();
int cnt = 0;
while (reader.next(key, value)) {
int fDocno = key.getRightElement();
int eDocno = key.getLeftElement();
if ((eDocno == 6127 && fDocno == 1000000074)
|| (eDocno == 6127 && fDocno == 1000000071)) {
sLogger.info(key);
}
if (langID == CLIRUtils.E) {
if (!pwsimMapping.containsKey(eDocno)) {
pwsimMapping.put(eDocno, new ArrayListOfIntsWritable());
}
pwsimMapping
.get(eDocno)
.add(
fDocno); // we add 1000000000 to foreign docnos to distinguish them during
// pwsim algo
} else {
if (!pwsimMapping.containsKey(fDocno)) {
pwsimMapping.put(fDocno, new ArrayListOfIntsWritable());
}
pwsimMapping
.get(fDocno)
.add(
eDocno); // we add 1000000000 to foreign docnos to distinguish them during
// pwsim algo
}
cnt++;
key = (PairOfInts) reader.getKeyClass().newInstance();
value = (IntWritable) reader.getValueClass().newInstance();
}
reader.close();
sLogger.info(pwsimMapping.size() + "," + cnt + " pairs loaded from " + localFile);
}
}
} catch (Exception e) {
throw new RuntimeException(e);
}
}
@Test
public void testIterable() {
Int2IntFrequencyDistribution fd = new Int2IntFrequencyDistributionOpen();
fd.set(1, 1);
fd.set(4, 3);
fd.set(2, 4);
fd.set(5, 7);
fd.set(6, 9);
fd.set(3, 2);
assertEquals(6, fd.getNumberOfEvents());
assertEquals(26, fd.getSumOfCounts());
SortedSet<PairOfInts> list = new TreeSet<PairOfInts>();
for (PairOfInts pair : fd) {
list.add(pair.clone());
}
assertEquals(6, list.size());
Iterator<PairOfInts> iter = list.iterator();
PairOfInts e = iter.next();
assertEquals(1, e.getLeftElement());
assertEquals(1, e.getRightElement());
e = iter.next();
assertEquals(2, e.getLeftElement());
assertEquals(4, e.getRightElement());
e = iter.next();
assertEquals(3, e.getLeftElement());
assertEquals(2, e.getRightElement());
e = iter.next();
assertEquals(4, e.getLeftElement());
assertEquals(3, e.getRightElement());
e = iter.next();
assertEquals(5, e.getLeftElement());
assertEquals(7, e.getRightElement());
e = iter.next();
assertEquals(6, e.getLeftElement());
assertEquals(9, e.getRightElement());
}
/** Runs this tool. */
@SuppressWarnings({"static-access"})
public int run(String[] args) throws Exception {
Options options = new Options();
options.addOption(
OptionBuilder.withArgName("path").hasArg().withDescription("input path").create(INDEX));
options.addOption(
OptionBuilder.withArgName("path")
.hasArg()
.withDescription("output path")
.create(COLLECTION));
CommandLine cmdline = null;
CommandLineParser parser = new GnuParser();
try {
cmdline = parser.parse(options, args);
} catch (ParseException exp) {
System.err.println("Error parsing command line: " + exp.getMessage());
System.exit(-1);
}
if (!cmdline.hasOption(INDEX) || !cmdline.hasOption(COLLECTION)) {
System.out.println("args: " + Arrays.toString(args));
HelpFormatter formatter = new HelpFormatter();
formatter.setWidth(120);
formatter.printHelp(LookupPostingsCompressed.class.getName(), options);
ToolRunner.printGenericCommandUsage(System.out);
System.exit(-1);
}
String indexPath = cmdline.getOptionValue(INDEX);
String collectionPath = cmdline.getOptionValue(COLLECTION);
if (collectionPath.endsWith(".gz")) {
System.out.println("gzipped collection is not seekable: use compressed version!");
System.exit(-1);
}
Configuration config = new Configuration();
FileSystem fs = FileSystem.get(config);
MapFile.Reader reader = new MapFile.Reader(new Path(indexPath + "/part-r-00000"), config);
FSDataInputStream collection = fs.open(new Path(collectionPath));
BufferedReader d = new BufferedReader(new InputStreamReader(collection));
Text key = new Text();
ArrayListWritable<PairOfInts> postings;
BytesWritable bytesValue = new BytesWritable();
System.out.println("Looking up postings for the term \"starcross'd\"");
key.set("starcross'd");
reader.get(key, bytesValue);
postings = deserializePosting(bytesValue);
// ArrayListWritable<PairOfVInts> postings = value;
for (PairOfInts pair : postings) {
System.out.println(pair);
collection.seek(pair.getLeftElement());
System.out.println(d.readLine());
}
bytesValue = new BytesWritable();
key.set("gold");
reader.get(key, bytesValue);
postings = deserializePosting(bytesValue);
System.out.println(
"Complete postings list for 'gold': (" + postings.size() + ", " + postings + ")");
Int2IntFrequencyDistribution goldHist = new Int2IntFrequencyDistributionEntry();
// postings = value;
for (PairOfInts pair : postings) {
goldHist.increment(pair.getRightElement());
}
System.out.println("histogram of tf values for gold");
for (PairOfInts pair : goldHist) {
System.out.println(pair.getLeftElement() + "\t" + pair.getRightElement());
}
bytesValue = new BytesWritable();
key.set("silver");
reader.get(key, bytesValue);
postings = deserializePosting(bytesValue);
System.out.println(
"Complete postings list for 'silver': (" + postings.size() + ", " + postings + ")");
Int2IntFrequencyDistribution silverHist = new Int2IntFrequencyDistributionEntry();
// postings = value;
for (PairOfInts pair : postings) {
silverHist.increment(pair.getRightElement());
}
System.out.println("histogram of tf values for silver");
for (PairOfInts pair : silverHist) {
System.out.println(pair.getLeftElement() + "\t" + pair.getRightElement());
}
bytesValue = new BytesWritable();
key.set("bronze");
Writable w = reader.get(key, bytesValue);
if (w == null) {
System.out.println("the term bronze does not appear in the collection");
}
collection.close();
reader.close();
return 0;
}
