1 | package org.greenstone.atea;
|
---|
2 |
|
---|
3 |
|
---|
4 | import java.io.*;
|
---|
5 | import java.util.Properties;
|
---|
6 | import java.util.zip.GZIPInputStream;
|
---|
7 | import java.util.Iterator;
|
---|
8 | import java.util.Set;
|
---|
9 | import java.util.TreeSet;
|
---|
10 |
|
---|
11 | import org.apache.log4j.Logger;
|
---|
12 |
|
---|
13 | /**
|
---|
14 | * The main() method of this class takes a folder of warc.wet(.gz) files and goes through
|
---|
15 | * the WET records in each, putting each WET record into a file. Each file is put into a
|
---|
16 | * keep or discard folder, based on content-length and number of lines.
|
---|
17 | * A single instance of the WETProcessor class processes a single unzipped warc.wet file.
|
---|
18 | *
|
---|
19 | * To compile, including the jars in lib/ for compiling.
|
---|
20 | * maori-lang-detection/src$ javac -cp ".:../lib/*" org/greenstone/atea/CCWETProcessor.java
|
---|
21 | *
|
---|
22 | * To run, passing the log4j and other properties files in conf/ folder:
|
---|
23 | * maori-lang-detection/src$ java -cp ".:../conf:../lib/*" org.greenstone.atea.CCWETProcessor <folder containing warc.wet(.gz) files> <outputFolder>
|
---|
24 | *
|
---|
25 | * e.g.
|
---|
26 | * - java -cp ".:../conf:../lib/*" org.greenstone.atea.CCWETProcessor ../tmp/processWET /Scratch/ak19/gs3-extensions/maori-lang-detection/tmp/processedWET
|
---|
27 | * - java -cp ".:../conf:../lib/*" org.greenstone.atea.CCWETProcessor ../tmp/processWET /Scratch/ak19/gs3-extensions/maori-lang-detection/tmp/processedWET 2>&1 | less
|
---|
28 | *
|
---|
29 | */
|
---|
30 |
|
---|
31 | public class CCWETProcessor {
|
---|
32 | private static Logger logger = Logger.getLogger(org.greenstone.atea.CCWETProcessor.class.getName());
|
---|
33 |
|
---|
34 | // Properties shared across WETProcessor instances
|
---|
35 | public final int MAX_WORD_LENGTH;
|
---|
36 | public final int MIN_NUM_WORDS;
|
---|
37 | public final int MAX_WORDS_CAMELCASE;
|
---|
38 |
|
---|
39 | private Properties configProperties = new Properties();
|
---|
40 |
|
---|
41 | // File paths shared across WETProcessor instances
|
---|
42 | public final File WETFilesDir;
|
---|
43 | public final File outputFolder;
|
---|
44 | public final File discardFolder;
|
---|
45 | public final File keepFolder;
|
---|
46 | public final File keepURLsFile;
|
---|
47 | public final File discardURLsFile;
|
---|
48 |
|
---|
49 | // Keep a count of all the records that all WETProcessors instantiated
|
---|
50 | // by our main method combined have processed
|
---|
51 | private int totalRecordCount = 0;
|
---|
52 |
|
---|
53 | private int wetFileCount = 0;
|
---|
54 |
|
---|
55 | public CCWETProcessor(File inFolder, File outFolder) {
|
---|
56 | this.WETFilesDir = inFolder;
|
---|
57 | this.outputFolder = outFolder;
|
---|
58 |
|
---|
59 | // load up the properties from the config file
|
---|
60 | try (InputStream infile = org.greenstone.atea.WETProcessor.class.getClassLoader().getResourceAsStream("config.properties")) {
|
---|
61 | configProperties = new Properties();
|
---|
62 | configProperties.load(infile);
|
---|
63 | //infile.close();
|
---|
64 |
|
---|
65 | } catch(Exception e) {
|
---|
66 | System.err.println("Exception attempting to read properties from config.properties.");
|
---|
67 | logger.error("Exception attempting to read properties from config.properties.");
|
---|
68 | e.printStackTrace();
|
---|
69 | }
|
---|
70 |
|
---|
71 | if(configProperties.size() == 0) {
|
---|
72 | System.err.println("*** Warning: no values read into config properties. Using defaults.");
|
---|
73 | }
|
---|
74 |
|
---|
75 | MAX_WORD_LENGTH = Integer.parseInt(configProperties.getProperty("WETprocessor.max.word.length", "15"));
|
---|
76 | MIN_NUM_WORDS = Integer.parseInt(configProperties.getProperty("WETprocessor.min.num.words", "20"));
|
---|
77 | MAX_WORDS_CAMELCASE = Integer.parseInt(configProperties.getProperty("WETprocessor.max.words.camelcase", "10"));
|
---|
78 |
|
---|
79 |
|
---|
80 | this.discardFolder = new File(outFolder, "discard");
|
---|
81 | if(!discardFolder.exists()) {
|
---|
82 | discardFolder.mkdir();
|
---|
83 | }
|
---|
84 | this.keepFolder = new File(outFolder, "keep");
|
---|
85 | if(!keepFolder.exists()) {
|
---|
86 | keepFolder.mkdir();
|
---|
87 | }
|
---|
88 |
|
---|
89 | this.keepURLsFile = new File(outFolder, "keepURLs.txt");
|
---|
90 | if(keepURLsFile.exists() && !keepURLsFile.delete()) {
|
---|
91 | System.err.println("Warning: Unable to delete " + this.keepURLsFile + ". Unable to proceed.");
|
---|
92 | //return;
|
---|
93 | }
|
---|
94 | this.discardURLsFile = new File(outFolder, "discardURLs.txt");
|
---|
95 | if(discardURLsFile.exists() && !discardURLsFile.delete()) {
|
---|
96 | System.err.println("Warning Unable to delete " + discardURLsFile + ". Unable to proceed.");
|
---|
97 | //return;
|
---|
98 | }
|
---|
99 |
|
---|
100 | }
|
---|
101 |
|
---|
102 | /**
|
---|
103 | * Takes as input the keepURLs.txt file generated by running WETProcessor instances.
|
---|
104 | * As output produces the URL seed list and regex-urlfilter text files required by nutch,
|
---|
105 | * https://cwiki.apache.org/confluence/display/nutch/NutchTutorial
|
---|
106 | */
|
---|
107 | public void createSeedURLsFiles(File seedURLsFile, File urlFilterFile) {
|
---|
108 | // Maintain Sets of unique domains and urls
|
---|
109 | // TreeSet: by default, "the elements are ordered using their natural ordering"
|
---|
110 | // (or by a Comparator provided at set creation time).
|
---|
111 | // Whereas HashSet doesn't guarantee ordering.
|
---|
112 | // So we get alphabetic sorting for free. And guaranteed log(n) for basic operations.
|
---|
113 |
|
---|
114 | Set<String> domainsSet = new TreeSet<String>();
|
---|
115 | Set<String> urlsSet = new TreeSet<String>();
|
---|
116 |
|
---|
117 | final String FILTER_REGEX_PREFIX = "+https?://([a-z0-9-]+\\.)*"; // https?://([a-z0-9-]+\.)*
|
---|
118 |
|
---|
119 | try (
|
---|
120 | BufferedReader reader = new BufferedReader(new FileReader(this.keepURLsFile));
|
---|
121 | ) {
|
---|
122 |
|
---|
123 | // read a URL at a time from urlsFile
|
---|
124 | String url = null;
|
---|
125 | String domain = null;
|
---|
126 | while((url = reader.readLine()) != null) { // readLine removes newline separator
|
---|
127 |
|
---|
128 | // work out domain. This retains any www. or subdomain prefix:
|
---|
129 | int startIndex = url.indexOf("//"); // http:// or https:// prefix
|
---|
130 | startIndex = (startIndex == -1) ? 0 : (startIndex+2); // skip past the protocol's // portion
|
---|
131 | domain = url.substring(startIndex);
|
---|
132 | int endIndex = domain.indexOf("/");
|
---|
133 | if(endIndex == -1) endIndex = domain.length();
|
---|
134 | domain = domain.substring(0, endIndex);
|
---|
135 |
|
---|
136 | //if(!domainsMap.containsKey(domain)) {
|
---|
137 | urlsSet.add(url);
|
---|
138 | domainsSet.add(domain);
|
---|
139 | //}
|
---|
140 | }
|
---|
141 | } catch (IOException ioe) {
|
---|
142 | ioe.printStackTrace();
|
---|
143 | System.err.println("\n@@@@@@@@@ Error reading in urls from file " + this.keepURLsFile);
|
---|
144 | }
|
---|
145 |
|
---|
146 | try (BufferedWriter seedURLsWriter = new BufferedWriter(new FileWriter(seedURLsFile))) {
|
---|
147 | Iterator<String> i = urlsSet.iterator();
|
---|
148 | while(i.hasNext()) {
|
---|
149 | String url = i.next();
|
---|
150 | seedURLsWriter.write(url + "\n");
|
---|
151 | }
|
---|
152 |
|
---|
153 | } catch (IOException ioe) {
|
---|
154 | ioe.printStackTrace();
|
---|
155 | System.err.println("\n@@@@@@@@@ Error writing to " + seedURLsFile);
|
---|
156 | }
|
---|
157 |
|
---|
158 | try (BufferedWriter urlFilterWriter = new BufferedWriter(new FileWriter(urlFilterFile))) {
|
---|
159 | Iterator<String> i = domainsSet.iterator();
|
---|
160 | // nutch.apache.org => +^https?://([a-z0-9-]+\.)*nutch\.apache\.org/
|
---|
161 | while(i.hasNext()) {
|
---|
162 | String domain = i.next();
|
---|
163 | domain = FILTER_REGEX_PREFIX + domain.replace(".", "\\.") + "/";
|
---|
164 | urlFilterWriter.write(domain + "\n");
|
---|
165 | }
|
---|
166 |
|
---|
167 | } catch (IOException ioe) {
|
---|
168 | ioe.printStackTrace();
|
---|
169 | System.err.println("\n@@@@@@@@@ Error writing to " + urlFilterFile);
|
---|
170 | }
|
---|
171 | }
|
---|
172 |
|
---|
173 | /*
|
---|
174 | public boolean isBlacklisted(String url) {
|
---|
175 | return false;
|
---|
176 | }
|
---|
177 | */
|
---|
178 |
|
---|
179 | public boolean isGreylisted(String url) {
|
---|
180 | // alexa top sites and auto-translated product sites
|
---|
181 | return false;
|
---|
182 | }
|
---|
183 |
|
---|
184 | /**
|
---|
185 | * Checks URL parameter against each line ("filter") of conf/url-discard-filter.txt to decide
|
---|
186 | * whether it is in the discard list.
|
---|
187 | * Filters don't represent actual regex, just ^ and $ as start and end terminators.
|
---|
188 | * By not having this method deal with actual regex for filters, this has the advantage that
|
---|
189 | * we don't have to remember to escape or double escape each filter to turn it into a regex.
|
---|
190 | */
|
---|
191 | //public boolean isInDiscardFilter(String url) {
|
---|
192 |
|
---|
193 | public boolean isBlacklisted(String url) {
|
---|
194 | String discardFilterFile = "url-discard-filter.txt"; // in conf folder
|
---|
195 |
|
---|
196 | try (
|
---|
197 | BufferedReader reader = new BufferedReader(new InputStreamReader(org.greenstone.atea.WETProcessor.class.getClassLoader().getResourceAsStream(discardFilterFile), "UTF-8"));
|
---|
198 | ) {
|
---|
199 | String filter = null;
|
---|
200 | while((filter = reader.readLine()) != null) {
|
---|
201 | if(filter.trim().equals("")) {
|
---|
202 | continue;
|
---|
203 | }
|
---|
204 | //System.err.println("Got filter: " + filter);
|
---|
205 | if(filter.startsWith("^") && filter.endsWith("$") && url.equals(filter.substring(1, filter.length()-1))) {
|
---|
206 | System.err.println("*** Discarding url " + url + "\n\tas it MATCHES filter " + filter);
|
---|
207 | }
|
---|
208 | else if(filter.startsWith("^") && url.startsWith(filter.substring(1))) {
|
---|
209 | System.err.println("*** Discarding url " + url + "\n\tas it STARTS WITH filter " + filter);
|
---|
210 | return true;
|
---|
211 | }
|
---|
212 | else if(filter.endsWith("$") && url.endsWith(filter.substring(0, filter.length()-1))) {
|
---|
213 | System.err.println("*** Discarding url " + url + "\n\tas it ENDS WITH filter " + filter);
|
---|
214 | return true;
|
---|
215 | }
|
---|
216 | else if(url.contains(filter)) {
|
---|
217 | System.err.println("*** Discarding url " + url + "\n\tas it CONTAINS filter " + filter);
|
---|
218 | return true;
|
---|
219 | }
|
---|
220 |
|
---|
221 | }
|
---|
222 |
|
---|
223 | } catch (IOException ioe) {
|
---|
224 | ioe.printStackTrace();
|
---|
225 | System.err.println("\n@@@@@@@@@ Error reading from " + discardFilterFile);
|
---|
226 | }
|
---|
227 |
|
---|
228 | return false;
|
---|
229 | }
|
---|
230 |
|
---|
231 | /** Maintain a count of all WET files processed. */
|
---|
232 | public void setWETFileCount(int count) { this.wetFileCount = count; }
|
---|
233 |
|
---|
234 | /** Maintain a count of all WET records processed. */
|
---|
235 | //public int getRecordCount() { return this.totalRecordCount; }
|
---|
236 | //public void addToRecordCount(int count) { this.totalRecordCount += count; }
|
---|
237 | public void setRecordCount(int count) { this.totalRecordCount = count; }
|
---|
238 |
|
---|
239 | public static void printUsage() {
|
---|
240 | System.err.println("Run this program as:");
|
---|
241 | System.err.println("\tWetProcessor <folder containing wet(.gz) files> <output folder path>");
|
---|
242 | }
|
---|
243 |
|
---|
244 | /** Filename filter to only list warc.wet files or else warc.wet.gz files
|
---|
245 | * for which unzipped warc.wet equivalents don't yet exist.
|
---|
246 | */
|
---|
247 | private static class WETFilenameFilter implements FilenameFilter {
|
---|
248 |
|
---|
249 | public boolean accept(File dir, String name) {
|
---|
250 | if(name.endsWith(".warc.wet")) {
|
---|
251 | logger.debug("Will include " + name + " for processing.");
|
---|
252 | return true;
|
---|
253 | }
|
---|
254 |
|
---|
255 | if(name.endsWith(".warc.wet.gz")) {
|
---|
256 | String nameWithoutGZext = name.substring(0, name.lastIndexOf(".gz"));
|
---|
257 | File unzippedVersion = new File(dir, nameWithoutGZext);
|
---|
258 | if(unzippedVersion.exists()) {
|
---|
259 | logger.debug("--- Unzipped version " + unzippedVersion + " exists.");
|
---|
260 | logger.debug("Skipping " + name);
|
---|
261 | return false; // don't count gzipped version if unzipped version exists.
|
---|
262 | }
|
---|
263 | else {
|
---|
264 | logger.debug("Only zipped version " + name + " exists.");
|
---|
265 | return true; // No unzipped version, so have to work with gzipped version
|
---|
266 | }
|
---|
267 | }
|
---|
268 |
|
---|
269 | // we're not even interested in any other file extensions
|
---|
270 | logger.debug("Not a WET file. Skipping " + name);
|
---|
271 | return false;
|
---|
272 | }
|
---|
273 | }
|
---|
274 |
|
---|
275 |
|
---|
276 | public static void main(String[] args) {
|
---|
277 | if(args.length != 2) {
|
---|
278 | printUsage();
|
---|
279 | return;
|
---|
280 | }
|
---|
281 |
|
---|
282 |
|
---|
283 | File WETFileDir = new File(args[0]);
|
---|
284 | if(!WETFileDir.exists() || !WETFileDir.isDirectory()) {
|
---|
285 | System.out.println("Error: " + args[0] + " does not exist or is not a directory");
|
---|
286 | return;
|
---|
287 | }
|
---|
288 |
|
---|
289 | File outFolder = new File(args[1]);
|
---|
290 | if(!outFolder.exists() || !outFolder.isDirectory()) {
|
---|
291 | System.out.println("Error: " + args[1] + " does not exist or is not a directory.");
|
---|
292 | return;
|
---|
293 | }
|
---|
294 |
|
---|
295 | CCWETProcessor ccWETFilesProcessor = new CCWETProcessor(WETFileDir, outFolder);
|
---|
296 |
|
---|
297 | //ccWETFilesProcessor.processAllWETFiles();
|
---|
298 |
|
---|
299 | // Will list all the warc.wet files in the input directory or else their gzipped versions
|
---|
300 | File[] WETFiles = WETFileDir.listFiles(new WETFilenameFilter());
|
---|
301 |
|
---|
302 | int wetRecordCount = 0;
|
---|
303 | int wetFileCount = 0;
|
---|
304 |
|
---|
305 | for(int i = 0; i < WETFiles.length; i++) {
|
---|
306 | File WETFile = WETFiles[i];
|
---|
307 | logger.debug("Processing WETfile: " + WETFile);
|
---|
308 |
|
---|
309 | // Any .gz files listed means they haven't been unzipped yet. So unzip.
|
---|
310 | String WETFilename = WETFile.toString();
|
---|
311 | if(WETFilename.endsWith(".gz")) {
|
---|
312 | File GZippedWETFile = WETFile;
|
---|
313 | String WETGZippedFilename = WETFilename;
|
---|
314 | WETFilename = WETFilename.substring(0, WETFilename.lastIndexOf(".gz"));
|
---|
315 |
|
---|
316 | WETFile = new File(WETFilename);
|
---|
317 | Utility.unzipFile(GZippedWETFile, WETFile);
|
---|
318 | }
|
---|
319 | // hereafter all WETFiles should refer to the unzipped version
|
---|
320 | // Check the unzipped WETFile exists
|
---|
321 |
|
---|
322 | if(!WETFile.exists() || !WETFile.isFile()) {
|
---|
323 | System.err.println("Error: " + WETFile + " does not exist (failure to unzip?)");
|
---|
324 | logger.error("Error: " + WETFile + " does not exist (failure to unzip?)");
|
---|
325 | return;
|
---|
326 | }
|
---|
327 |
|
---|
328 | // Finally, we can process this WETFile's records into the keep and discard pile
|
---|
329 | wetFileCount++;
|
---|
330 | logger.debug("Off to process " + WETFile);
|
---|
331 | WETProcessor wetFileProcessor = new WETProcessor(WETFile, ccWETFilesProcessor);
|
---|
332 | wetFileProcessor.processWETFile();
|
---|
333 | wetRecordCount += wetFileProcessor.getRecordCount();
|
---|
334 | }
|
---|
335 |
|
---|
336 | // for information purposes
|
---|
337 | ccWETFilesProcessor.setWETFileCount(wetFileCount);
|
---|
338 | ccWETFilesProcessor.setRecordCount(wetRecordCount);
|
---|
339 |
|
---|
340 | File seedURLsFile = new File(outFolder, "seedURLs.txt");
|
---|
341 | File urlFilterFile = new File(outFolder, "regex-urlfilter.txt");
|
---|
342 | ccWETFilesProcessor.createSeedURLsFiles(seedURLsFile, urlFilterFile);
|
---|
343 |
|
---|
344 | return;
|
---|
345 |
|
---|
346 | }
|
---|
347 | }
|
---|