crawljobhandler.html

来自「网络爬虫开源代码」· HTML 代码 · 共 744 行 · 第 1/5 页

<a name="141" href="#141">141</a> <em>	 * Thread to start the next job in background</em>
<a name="142" href="#142">142</a> <em>     */</em>    
<a name="143" href="#143">143</a>     <strong>private</strong> Thread startingNextJob = <strong>null</strong>;
<a name="144" href="#144">144</a> 
<a name="145" href="#145">145</a>     <em>/**<em>*</em></em>
<a name="146" href="#146">146</a> <em>     * A list of pending CrawlJobs.</em>
<a name="147" href="#147">147</a> <em>     */</em>
<a name="148" href="#148">148</a>     <strong>private</strong> TreeSet&lt;CrawlJob> pendingCrawlJobs;
<a name="149" href="#149">149</a> 
<a name="150" href="#150">150</a>     <em>/**<em>*</em></em>
<a name="151" href="#151">151</a> <em>     * A list of completed CrawlJobs.</em>
<a name="152" href="#152">152</a> <em>     */</em>
<a name="153" href="#153">153</a>     <em class="comment">//private Vector completedCrawlJobs = new Vector();</em>
<a name="154" href="#154">154</a>     <strong>private</strong> TreeSet&lt;CrawlJob> completedCrawlJobs;
<a name="155" href="#155">155</a> 
<a name="156" href="#156">156</a>     <em>/**<em>*</em></em>
<a name="157" href="#157">157</a> <em>     * A list of profile CrawlJobs.</em>
<a name="158" href="#158">158</a> <em>     */</em>
<a name="159" href="#159">159</a>     <strong>private</strong> TreeSet&lt;CrawlJob> profileJobs;
<a name="160" href="#160">160</a>     
<a name="161" href="#161">161</a>     <em class="comment">// The UIDs of profiles should be NOT be timestamps. A descriptive name is</em>
<a name="162" href="#162">162</a>     <em class="comment">// ideal.</em>
<a name="163" href="#163">163</a>     <strong>private</strong> String defaultProfile = <strong>null</strong>;
<a name="164" href="#164">164</a> 
<a name="165" href="#165">165</a>     <em>/**<em>*</em></em>
<a name="166" href="#166">166</a> <em>     * If true the crawler is 'running'. That is the next pending job will start</em>
<a name="167" href="#167">167</a> <em>     * crawling as soon as the current job (if any) is completed.</em>
<a name="168" href="#168">168</a> <em>     */</em>
<a name="169" href="#169">169</a>     <strong>private</strong> <strong>boolean</strong> running = false;
<a name="170" href="#170">170</a>     
<a name="171" href="#171">171</a>     <em>/**<em>*</em></em>
<a name="172" href="#172">172</a> <em>     * String to indicate recovery should be based on the recovery log, not</em>
<a name="173" href="#173">173</a> <em>     * based on checkpointing.</em>
<a name="174" href="#174">174</a> <em>     */</em>
<a name="175" href="#175">175</a>     <strong>public</strong> <strong>static</strong> <strong>final</strong> String RECOVER_LOG = <span class="string">"recover"</span>;
<a name="176" href="#176">176</a>     
<a name="177" href="#177">177</a>     <em>/**<em>*</em></em>
<a name="178" href="#178">178</a> <em>     * Jobs directory.</em>
<a name="179" href="#179">179</a> <em>     */</em>
<a name="180" href="#180">180</a>     <strong>private</strong> <strong>final</strong> File jobsDir;
<a name="181" href="#181">181</a>     
<a name="182" href="#182">182</a>     <em>/**<em>*</em></em>
<a name="183" href="#183">183</a> <em>     * Constructor.</em>
<a name="184" href="#184">184</a> <em>     * @param jobsDir Jobs directory.</em>
<a name="185" href="#185">185</a> <em>     */</em>
<a name="186" href="#186">186</a>     <strong>public</strong> <a href="../../../../org/archive/crawler/admin/CrawlJobHandler.html">CrawlJobHandler</a>(<strong>final</strong> File jobsDir) {
<a name="187" href="#187">187</a>         <strong>this</strong>(jobsDir, <strong>true</strong>, <strong>true</strong>);
<a name="188" href="#188">188</a>     }
<a name="189" href="#189">189</a> 
<a name="190" href="#190">190</a>     <em>/**<em>*</em></em>
<a name="191" href="#191">191</a> <em>     * Constructor allowing for optional loading of profiles and jobs.</em>
<a name="192" href="#192">192</a> <em>     * @param jobsDir Jobs directory.</em>
<a name="193" href="#193">193</a> <em>     * @param loadJobs If true then any applicable jobs will be loaded.</em>
<a name="194" href="#194">194</a> <em>     * @param loadProfiles If true then any applicable profiles will be loaded.</em>
<a name="195" href="#195">195</a> <em>     */</em>
<a name="196" href="#196">196</a>     <strong>public</strong> <a href="../../../../org/archive/crawler/admin/CrawlJobHandler.html">CrawlJobHandler</a>(<strong>final</strong> File jobsDir,
<a name="197" href="#197">197</a>             <strong>final</strong> <strong>boolean</strong> loadJobs, <strong>final</strong> <strong>boolean</strong> loadProfiles) {
<a name="198" href="#198">198</a>         <strong>this</strong>.jobsDir = jobsDir;
<a name="199" href="#199">199</a>         <em class="comment">// Make a comparator for CrawlJobs.</em>
<a name="200" href="#200">200</a>         Comparator&lt;CrawlJob> comp = <strong>new</strong> Comparator&lt;CrawlJob>(){
<a name="201" href="#201">201</a>             <strong>public</strong> <strong>int</strong> compare(<a href="../../../../org/archive/crawler/admin/CrawlJob.html">CrawlJob</a> job1, <a href="../../../../org/archive/crawler/admin/CrawlJob.html">CrawlJob</a> job2) {
<a name="202" href="#202">202</a>                 <strong>if</strong>( job1.getJobPriority() &lt; job2.getJobPriority() ){
<a name="203" href="#203">203</a>                     <strong>return</strong> -1;
<a name="204" href="#204">204</a>                 } <strong>else</strong> <strong>if</strong>( job1.getJobPriority() > job2.getJobPriority() ){
<a name="205" href="#205">205</a>                     <strong>return</strong> 1;
<a name="206" href="#206">206</a>                 } <strong>else</strong> {
<a name="207" href="#207">207</a>                     <em class="comment">// Same priority, use UID (which should be a timestamp).</em>
<a name="208" href="#208">208</a>                     <em class="comment">// Lower UID (string compare) means earlier time.</em>
<a name="209" href="#209">209</a>                     <strong>return</strong> job1.getUID().compareTo(job2.getUID());
<a name="210" href="#210">210</a>                 }
<a name="211" href="#211">211</a>             }
<a name="212" href="#212">212</a>         };
<a name="213" href="#213">213</a>         <strong>this</strong>.pendingCrawlJobs = <strong>new</strong> TreeSet&lt;CrawlJob>(comp);
<a name="214" href="#214">214</a>         <strong>this</strong>.completedCrawlJobs = <strong>new</strong> TreeSet&lt;CrawlJob>(comp);
<a name="215" href="#215">215</a>         <em class="comment">// Profiles always have the same priority so it will be sorted by name</em>
<a name="216" href="#216">216</a>         <strong>this</strong>.profileJobs = <strong>new</strong> TreeSet&lt;CrawlJob>(comp);
<a name="217" href="#217">217</a>         <strong>if</strong> (loadProfiles){
<a name="218" href="#218">218</a>             loadProfiles();
<a name="219" href="#219">219</a>         }
<a name="220" href="#220">220</a>         <strong>if</strong> (loadJobs){
<a name="221" href="#221">221</a>             loadJobs();
<a name="222" href="#222">222</a>         }
<a name="223" href="#223">223</a>     }
<a name="224" href="#224">224</a>     
<a name="225" href="#225">225</a>     <em>/**<em>*</em></em>
<a name="226" href="#226">226</a> <em>     * Find the state.job file in the job directory.</em>
<a name="227" href="#227">227</a> <em>     * @param jobDir Directory to look in.</em>
<a name="228" href="#228">228</a> <em>     * @return Full path to 'state.job' file or null if none found.</em>
<a name="229" href="#229">229</a> <em>     */</em>
<a name="230" href="#230">230</a>     <strong>protected</strong> File getStateJobFile(<strong>final</strong> File jobDir) {
<a name="231" href="#231">231</a>         <em class="comment">// Need to find job file ('state.job').</em>
<a name="232" href="#232">232</a>         File[] jobFiles = jobDir.listFiles(<strong>new</strong> FilenameFilter() {
<a name="233" href="#233">233</a>             <strong>public</strong> <strong>boolean</strong> accept(File dir, String name) {
<a name="234" href="#234">234</a>                 <strong>return</strong> name.toLowerCase().endsWith(<span class="string">".job"</span>) &amp;&amp;
<a name="235" href="#235">235</a>                     (<strong>new</strong> File(dir, name)).canRead();
<a name="236" href="#236">236</a>             }
<a name="237" href="#237">237</a>             
<a name="238" href="#238">238</a>         });
<a name="239" href="#239">239</a>         <strong>return</strong> (jobFiles.length == 1)? jobFiles[0]: <strong>null</strong>;
<a name="240" href="#240">240</a>     }
<a name="241" href="#241">241</a> 
<a name="242" href="#242">242</a>     <em>/**<em>*</em></em>
<a name="243" href="#243">243</a> <em>     * Loads any availible jobs in the jobs directory.</em>
<a name="244" href="#244">244</a> <em>     * &lt;p></em>
<a name="245" href="#245">245</a> <em>     * Availible jobs are any directory containing a file called</em>
<a name="246" href="#246">246</a> <em>     * &lt;code>state.job&lt;/code>. The file must contain valid job information.</em>
<a name="247" href="#247">247</a> <em>     */</em>
<a name="248" href="#248">248</a>     <strong>private</strong> <strong>void</strong> loadJobs() {
<a name="249" href="#249">249</a>         <strong>this</strong>.jobsDir.mkdirs();
<a name="250" href="#250">250</a>         File[] jobs = <strong>this</strong>.jobsDir.listFiles();
<a name="251" href="#251">251</a>         <strong>for</strong> (<strong>int</strong> i = 0; i &lt; jobs.length; i++) {
<a name="252" href="#252">252</a>             <strong>if</strong> (jobs[i].isDirectory()) {
<a name="253" href="#253">253</a>                 File jobFile = getStateJobFile(jobs[i]);
<a name="254" href="#254">254</a>                 <strong>if</strong> (jobFile != <strong>null</strong>) {
<a name="255" href="#255">255</a>                     loadJob(jobFile);
<a name="256" href="#256">256</a>                 }
<a name="257" href="#257">257</a>             }
<a name="258" href="#258">258</a>         }
<a name="259" href="#259">259</a>     }
<a name="260" href="#260">260</a> 
<a name="261" href="#261">261</a>     <em>/**<em>*</em></em>
<a name="262" href="#262">262</a> <em>     * Loads a job given a specific job file. The loaded job will be placed in</em>
<a name="263" href="#263">263</a> <em>     * the list of completed jobs or pending queue depending on its status.</em>
<a name="264" href="#264">264</a> <em>     * Running jobs will have their status set to 'finished abnormally' and put</em>
<a name="265" href="#265">265</a> <em>     * into the completed list.</em>
<a name="266" href="#266">266</a> <em>     * @param job The job file of the job to load.</em>
<a name="267" href="#267">267</a> <em>     */</em>
<a name="268" href="#268">268</a>     <strong>protected</strong> <strong>void</strong> loadJob(<strong>final</strong> File job) {
<a name="269" href="#269">269</a>         <a href="../../../../org/archive/crawler/admin/CrawlJob.html">CrawlJob</a> cjob = <strong>null</strong>;
<a name="270" href="#270">270</a>         <strong>try</strong> {
<a name="271" href="#271">271</a>             <em class="comment">// Load the CrawlJob</em>
<a name="272" href="#272">272</a>             cjob = <strong>new</strong> <a href="../../../../org/archive/crawler/admin/CrawlJob.html">CrawlJob</a>(job, <strong>new</strong> <a href="../../../../org/archive/crawler/admin/CrawlJobErrorHandler.html">CrawlJobErrorHandler</a>());
<a name="273" href="#273">273</a>         } <strong>catch</strong> (InvalidJobFileException e) {
<a name="274" href="#274">274</a>             logger.log(Level.INFO,
<a name="275" href="#275">275</a>                     <span class="string">"Invalid job file for "</span> + job.getAbsolutePath(), e);
<a name="276" href="#276">276</a>             <strong>return</strong>;
<a name="277" href="#277">277</a>         } <strong>catch</strong> (IOException e) {
<a name="278" href="#278">278</a>             logger.log(Level.INFO, <span class="string">"IOException for "</span> + job.getName() +
<a name="279" href="#279">279</a>                     <span class="string">", "</span> + job.getAbsolutePath(), e);
<a name="280" href="#280">280</a>             <strong>return</strong>;
<a name="281" href="#281">281</a>         }
<a name="282" href="#282">282</a>         
<a name="283" href="#283">283</a>         <em class="comment">// TODO: Move test into CrawlJob.</em>
<a name="284" href="#284">284</a>         <em class="comment">// Check job status and place it accordingly.</em>
<a name="285" href="#285">285</a>         <strong>if</strong> (cjob.getStatus().equals(CrawlJob.STATUS_RUNNING)
<a name="286" href="#286">286</a>                 || cjob.getStatus().equals(CrawlJob.STATUS_PAUSED)
<a name="287" href="#287">287</a>                 || cjob.getStatus().equals(CrawlJob.STATUS_CHECKPOINTING)
<a name="288" href="#288">288</a>                 || cjob.getStatus().equals(CrawlJob.STATUS_WAITING_FOR_PAUSE) ){
<a name="289" href="#289">289</a>             <em class="comment">// Was a running job.</em>