experimentalwarcwriterprocessor.html

来自「网络爬虫开源代码」· HTML 代码 · 共 495 行 · 第 1/3 页

<a name="157" href="#157">157</a> <em>     * @param curi CrawlURI to process.</em>
<a name="158" href="#158">158</a> <em>     * </em>
<a name="159" href="#159">159</a> <em>     */</em>
<a name="160" href="#160">160</a>     <strong>protected</strong> <strong>void</strong> innerProcess(<a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi) {
<a name="161" href="#161">161</a>         <em class="comment">// If failure, or we haven't fetched the resource yet, return</em>
<a name="162" href="#162">162</a>         <strong>if</strong> (curi.getFetchStatus() &lt;= 0) {
<a name="163" href="#163">163</a>             <strong>return</strong>;
<a name="164" href="#164">164</a>         }
<a name="165" href="#165">165</a>         
<a name="166" href="#166">166</a>         <em class="comment">// If no recorded content at all, don't write record.</em>
<a name="167" href="#167">167</a>         <strong>long</strong> recordLength = curi.getContentSize();
<a name="168" href="#168">168</a>         <strong>if</strong> (recordLength &lt;= 0) {
<a name="169" href="#169">169</a>             <em class="comment">// getContentSize() should be > 0 if any material (even just</em>
<a name="170" href="#170">170</a>             <em class="comment">// HTTP headers with zero-length body) is available. </em>
<a name="171" href="#171">171</a>         	<strong>return</strong>;
<a name="172" href="#172">172</a>         }
<a name="173" href="#173">173</a>         
<a name="174" href="#174">174</a>         String scheme = curi.getUURI().getScheme().toLowerCase();
<a name="175" href="#175">175</a>         <strong>try</strong> {
<a name="176" href="#176">176</a>             <strong>if</strong> (shouldWrite(curi)) {
<a name="177" href="#177">177</a>                 write(scheme, curi);
<a name="178" href="#178">178</a>             } <strong>else</strong> {
<a name="179" href="#179">179</a>                 logger.info(<span class="string">"This writer does not write out scheme "</span> +
<a name="180" href="#180">180</a>                         scheme + <span class="string">" content"</span>);
<a name="181" href="#181">181</a>             }
<a name="182" href="#182">182</a>         } <strong>catch</strong> (IOException e) {
<a name="183" href="#183">183</a>             curi.addLocalizedError(<strong>this</strong>.getName(), e, <span class="string">"WriteRecord: "</span> +
<a name="184" href="#184">184</a>                 curi.toString());
<a name="185" href="#185">185</a>             logger.log(Level.SEVERE, <span class="string">"Failed write of Record: "</span> +
<a name="186" href="#186">186</a>                 curi.toString(), e);
<a name="187" href="#187">187</a>         }
<a name="188" href="#188">188</a>     }
<a name="189" href="#189">189</a>     
<a name="190" href="#190">190</a>     <strong>protected</strong> <strong>void</strong> write(<strong>final</strong> String lowerCaseScheme, <strong>final</strong> <a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi)
<a name="191" href="#191">191</a>     throws IOException {
<a name="192" href="#192">192</a>         <a href="../../../../org/archive/io/WriterPoolMember.html">WriterPoolMember</a> writer = getPool().borrowFile();
<a name="193" href="#193">193</a>         <strong>long</strong> position = writer.getPosition();
<a name="194" href="#194">194</a>         <em class="comment">// See if we need to open a new file because we've exceeed maxBytes.</em>
<a name="195" href="#195">195</a>         <em class="comment">// Call to checkFileSize will open new file if we're at maximum for</em>
<a name="196" href="#196">196</a>         <em class="comment">// current file.</em>
<a name="197" href="#197">197</a>         writer.checkSize();
<a name="198" href="#198">198</a>         <strong>if</strong> (writer.getPosition() != position) {
<a name="199" href="#199">199</a>             <em class="comment">// We just closed the file because it was larger than maxBytes.</em>
<a name="200" href="#200">200</a>             <em class="comment">// Add to the totalBytesWritten the size of the first record</em>
<a name="201" href="#201">201</a>             <em class="comment">// in the file, if any.</em>
<a name="202" href="#202">202</a>             setTotalBytesWritten(getTotalBytesWritten() +
<a name="203" href="#203">203</a>             	(writer.getPosition() - position));
<a name="204" href="#204">204</a>             position = writer.getPosition();
<a name="205" href="#205">205</a>         }
<a name="206" href="#206">206</a>         
<a name="207" href="#207">207</a>         <a href="../../../../org/archive/io/warc/ExperimentalWARCWriter.html">ExperimentalWARCWriter</a> w = (ExperimentalWARCWriter)writer;
<a name="208" href="#208">208</a>         <strong>try</strong> {
<a name="209" href="#209">209</a>             <em class="comment">// Write a request, response, and metadata all in the one</em>
<a name="210" href="#210">210</a>             <em class="comment">// 'transaction'.</em>
<a name="211" href="#211">211</a>             <strong>final</strong> URI baseid = getRecordID();
<a name="212" href="#212">212</a>             <strong>final</strong> String timestamp =
<a name="213" href="#213">213</a>                 ArchiveUtils.getLog14Date(curi.getLong(A_FETCH_BEGAN_TIME));
<a name="214" href="#214">214</a>             <strong>if</strong> (lowerCaseScheme.startsWith(<span class="string">"http"</span>)) {
<a name="215" href="#215">215</a>                 <em class="comment">// Add named fields for ip, checksum, and relate the metadata</em>
<a name="216" href="#216">216</a>                 <em class="comment">// and request to the resource field.</em>
<a name="217" href="#217">217</a>                 <em class="comment">// TODO: Use other than ANVL (or rename ANVL as NameValue or</em>
<a name="218" href="#218">218</a>                 <em class="comment">// use RFC822 (commons-httpclient?).</em>
<a name="219" href="#219">219</a>                 <a href="../../../../org/archive/util/anvl/ANVLRecord.html">ANVLRecord</a> headers = <strong>new</strong> <a href="../../../../org/archive/util/anvl/ANVLRecord.html">ANVLRecord</a>(5);
<a name="220" href="#220">220</a>                 <strong>if</strong> (curi.getContentDigest() != <strong>null</strong>) {
<a name="221" href="#221">221</a>                     headers.addLabelValue(HEADER_KEY_CHECKSUM,
<a name="222" href="#222">222</a>                         curi.getContentDigestSchemeString());
<a name="223" href="#223">223</a>                 }
<a name="224" href="#224">224</a>                 headers.addLabelValue(HEADER_KEY_IP, getHostAddress(curi));
<a name="225" href="#225">225</a>                 URI rid;
<a name="226" href="#226">226</a>                 
<a name="227" href="#227">227</a>                 <strong>if</strong> (IdenticalDigestDecideRule.hasIdenticalDigest(curi) &amp;&amp; 
<a name="228" href="#228">228</a>                         ((Boolean)getUncheckedAttribute(curi, 
<a name="229" href="#229">229</a>                                 ATTR_WRITE_REVISIT_FOR_IDENTICAL_DIGESTS))) {
<a name="230" href="#230">230</a>                     rid = writeRevisitDigest(w, timestamp, HTTP_RESPONSE_MIMETYPE,
<a name="231" href="#231">231</a>                             baseid, curi, headers);
<a name="232" href="#232">232</a>                 } <strong>else</strong> <strong>if</strong> (curi.getFetchStatus() == HttpStatus.SC_NOT_MODIFIED &amp;&amp; 
<a name="233" href="#233">233</a>                         ((Boolean)getUncheckedAttribute(curi, 
<a name="234" href="#234">234</a>                                 ATTR_WRITE_REVISIT_FOR_NOT_MODIFIED))) {
<a name="235" href="#235">235</a>                     rid = writeRevisitNotModified(w, timestamp,
<a name="236" href="#236">236</a>                             baseid, curi, headers);
<a name="237" href="#237">237</a>                 } <strong>else</strong> {
<a name="238" href="#238">238</a>                     <strong>if</strong> (curi.isTruncatedFetch()) {
<a name="239" href="#239">239</a>                         String value = curi.isTimeTruncatedFetch()?
<a name="240" href="#240">240</a>                             NAMED_FIELD_TRUNCATED_VALUE_TIME:
<a name="241" href="#241">241</a>                             curi.isLengthTruncatedFetch()?
<a name="242" href="#242">242</a>                                 NAMED_FIELD_TRUNCATED_VALUE_LEN:
<a name="243" href="#243">243</a>                                 curi.isHeaderTruncatedFetch()?
<a name="244" href="#244">244</a>                                     NAMED_FIELD_TRUNCATED_VALUE_HEAD:
<a name="245" href="#245">245</a>                             <em class="comment">// TODO: Add this to spec.</em>
<a name="246" href="#246">246</a>                             TRUNCATED_VALUE_UNSPECIFIED;
<a name="247" href="#247">247</a>                         headers.addLabelValue(HEADER_KEY_TRUNCATED, value);
<a name="248" href="#248">248</a>                     }
<a name="249" href="#249">249</a>                     rid = writeResponse(w, timestamp, HTTP_RESPONSE_MIMETYPE,
<a name="250" href="#250">250</a>                     	baseid, curi, headers);
<a name="251" href="#251">251</a>                 }
<a name="252" href="#252">252</a>                 
<a name="253" href="#253">253</a>                 headers = <strong>new</strong> <a href="../../../../org/archive/util/anvl/ANVLRecord.html">ANVLRecord</a>(1);
<a name="254" href="#254">254</a>                 headers.addLabelValue(HEADER_KEY_CONCURRENT_TO,
<a name="255" href="#255">255</a>                     '&lt;' + rid.toString() + '>');
<a name="256" href="#256">256</a> 
<a name="257" href="#257">257</a>                 <strong>if</strong>(((Boolean)getUncheckedAttribute(curi, ATTR_WRITE_REQUESTS))) {
<a name="258" href="#258">258</a>                     writeRequest(w, timestamp, HTTP_REQUEST_MIMETYPE,
<a name="259" href="#259">259</a>                             baseid, curi, headers);
<a name="260" href="#260">260</a>                 }
<a name="261" href="#261">261</a>                 <strong>if</strong>(((Boolean)getUncheckedAttribute(curi, ATTR_WRITE_METADATA))) {
<a name="262" href="#262">262</a>                     writeMetadata(w, timestamp, baseid, curi, headers);
<a name="263" href="#263">263</a>                 } 
<a name="264" href="#264">264</a>             } <strong>else</strong> <strong>if</strong> (lowerCaseScheme.equals(<span class="string">"dns"</span>)) {
<a name="265" href="#265">265</a>                 <a href="../../../../org/archive/util/anvl/ANVLRecord.html">ANVLRecord</a> headers = <strong>null</strong>;
<a name="266" href="#266">266</a>                 String ip = curi.getString(A_DNS_SERVER_IP_LABEL);
<a name="267" href="#267">267</a>                 <strong>if</strong> (ip != <strong>null</strong> &amp;&amp; ip.length() > 0) {
<a name="268" href="#268">268</a>                     headers = <strong>new</strong> <a href="../../../../org/archive/util/anvl/ANVLRecord.html">ANVLRecord</a>(1);
<a name="269" href="#269">269</a>                     headers.addLabelValue(HEADER_KEY_IP, ip);
<a name="270" href="#270">270</a>                 }
<a name="271" href="#271">271</a>                 writeResponse(w, timestamp, curi.getContentType(), baseid,
<a name="272" href="#272">272</a>                     curi, headers);
<a name="273" href="#273">273</a>             } <strong>else</strong> {
<a name="274" href="#274">274</a>                 logger.warning(<span class="string">"No handler for scheme "</span> + lowerCaseScheme);
<a name="275" href="#275">275</a>             }
<a name="276" href="#276">276</a>         } <strong>catch</strong> (IOException e) {
<a name="277" href="#277">277</a>             <em class="comment">// Invalidate this file (It gets a '.invalid' suffix).</em>
<a name="278" href="#278">278</a>             getPool().invalidateFile(writer);
<a name="279" href="#279">279</a>             <em class="comment">// Set the writer to null otherwise the pool accounting</em>
<a name="280" href="#280">280</a>             <em class="comment">// of how many active writers gets skewed if we subsequently</em>
<a name="281" href="#281">281</a>             <em class="comment">// do a returnWriter call on this object in the finally block.</em>
<a name="282" href="#282">282</a>             writer = <strong>null</strong>;
<a name="283" href="#283">283</a>             <strong>throw</strong> e;
<a name="284" href="#284">284</a>         } <strong>finally</strong> {
<a name="285" href="#285">285</a>             <strong>if</strong> (writer != <strong>null</strong>) {
<a name="286" href="#286">286</a>             	setTotalBytesWritten(getTotalBytesWritten() +
<a name="287" href="#287">287</a>             	     (writer.getPosition() - position));
<a name="288" href="#288">288</a>                 getPool().returnFile(writer);
<a name="289" href="#289">289</a>             }
<a name="290" href="#290">290</a>         }
<a name="291" href="#291">291</a>         checkBytesWritten();
<a name="292" href="#292">292</a>     }
<a name="293" href="#293">293</a>     
<a name="294" href="#294">294</a>     <strong>protected</strong> URI writeRequest(<strong>final</strong> <a href="../../../../org/archive/io/warc/ExperimentalWARCWriter.html">ExperimentalWARCWriter</a> w,
<a name="295" href="#295">295</a>             <strong>final</strong> String timestamp, <strong>final</strong> String mimetype,
<a name="296" href="#296">296</a>             <strong>final</strong> URI baseid, <strong>final</strong> <a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi,
<a name="297" href="#297">297</a>             <strong>final</strong> <a href="../../../../org/archive/util/anvl/ANVLRecord.html">ANVLRecord</a> namedFields) 
<a name="298" href="#298">298</a>     throws IOException {
<a name="299" href="#299">299</a>         <strong>final</strong> URI uid = qualifyRecordID(baseid, TYPE, REQUEST);
<a name="300" href="#300">300</a>         <a href="../../../../org/archive/io/ReplayInputStream.html">ReplayInputStream</a> ris =
<a name="301" href="#301">301</a>             curi.getHttpRecorder().getRecordedOutput().getReplayInputStream();
<a name="302" href="#302">302</a>         <strong>try</strong> {
<a name="303" href="#303">303</a>             w.writeRequestRecord(curi.toString(), timestamp, mimetype, uid,
<a name="304" href="#304">304</a>                 namedFields, ris,
<a name="305" href="#305">305</a>                 curi.getHttpRecorder().getRecordedOutput().getSize());
<a name="306" href="#306">306</a>         } <strong>finally</strong> {
<a name="307" href="#307">307</a>             <strong>if</strong> (ris != <strong>null</strong>) {
<a name="308" href="#308">308</a>                 ris.close();
<a name="309" href="#309">309</a>             }
<a name="310" href="#310">310</a>         }
<a name="311" href="#311">311</a>         <strong>return</strong> uid;
<a name="312" href="#312">312</a>     }
<a name="313" href="#313">313</a>     
<a name="314" href="#314">314</a>     <strong>protected</strong> URI writeResponse(<strong>final</strong> <a href="../../../../org/archive/io/warc/ExperimentalWARCWriter.html">ExperimentalWARCWriter</a> w,
<a name="315" href="#315">315</a>             <strong>final</strong> String timestamp, <strong>final</strong> String mimetype,
<a name="316" href="#316">316</a>             <strong>final</strong> URI baseid, <strong>final</strong> <a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi,
<a name="317" href="#317">317</a>             <strong>final</strong> <a href="../../../../org/archive/util/anvl/ANVLRecord.html">ANVLRecord</a> namedFields) 
<a name="318" href="#318">318</a>     throws IOException {
<a name="319" href="#319">319</a>         <a href="../../../../org/archive/io/ReplayInputStream.html">ReplayInputStream</a> ris =
<a name="320" href="#320">320</a>             curi.getHttpRecorder().getRecordedInput().getReplayInputStream();
<a name="321" href="#321">321</a>         <strong>try</strong> {