📄 surtprefixeddeciderule.html
字号:
<a name="142" href="#142">142</a> .booleanValue()) {<a name="143" href="#143">143</a> <strong>if</strong>(evaluate(((CandidateURI)object).getVia())) {<a name="144" href="#144">144</a> <strong>return</strong> <strong>true</strong>;<a name="145" href="#145">145</a> }<a name="146" href="#146">146</a> }<a name="147" href="#147">147</a> String candidateSurt;<a name="148" href="#148">148</a> candidateSurt = SurtPrefixSet.getCandidateSurt(object);<a name="149" href="#149">149</a> <strong>if</strong> (candidateSurt == <strong>null</strong>) {<a name="150" href="#150">150</a> <strong>return</strong> false;<a name="151" href="#151">151</a> }<a name="152" href="#152">152</a> <strong>return</strong> getPrefixes().containsPrefixOf(candidateSurt);<a name="153" href="#153">153</a> }<a name="154" href="#154">154</a> <a name="155" href="#155">155</a> <em>/**<em>*</em></em><a name="156" href="#156">156</a> <em> * Synchronized get of prefix set to use</em><a name="157" href="#157">157</a> <em> * </em><a name="158" href="#158">158</a> <em> * @return SurtPrefixSet to use for check</em><a name="159" href="#159">159</a> <em> */</em><a name="160" href="#160">160</a> <strong>private</strong> <strong>synchronized</strong> <a href="../../../../org/archive/util/SurtPrefixSet.html">SurtPrefixSet</a> getPrefixes() {<a name="161" href="#161">161</a> <strong>if</strong> (surtPrefixes == <strong>null</strong>) {<a name="162" href="#162">162</a> readPrefixes();<a name="163" href="#163">163</a> }<a name="164" href="#164">164</a> <strong>return</strong> surtPrefixes;<a name="165" href="#165">165</a> }<a name="166" href="#166">166</a> <a name="167" href="#167">167</a> <strong>protected</strong> <strong>void</strong> readPrefixes() {<a name="168" href="#168">168</a> buildSurtPrefixSet();<a name="169" href="#169">169</a> dumpSurtPrefixSet();<a name="170" href="#170">170</a> }<a name="171" href="#171">171</a> <a name="172" href="#172">172</a> <em>/**<em>*</em></em><a name="173" href="#173">173</a> <em> * Dump the current prefixes in use to configured dump file (if any)</em><a name="174" href="#174">174</a> <em> */</em><a name="175" href="#175">175</a> <strong>protected</strong> <strong>void</strong> dumpSurtPrefixSet() {<a name="176" href="#176">176</a> <em class="comment">// dump surts to file, if appropriate</em><a name="177" href="#177">177</a> String dumpPath = (String)getUncheckedAttribute(<strong>null</strong>,<a name="178" href="#178">178</a> ATTR_SURTS_DUMP_FILE);<a name="179" href="#179">179</a> <strong>if</strong> (dumpPath.length() > 0) {<a name="180" href="#180">180</a> File dump = <strong>new</strong> File(dumpPath);<a name="181" href="#181">181</a> <strong>if</strong> (!dump.isAbsolute()) {<a name="182" href="#182">182</a> dump = <strong>new</strong> File(getSettingsHandler().getOrder().getController()<a name="183" href="#183">183</a> .getDisk(), dumpPath);<a name="184" href="#184">184</a> }<a name="185" href="#185">185</a> <strong>try</strong> {<a name="186" href="#186">186</a> FileWriter fw = <strong>new</strong> FileWriter(dump);<a name="187" href="#187">187</a> <strong>try</strong> {<a name="188" href="#188">188</a> surtPrefixes.exportTo(fw);<a name="189" href="#189">189</a> } <strong>finally</strong> {<a name="190" href="#190">190</a> fw.close();<a name="191" href="#191">191</a> }<a name="192" href="#192">192</a> } <strong>catch</strong> (IOException e) {<a name="193" href="#193">193</a> e.printStackTrace();<a name="194" href="#194">194</a> <strong>throw</strong> <strong>new</strong> RuntimeException(e);<a name="195" href="#195">195</a> }<a name="196" href="#196">196</a> }<a name="197" href="#197">197</a> }<a name="198" href="#198">198</a> <a name="199" href="#199">199</a> <em>/**<em>*</em></em><a name="200" href="#200">200</a> <em> * Construct the set of prefixes to use, from the seed list (</em><a name="201" href="#201">201</a> <em> * which may include both URIs and '+'-prefixed directives).</em><a name="202" href="#202">202</a> <em> */</em><a name="203" href="#203">203</a> <strong>protected</strong> <strong>void</strong> buildSurtPrefixSet() {<a name="204" href="#204">204</a> <a href="../../../../org/archive/util/SurtPrefixSet.html">SurtPrefixSet</a> <strong>new</strong>SurtPrefixes = <strong>new</strong> <a href="../../../../org/archive/util/SurtPrefixSet.html">SurtPrefixSet</a>();<a name="205" href="#205">205</a> FileReader fr = <strong>null</strong>;<a name="206" href="#206">206</a> <a name="207" href="#207">207</a> <em class="comment">// read SURTs from file, if appropriate</em><a name="208" href="#208">208</a> String sourcePath = (String)getUncheckedAttribute(<strong>null</strong>,<a name="209" href="#209">209</a> ATTR_SURTS_SOURCE_FILE);<a name="210" href="#210">210</a> <strong>if</strong> (sourcePath.length() > 0) {<a name="211" href="#211">211</a> File source = <strong>new</strong> File(sourcePath);<a name="212" href="#212">212</a> <strong>if</strong> (!source.isAbsolute()) {<a name="213" href="#213">213</a> source = <strong>new</strong> File(getSettingsHandler().getOrder()<a name="214" href="#214">214</a> .getController().getDisk(), sourcePath);<a name="215" href="#215">215</a> }<a name="216" href="#216">216</a> <strong>try</strong> {<a name="217" href="#217">217</a> fr = <strong>new</strong> FileReader(source);<a name="218" href="#218">218</a> <strong>try</strong> {<a name="219" href="#219">219</a> newSurtPrefixes.importFromMixed(fr, <strong>true</strong>);<a name="220" href="#220">220</a> } <strong>finally</strong> {<a name="221" href="#221">221</a> fr.close();<a name="222" href="#222">222</a> }<a name="223" href="#223">223</a> } <strong>catch</strong> (IOException e) {<a name="224" href="#224">224</a> e.printStackTrace();<a name="225" href="#225">225</a> <strong>throw</strong> <strong>new</strong> RuntimeException(e);<a name="226" href="#226">226</a> }<a name="227" href="#227">227</a> }<a name="228" href="#228">228</a> <a name="229" href="#229">229</a> <em class="comment">// interpret seeds as surts, if appropriate</em><a name="230" href="#230">230</a> <strong>boolean</strong> deduceFromSeeds = ((Boolean)getUncheckedAttribute(<strong>null</strong>,<a name="231" href="#231">231</a> ATTR_SEEDS_AS_SURT_PREFIXES)).booleanValue();<a name="232" href="#232">232</a> <strong>if</strong>(deduceFromSeeds) {<a name="233" href="#233">233</a> <strong>try</strong> {<a name="234" href="#234">234</a> fr = <strong>new</strong> FileReader(getSeedfile());<a name="235" href="#235">235</a> <strong>try</strong> {<a name="236" href="#236">236</a> newSurtPrefixes.importFromMixed(fr, deduceFromSeeds);<a name="237" href="#237">237</a> } <strong>finally</strong> {<a name="238" href="#238">238</a> fr.close();<a name="239" href="#239">239</a> }<a name="240" href="#240">240</a> } <strong>catch</strong> (IOException e) {<a name="241" href="#241">241</a> e.printStackTrace();<a name="242" href="#242">242</a> <strong>throw</strong> <strong>new</strong> RuntimeException(e);<a name="243" href="#243">243</a> }<a name="244" href="#244">244</a> }<a name="245" href="#245">245</a> <a name="246" href="#246">246</a> surtPrefixes = newSurtPrefixes;<a name="247" href="#247">247</a> }<a name="248" href="#248">248</a> <a name="249" href="#249">249</a> <em>/**<em>*</em></em><a name="250" href="#250">250</a> <em> * Re-read prefixes after an update.</em><a name="251" href="#251">251</a> <em> * </em><a name="252" href="#252">252</a> <em> * @see org.archive.crawler.framework.CrawlScope#kickUpdate()</em><a name="253" href="#253">253</a> <em> */</em><a name="254" href="#254">254</a> <strong>public</strong> <strong>synchronized</strong> <strong>void</strong> kickUpdate() {<a name="255" href="#255">255</a> <strong>super</strong>.kickUpdate();<a name="256" href="#256">256</a> <strong>if</strong> (((Boolean) getUncheckedAttribute(<strong>null</strong>, ATTR_REBUILD_ON_RECONFIG))<a name="257" href="#257">257</a> .booleanValue()) {<a name="258" href="#258">258</a> readPrefixes();<a name="259" href="#259">259</a> }<a name="260" href="#260">260</a> <em class="comment">// TODO: make conditional on file having actually changed,</em><a name="261" href="#261">261</a> <em class="comment">// perhaps by remembering mod-time</em><a name="262" href="#262">262</a> }<a name="263" href="#263">263</a> <a name="264" href="#264">264</a> <em>/**<em>*</em></em><a name="265" href="#265">265</a> <em> * Dig through everything to get the crawl-global seeds file. </em><a name="266" href="#266">266</a> <em> * Add self as listener while at it. </em><a name="267" href="#267">267</a> <em> * </em><a name="268" href="#268">268</a> <em> * @return Seed list file</em><a name="269" href="#269">269</a> <em> */</em><a name="270" href="#270">270</a> <strong>protected</strong> File getSeedfile() {<a name="271" href="#271">271</a> <a href="../../../../org/archive/crawler/framework/CrawlScope.html">CrawlScope</a> scope =<a name="272" href="#272">272</a> getSettingsHandler().getOrder().getController().getScope();<a name="273" href="#273">273</a> scope.addSeedListener(<strong>this</strong>);<a name="274" href="#274">274</a> <strong>return</strong> scope.getSeedfile();<a name="275" href="#275">275</a> }<a name="276" href="#276">276</a> <a name="277" href="#277">277</a> <strong>public</strong> <strong>synchronized</strong> <strong>void</strong> addedSeed(<strong>final</strong> <a href="../../../../org/archive/crawler/datamodel/CandidateURI.html">CandidateURI</a> curi) {<a name="278" href="#278">278</a> <a href="../../../../org/archive/util/SurtPrefixSet.html">SurtPrefixSet</a> newSurtPrefixes = (SurtPrefixSet) surtPrefixes.clone();<a name="279" href="#279">279</a> newSurtPrefixes.add(prefixFrom(curi.toString()));<a name="280" href="#280">280</a> surtPrefixes = newSurtPrefixes;<a name="281" href="#281">281</a> }<a name="282" href="#282">282</a> <a name="283" href="#283">283</a> <strong>protected</strong> String prefixFrom(String uri) {<a name="284" href="#284">284</a> <strong>return</strong> SurtPrefixSet.prefixFromPlain(uri);<a name="285" href="#285">285</a> }<a name="286" href="#286">286</a> }</pre><hr/><div id="footer">This page was automatically generated by <a href="http://maven.apache.org/">Maven</a></div></body></html>
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -