📄 quotaenforcer.html
字号:
<a name="200" href="#200">200</a> addElementToDefinition(<strong>new</strong> <a href="../../../../org/archive/crawler/settings/SimpleType.html">SimpleType</a>(ATTR_HOST_MAX_FETCH_SUCCESSES,<a name="201" href="#201">201</a> maxFetchSuccessesDesc.replaceAll(<span class="string">"$CATEGORY"</span>,<span class="string">"host"</span>),<a name="202" href="#202">202</a> DEFAULT_HOST_MAX_FETCH_SUCCESSES));<a name="203" href="#203">203</a> addElementToDefinition(<strong>new</strong> <a href="../../../../org/archive/crawler/settings/SimpleType.html">SimpleType</a>(ATTR_HOST_MAX_SUCCESS_KB,<a name="204" href="#204">204</a> maxSuccessKbDesc.replaceAll(<span class="string">"$CATEGORY"</span>,<span class="string">"host"</span>),<a name="205" href="#205">205</a> DEFAULT_HOST_MAX_SUCCESS_KB));<a name="206" href="#206">206</a> <em class="comment">// host all-responses</em><a name="207" href="#207">207</a> addElementToDefinition(<strong>new</strong> <a href="../../../../org/archive/crawler/settings/SimpleType.html">SimpleType</a>(ATTR_HOST_MAX_FETCH_RESPONSES,<a name="208" href="#208">208</a> maxFetchResponsesDesc.replaceAll(<span class="string">"$CATEGORY"</span>,<span class="string">"host"</span>),<a name="209" href="#209">209</a> DEFAULT_HOST_MAX_FETCH_RESPONSES));<a name="210" href="#210">210</a> addElementToDefinition(<strong>new</strong> <a href="../../../../org/archive/crawler/settings/SimpleType.html">SimpleType</a>(ATTR_HOST_MAX_ALL_KB,<a name="211" href="#211">211</a> maxAllKbDesc.replaceAll(<span class="string">"$CATEGORY"</span>,<span class="string">"host"</span>),<a name="212" href="#212">212</a> DEFAULT_HOST_MAX_ALL_KB)); <a name="213" href="#213">213</a> <em class="comment">// group successes</em><a name="214" href="#214">214</a> addElementToDefinition(<strong>new</strong> <a href="../../../../org/archive/crawler/settings/SimpleType.html">SimpleType</a>(ATTR_GROUP_MAX_FETCH_SUCCESSES,<a name="215" href="#215">215</a> maxFetchSuccessesDesc.replaceAll(<span class="string">"$CATEGORY"</span>,<span class="string">"group (queue)"</span>),<a name="216" href="#216">216</a> DEFAULT_GROUP_MAX_FETCH_SUCCESSES));<a name="217" href="#217">217</a> addElementToDefinition(<strong>new</strong> <a href="../../../../org/archive/crawler/settings/SimpleType.html">SimpleType</a>(ATTR_GROUP_MAX_SUCCESS_KB,<a name="218" href="#218">218</a> maxSuccessKbDesc.replaceAll(<span class="string">"$CATEGORY"</span>,<span class="string">"group (queue)"</span>),<a name="219" href="#219">219</a> DEFAULT_GROUP_MAX_SUCCESS_KB));<a name="220" href="#220">220</a> <em class="comment">// group all-responses</em><a name="221" href="#221">221</a> addElementToDefinition(<strong>new</strong> <a href="../../../../org/archive/crawler/settings/SimpleType.html">SimpleType</a>(ATTR_GROUP_MAX_FETCH_RESPONSES,<a name="222" href="#222">222</a> maxFetchResponsesDesc.replaceAll(<span class="string">"$CATEGORY"</span>,<span class="string">"group (queue)"</span>),<a name="223" href="#223">223</a> DEFAULT_GROUP_MAX_FETCH_RESPONSES));<a name="224" href="#224">224</a> addElementToDefinition(<strong>new</strong> <a href="../../../../org/archive/crawler/settings/SimpleType.html">SimpleType</a>(ATTR_GROUP_MAX_ALL_KB,<a name="225" href="#225">225</a> maxAllKbDesc.replaceAll(<span class="string">"$CATEGORY"</span>,<span class="string">"group (queue)"</span>),<a name="226" href="#226">226</a> DEFAULT_GROUP_MAX_ALL_KB)); <a name="227" href="#227">227</a> <a name="228" href="#228">228</a> }<a name="229" href="#229">229</a> <a name="230" href="#230">230</a> <strong>protected</strong> <strong>void</strong> innerProcess(<a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi) {<a name="231" href="#231">231</a> CrawlSubstats.HasCrawlSubstats[] haveStats = <a name="232" href="#232">232</a> <strong>new</strong> CrawlSubstats.HasCrawlSubstats[] {<a name="233" href="#233">233</a> getController().getServerCache().getServerFor(curi), <em class="comment">// server</em><a name="234" href="#234">234</a> getController().getServerCache().getHostFor(curi), <em class="comment">// host</em><a name="235" href="#235">235</a> getController().getFrontier().getGroup(curi) <em class="comment">// group</em><a name="236" href="#236">236</a> };<a name="237" href="#237">237</a> <a name="238" href="#238">238</a> <strong>for</strong>(<strong>int</strong> cat=SERVER;cat<=GROUP;cat++) {<a name="239" href="#239">239</a> <strong>if</strong> (checkQuotas(curi,haveStats[cat],cat)) {<a name="240" href="#240">240</a> <strong>return</strong>;<a name="241" href="#241">241</a> }<a name="242" href="#242">242</a> }<a name="243" href="#243">243</a> }<a name="244" href="#244">244</a> <a name="245" href="#245">245</a> <em>/**<em>*</em></em><a name="246" href="#246">246</a> <em> * Check all quotas for the given substats and category (server, host, or</em><a name="247" href="#247">247</a> <em> * group). </em><a name="248" href="#248">248</a> <em> * </em><a name="249" href="#249">249</a> <em> * @param curi CrawlURI to mark up with results</em><a name="250" href="#250">250</a> <em> * @param hasStats holds CrawlSubstats with actual values to test</em><a name="251" href="#251">251</a> <em> * @param CAT category index (SERVER, HOST, GROUP) to quota settings keys</em><a name="252" href="#252">252</a> <em> * @return true if quota precludes fetching of CrawlURI</em><a name="253" href="#253">253</a> <em> */</em><a name="254" href="#254">254</a> <strong>protected</strong> <strong>boolean</strong> checkQuotas(<strong>final</strong> <a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi,<a name="255" href="#255">255</a> <strong>final</strong> CrawlSubstats.HasCrawlSubstats hasStats,<a name="256" href="#256">256</a> <strong>final</strong> <strong>int</strong> CAT) {<a name="257" href="#257">257</a> <a href="../../../../org/archive/crawler/datamodel/CrawlSubstats.html">CrawlSubstats</a> substats = hasStats.getSubstats();<a name="258" href="#258">258</a> <strong>long</strong>[] actuals = <strong>new</strong> <strong>long</strong>[] {<a name="259" href="#259">259</a> -1, <em class="comment">// dummy</em><a name="260" href="#260">260</a> substats.getFetchSuccesses(),<a name="261" href="#261">261</a> substats.getSuccessBytes()/1024,<a name="262" href="#262">262</a> substats.getFetchResponses(),<a name="263" href="#263">263</a> substats.getTotalBytes()/1024,<a name="264" href="#264">264</a> };<a name="265" href="#265">265</a> <strong>for</strong>(<strong>int</strong> q=SUCCESSES; q<=RESPONSE_KB;q++) {<a name="266" href="#266">266</a> <strong>if</strong>(applyQuota(curi,keys[CAT][q],actuals[q])) {<a name="267" href="#267">267</a> <strong>return</strong> <strong>true</strong>; <a name="268" href="#268">268</a> }<a name="269" href="#269">269</a> }<a name="270" href="#270">270</a> <strong>return</strong> false; <a name="271" href="#271">271</a> }<a name="272" href="#272">272</a> <a name="273" href="#273">273</a> <em>/**<em>*</em></em><a name="274" href="#274">274</a> <em> * Apply the quota specified by the given key against the actual </em><a name="275" href="#275">275</a> <em> * value provided. If the quota and actual values rule out processing the </em><a name="276" href="#276">276</a> <em> * given CrawlURI, mark up the CrawlURI appropriately. </em><a name="277" href="#277">277</a> <em> * </em><a name="278" href="#278">278</a> <em> * @param curi CrawlURI whose processing is subject to a potential quota</em><a name="279" href="#279">279</a> <em> * limitation</em><a name="280" href="#280">280</a> <em> * @param quotaKey settings key to get applicable quota</em><a name="281" href="#281">281</a> <em> * @param actual current value to compare to quota </em><a name="282" href="#282">282</a> <em> * @return true is CrawlURI is blocked by a quota, false otherwise</em><a name="283" href="#283">283</a> <em> */</em><a name="284" href="#284">284</a> <strong>protected</strong> <strong>boolean</strong> applyQuota(<a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi, String quotaKey, <strong>long</strong> actual) {<a name="285" href="#285">285</a> <strong>long</strong> quota = ((Long)getUncheckedAttribute(curi, quotaKey)).<strong>long</strong>Value();<a name="286" href="#286">286</a> <strong>if</strong> (quota >= 0 && actual >= quota) {<a name="287" href="#287">287</a> curi.setFetchStatus(S_BLOCKED_BY_QUOTA);<a name="288" href="#288">288</a> curi.addAnnotation(<span class="string">"Q:"</span>+quotaKey);<a name="289" href="#289">289</a> curi.skipToProcessorChain(getController().getPostprocessorChain());<a name="290" href="#290">290</a> <strong>if</strong>((Boolean)getUncheckedAttribute(curi,ATTR_FORCE_RETIRE)) {<a name="291" href="#291">291</a> curi.putObject(CoreAttributeConstants.A_FORCE_RETIRE, (Boolean) <strong>true</strong>);<a name="292" href="#292">292</a> }<a name="293" href="#293">293</a> <strong>return</strong> <strong>true</strong>;<a name="294" href="#294">294</a> }<a name="295" href="#295">295</a> <strong>return</strong> false; <a name="296" href="#296">296</a> }<a name="297" href="#297">297</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 + -