?? bdbmultipleworkqueues.html
字號:
<a name="282" href="#282">282</a> OperationStatus status;<a name="283" href="#283">283</a> <strong>try</strong> {<a name="284" href="#284">284</a> cursor = <strong>this</strong>.pendingUrisDB.openCursor(<strong>null</strong>, <strong>null</strong>);<a name="285" href="#285">285</a> <em class="comment">// get cap; headKey at this point should always point to </em><a name="286" href="#286">286</a> <em class="comment">// a queue-beginning cap entry (zero-length value)</em><a name="287" href="#287">287</a> status = cursor.getSearchKey(headKey, result, <strong>null</strong>);<a name="288" href="#288">288</a> <strong>if</strong>(status!=OperationStatus.SUCCESS || result.getData().length > 0) {<a name="289" href="#289">289</a> <em class="comment">// cap missing</em><a name="290" href="#290">290</a> <strong>throw</strong> <strong>new</strong> DatabaseException(<span class="string">"bdb queue cap missing"</span>);<a name="291" href="#291">291</a> }<a name="292" href="#292">292</a> <em class="comment">// get next item (real first item of queue)</em><a name="293" href="#293">293</a> status = cursor.getNext(headKey,result,<strong>null</strong>);<a name="294" href="#294">294</a> } <strong>finally</strong> { <a name="295" href="#295">295</a> <strong>if</strong>(cursor!=<strong>null</strong>) {<a name="296" href="#296">296</a> cursor.close();<a name="297" href="#297">297</a> }<a name="298" href="#298">298</a> }<a name="299" href="#299">299</a> <strong>return</strong> status;<a name="300" href="#300">300</a> }<a name="301" href="#301">301</a> <a name="302" href="#302">302</a> <em>/**<em>*</em></em><a name="303" href="#303">303</a> <em> * Put the given CrawlURI in at the appropriate place. </em><a name="304" href="#304">304</a> <em> * </em><a name="305" href="#305">305</a> <em> * @param curi</em><a name="306" href="#306">306</a> <em> * @throws DatabaseException</em><a name="307" href="#307">307</a> <em> */</em><a name="308" href="#308">308</a> <strong>public</strong> <strong>void</strong> put(<a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi, <strong>boolean</strong> overwriteIfPresent) <a name="309" href="#309">309</a> throws DatabaseException {<a name="310" href="#310">310</a> DatabaseEntry insertKey = (DatabaseEntry)curi.getHolderKey();<a name="311" href="#311">311</a> <strong>if</strong> (insertKey == <strong>null</strong>) {<a name="312" href="#312">312</a> insertKey = calculateInsertKey(curi);<a name="313" href="#313">313</a> curi.setHolderKey(insertKey);<a name="314" href="#314">314</a> }<a name="315" href="#315">315</a> DatabaseEntry value = <strong>new</strong> DatabaseEntry();<a name="316" href="#316">316</a> crawlUriBinding.objectToEntry(curi, value);<a name="317" href="#317">317</a> <em class="comment">// Output tally on avg. size if level is FINE or greater.</em><a name="318" href="#318">318</a> <strong>if</strong> (LOGGER.isLoggable(Level.FINE)) {<a name="319" href="#319">319</a> tallyAverageEntrySize(curi, value);<a name="320" href="#320">320</a> }<a name="321" href="#321">321</a> OperationStatus status; <a name="322" href="#322">322</a> <strong>if</strong>(overwriteIfPresent) {<a name="323" href="#323">323</a> status = pendingUrisDB.put(<strong>null</strong>, insertKey, value);<a name="324" href="#324">324</a> } <strong>else</strong> {<a name="325" href="#325">325</a> status = pendingUrisDB.putNoOverwrite(<strong>null</strong>, insertKey, value);<a name="326" href="#326">326</a> }<a name="327" href="#327">327</a> <strong>if</strong>(status!=OperationStatus.SUCCESS) {<a name="328" href="#328">328</a> LOGGER.severe(<span class="string">"failed; "</span>+status+ <span class="string">" "</span>+curi);<a name="329" href="#329">329</a> }<a name="330" href="#330">330</a> }<a name="331" href="#331">331</a> <a name="332" href="#332">332</a> <strong>private</strong> <strong>long</strong> entryCount = 0;<a name="333" href="#333">333</a> <strong>private</strong> <strong>long</strong> entrySizeSum = 0;<a name="334" href="#334">334</a> <strong>private</strong> <strong>int</strong> largestEntry = 0;<a name="335" href="#335">335</a> <a name="336" href="#336">336</a> <em>/**<em>*</em></em><a name="337" href="#337">337</a> <em> * Log average size of database entry.</em><a name="338" href="#338">338</a> <em> * @param curi CrawlURI this entry is for.</em><a name="339" href="#339">339</a> <em> * @param value Database entry value.</em><a name="340" href="#340">340</a> <em> */</em><a name="341" href="#341">341</a> <strong>private</strong> <strong>synchronized</strong> <strong>void</strong> tallyAverageEntrySize(<a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi,<a name="342" href="#342">342</a> DatabaseEntry value) {<a name="343" href="#343">343</a> entryCount++;<a name="344" href="#344">344</a> <strong>int</strong> length = value.getData().length;<a name="345" href="#345">345</a> entrySizeSum += length;<a name="346" href="#346">346</a> <strong>int</strong> avg = (<strong>int</strong>) (entrySizeSum/entryCount);<a name="347" href="#347">347</a> <strong>if</strong>(entryCount % 1000 == 0) {<a name="348" href="#348">348</a> LOGGER.fine(<span class="string">"Average entry size at "</span>+entryCount+<span class="string">": "</span>+avg);<a name="349" href="#349">349</a> }<a name="350" href="#350">350</a> <strong>if</strong> (length>largestEntry) {<a name="351" href="#351">351</a> largestEntry = length; <a name="352" href="#352">352</a> LOGGER.fine(<span class="string">"Largest entry: "</span>+length+<span class="string">" "</span>+curi);<a name="353" href="#353">353</a> <strong>if</strong>(length>(2*avg)) {<a name="354" href="#354">354</a> LOGGER.fine(<span class="string">"excessive?"</span>);<a name="355" href="#355">355</a> }<a name="356" href="#356">356</a> }<a name="357" href="#357">357</a> }<a name="358" href="#358">358</a> <a name="359" href="#359">359</a> <em>/**<em>*</em></em><a name="360" href="#360">360</a> <em> * Calculate the 'origin' key for a virtual queue of items</em><a name="361" href="#361">361</a> <em> * with the given classKey. This origin key will be a </em><a name="362" href="#362">362</a> <em> * prefix of the keys for all items in the queue. </em><a name="363" href="#363">363</a> <em> * </em><a name="364" href="#364">364</a> <em> * @param classKey String key to derive origin byte key from </em><a name="365" href="#365">365</a> <em> * @return a byte array key </em><a name="366" href="#366">366</a> <em> */</em><a name="367" href="#367">367</a> <strong>static</strong> byte[] calculateOriginKey(String classKey) {<a name="368" href="#368">368</a> byte[] classKeyBytes = <strong>null</strong>;<a name="369" href="#369">369</a> <strong>int</strong> len = 0;<a name="370" href="#370">370</a> <strong>try</strong> {<a name="371" href="#371">371</a> classKeyBytes = classKey.getBytes(<span class="string">"UTF-8"</span>);<a name="372" href="#372">372</a> len = classKeyBytes.length;<a name="373" href="#373">373</a> } <strong>catch</strong> (UnsupportedEncodingException e) {<a name="374" href="#374">374</a> <em class="comment">// should be impossible; all JVMs must support UTF-8</em><a name="375" href="#375">375</a> e.printStackTrace();<a name="376" href="#376">376</a> }<a name="377" href="#377">377</a> byte[] keyData = <strong>new</strong> byte[len+1];<a name="378" href="#378">378</a> System.arraycopy(classKeyBytes,0,keyData,0,len);<a name="379" href="#379">379</a> keyData[len]=0;<a name="380" href="#380">380</a> <strong>return</strong> keyData;<a name="381" href="#381">381</a> }<a name="382" href="#382">382</a> <a name="383" href="#383">383</a> <em>/**<em>*</em></em><a name="384" href="#384">384</a> <em> * Calculate the insertKey that places a CrawlURI in the</em><a name="385" href="#385">385</a> <em> * desired spot. First bytes are always classKey (usu. host)</em><a name="386" href="#386">386</a> <em> * based -- ensuring grouping by host -- terminated by a zero</em><a name="387" href="#387">387</a> <em> * byte. Then 8 bytes of data ensuring desired ordering </em><a name="388" href="#388">388</a> <em> * within that 'queue' are used. The first byte of these 8 is</em><a name="389" href="#389">389</a> <em> * priority -- allowing 'immediate' and 'soon' items to </em><a name="390" href="#390">390</a> <em> * sort above regular. Next 1 byte is 'cost'. Last 6 bytes </em><a name="391" href="#391">391</a> <em> * are ordinal serial number, ensuring earlier-discovered </em><a name="392" href="#392">392</a> <em> * URIs sort before later. </em><a name="393" href="#393">393</a> <em> * </em><a name="394" href="#394">394</a> <em> * NOTE: Dangers here are:</em><a name="395" href="#395">395</a> <em> * (1) priorities or costs over 2^7 (signed byte comparison)</em><a name="396" href="#396">396</a> <em> * (2) ordinals over 2^48</em><a name="397" href="#397">397</a> <em> * </em><a name="398" href="#398">398</a> <em> * Package access & static for testing purposes. </em><a name="399" href="#399">399</a> <em> * </em><a name="400" href="#400">400</a> <em> * @param curi</em><a name="401" href="#401">401</a> <em> * @return a DatabaseEntry key for the CrawlURI</em><a name="402" href="#402">402</a> <em> */</em><a name="403" href="#403">403</a> <strong>static</strong> DatabaseEntry calculateInsertKey(<a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi) {<a name="404" href="#404">404</a> byte[] classKeyBytes = <strong>null</strong>;<a name="405" href="#405">405</a> <strong>int</strong> len = 0;<a name="406" href="#406">406</a> <strong>try</strong> {<a name="407" href="#407">407</a> classKeyBytes = curi.getClassKey().getBytes(<span class="string">"UTF-8"</span>);<a name="408" href="#408">408</a> len = classKeyBytes.length;<a name="409" href="#409">409</a> } <strong>catch</strong> (UnsupportedEncodingException e) {<a name="410" href="#410">410</a> <em class="comment">// should be impossible; all JVMs must support UTF-8</em><a name="411" href="#411">411</a> e.printStackTrace();<a name="412" href="#412">412</a> }<a name="413" href="#413">413</a> byte[] keyData = <strong>new</strong> byte[len+9];<a name="414" href="#414">414</a> System.arraycopy(classKeyBytes,0,keyData,0,len);<a name="415" href="#415">415</a> keyData[len]=0;<a name="416" href="#416">416</a> <strong>long</strong> ordinalPlus = curi.getOrdinal() & 0x0000FFFFFFFFFFFFL;<a name="417" href="#417">417</a> ordinalPlus = <a name="418" href="#418">418</a> ((<strong>long</strong>)curi.getSchedulingDirective() << 56) | ordinalPlus;<a name="419" href="#419">419</a> ordinalPlus = <a name="420" href="#420">420</a> ((((<strong>long</strong>)curi.getHolderCost()) & 0xFFL) << 48) | ordinalPlus;<a name="421" href="#421">421</a> ArchiveUtils.longIntoByteArray(ordinalPlus, keyData, len+1);<a name="422" href="#422">422</a> <strong>return</strong> <strong>new</strong> DatabaseEntry(keyData);<a name="423" href="#423">423</a> }<a name="424" href="#424">424</a> <a name="425" href="#425">425</a> <em>/**<em>*</em></em><a name="426" href="#426">426</a> <em> * Delete the given CrawlURI from persistent store. Requires</em>
?? 快捷鍵說明
復制代碼
Ctrl + C
搜索代碼
Ctrl + F
全屏模式
F11
切換主題
Ctrl + Shift + D
顯示快捷鍵
?
增大字號
Ctrl + =
減小字號
Ctrl + -