I have an error with the maven build, so I have test with an external 1.17 Tika Server but, POI not included. If you success a mvn package with 1.17 Tika, I am interested.
Today, I have not had much time to deal with it. I found some bugs that I would declare tomorrow if they are not already. They concern log4j2, local_fr and a bug with the web interface and the keyboard input key. I continu my investigation. De : Karl Wright [mailto:daddy...@gmail.com] Envoyé : mercredi 10 janvier 2018 17:15 À : user@manifoldcf.apache.org Objet : Re: Document connector excluding mime type and size - Tika Parser error Any news? Karl On Tue, Jan 9, 2018 at 1:10 PM, Karl Wright <daddy...@gmail.com <mailto:daddy...@gmail.com> > wrote: Let me know what happens. If it works for you, I'll see if we can put together a patch release of 2.9 with the fix. Karl On Tue, Jan 9, 2018 at 11:07 AM, msaunier <msaun...@citya.com <mailto:msaun...@citya.com> > wrote: Test check out and building with POI 3.17 and Tika 1.17? It’s possible. I finish a project and I test that. De : Karl Wright [mailto:daddy...@gmail.com <mailto:daddy...@gmail.com> ] Envoyé : mardi 9 janvier 2018 16:57 À : user@manifoldcf.apache.org <mailto:user@manifoldcf.apache.org> Objet : Re: Document connector excluding mime type and size - Tika Parser error So here's the problem; we used POI 3.17 with Tika 3.16 in 2.9, in order to deal with the classloader issue present in POI 3.15, and because POI 3.16 has a severe security issue that made it impossible to ship with. Unfortunately that doesn't quite work; POI 3.17 is not backwards compatible with 3.16 completely and therefore problems occur with this combination. The probable solution is to check out and build trunk and see if that works for you. It very well might. The question then is what to do next, because we are not scheduled to release again until April. We might have to do a point release to deal with this. Please give it a try and let me know what happens. Thanks, Karl On Tue, Jan 9, 2018 at 10:29 AM, Karl Wright <daddy...@gmail.com <mailto:daddy...@gmail.com> > wrote: Ok, never mind that last email. We patched it in part in 2.9 by including the latest POI. So clearly it's still an existing problem in POI. I'll have to open a ticket there and await a patch from them. Karl On Tue, Jan 9, 2018 at 10:27 AM, Karl Wright <daddy...@gmail.com <mailto:daddy...@gmail.com> > wrote: This screenshot cannot be MCF 2.9 since the version of poi was not 3.17 for the 2.9 release. Karl On Tue, Jan 9, 2018 at 10:02 AM, msaunier <msaun...@citya.com <mailto:msaun...@citya.com> > wrote: They 2 versions (2.8.1 and 2.9) of ManifoldCF are on 2 differents servers. De : Karl Wright [mailto:daddy...@gmail.com <mailto:daddy...@gmail.com> ] Envoyé : mardi 9 janvier 2018 15:54 À : user@manifoldcf.apache.org <mailto:user@manifoldcf.apache.org> Objet : Re: Document connector excluding mime type and size - Tika Parser error As for the Tika issue, we explicitly tested documents of that type when rolling out 2.8.1. When we updated 2.8.1 to a new Tika in 2.9 I believe we also tested this. One of the potential issues is that if you are dropping down different versions of ManifoldCF into the same directories you *might* have a poi* jar in the wrong place because of the way we had to do the patch. Please have a look at where the poi* jars are in your directory structure; they should all be in one directory (connector-common-lib). If you see any anywhere else, that's the cause of the issue. Karl On Tue, Jan 9, 2018 at 9:43 AM, Karl Wright <daddy...@gmail.com <mailto:daddy...@gmail.com> > wrote: Since the Tika extractor essentially filters out the content mime type (other than presenting it as metadata), you need to put an "allowed documents" transformation connection into your job pipeline BEFORE the Tika connector: https://manifoldcf.apache.org/release/release-2.9/en_US/end-user-documentation.html#alloweddocuments In fact, mime type exclusion is actually disabled in the Solr output connector *unless* you are using the extracting update handler. That should resolve the one problem for you. Thanks, Karl On Tue, Jan 9, 2018 at 9:35 AM, msaunier <msaun...@citya.com <mailto:msaun...@citya.com> > wrote: They document for Tika are : · Microsoft Word 97-2003 · Application/msword I can’t have more informations, they are in SCO servers and SCO do not have ls –lisan or stat command. For SolR connecting, I seem to have emptied the index before the last indexation. (ManifoldCF and Solr) I do it again to be sure. De : Karl Wright [mailto:daddy...@gmail.com <mailto:daddy...@gmail.com> ] Envoyé : mardi 9 janvier 2018 15:26 À : user@manifoldcf.apache.org <mailto:user@manifoldcf.apache.org> Objet : Re: Document connector excluding mime type and size - Tika Parser error CONNECTORS-1482 is for the Solr connector filtering issue. A question: When you changed these fields in the output connection, had you already indexed any documents? Those would only get cleaned up if you did a subsequent full crawl, after you made the connection change. Karl On Tue, Jan 9, 2018 at 9:22 AM, Karl Wright <daddy...@gmail.com <mailto:daddy...@gmail.com> > wrote: If you let me know what kind of file they are (extension and what application created them) that is probably good enough. Karl On Tue, Jan 9, 2018 at 9:19 AM, msaunier <msaun...@citya.com <mailto:msaun...@citya.com> > wrote: Okay good. I look if I can test 1.17 Tika version. I can’t transfert a document with this error, they are privates. Sorry. If I encounter the error again on a non-private document, I'll come back to you. De : Karl Wright [mailto:daddy...@gmail.com <mailto:daddy...@gmail.com> ] Envoyé : mardi 9 janvier 2018 15:12 À : user@manifoldcf.apache.org <mailto:user@manifoldcf.apache.org> Objet : Re: Document connector excluding mime type and size - Tika Parser error CONNECTORS-1481 is the ticket for the Tika problem. Karl On Tue, Jan 9, 2018 at 8:34 AM, Karl Wright <daddy...@gmail.com <mailto:daddy...@gmail.com> > wrote: Ok, if you are in a position to build trunk, that's a newer version of Tika (1.17) which might (or might not) address this problem. If you could create a ticket, I'd greatly appreciate attaching one document to it that causes the failure. Thanks! Karl On Tue, Jan 9, 2018 at 8:02 AM, msaunier <msaun...@citya.com <mailto:msaun...@citya.com> > wrote: It’s a 2.9 version. I have a 2.8.1 in an other server with same job and same documents. I will test on this other server and make you a return. Thanks for your help. De : Karl Wright [mailto:daddy...@gmail.com <mailto:daddy...@gmail.com> ] Envoyé : mardi 9 janvier 2018 13:15 À : user@manifoldcf.apache.org <mailto:user@manifoldcf.apache.org> Objet : Re: Document connector excluding mime type and size - Tika Parser error I looked at the history of this. We had to release a patch (2.8.1) that put various poi jars at root level in order to work around a Tika problem. That patch may not have been entirely correct in that it looks like it may have blocked access by one of the deeper jars to a higher level. Release 2.9 should fix this if I am correct. Karl On Tue, Jan 9, 2018 at 6:39 AM, Karl Wright <daddy...@gmail.com <mailto:daddy...@gmail.com> > wrote: What version of MCF is this? That's important to know since Tika has had problems with this kind of thing in the past and this looks like something similar. The problem you are reporting is due to either a missing jar, or a bug in an internal tika classloader. But I need to know whether this is a current bug or not, since we just went to a new Tika version. Karl On Tue, Jan 9, 2018 at 4:32 AM, msaunier <msaun...@citya.com <mailto:msaun...@citya.com> > wrote: Hello Karl, I hope you are well today. I have 2 problems with ManifoldCF. ----------- In *Outputs connectors* with Solr connector. I have add a « Maximum document length and I have « Excluded 5 mime types » but it not work. I join capture. ---------- And in second, I have a *Tika exception* in ManifoldCF. 3 documents are blocked : FATAL 2018-01-09T10:19:54,992 (Worker thread '5') - Error tossed: org.apache.poi.hwmf.record.HwmfFont.getCharSet()Lorg/apache/poi/hwmf/record/HwmfFont$WmfCharset; java.lang.NoSuchMethodError: org.apache.poi.hwmf.record.HwmfFont.getCharSet()Lorg/apache/poi/hwmf/record/HwmfFont$WmfCharset; at org.apache.tika.parser.microsoft.WMFParser.parse(WMFParser.java:74) ~[?:?] at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:280) ~[?:?] at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:280) ~[?:?] at org.apache.tika.parser.AutoDetectParser.parse(AutoDetectParser.java:135) ~[?:?] at org.apache.tika.parser.DelegatingParser.parse(DelegatingParser.java:72) ~[?:?] at org.apache.tika.extractor.ParsingEmbeddedDocumentExtractor.parseEmbedded(ParsingEmbeddedDocumentExtractor.java:102) ~[?:?] at org.apache.tika.parser.microsoft.ooxml.AbstractOOXMLExtractor.handleEmbeddedFile(AbstractOOXMLExtractor.java:375) ~[?:?] at org.apache.tika.parser.microsoft.ooxml.AbstractOOXMLExtractor.handleEmbeddedPart(AbstractOOXMLExtractor.java:260) ~[?:?] at org.apache.tika.parser.microsoft.ooxml.AbstractOOXMLExtractor.handleEmbeddedParts(AbstractOOXMLExtractor.java:205) ~[?:?] at org.apache.tika.parser.microsoft.ooxml.AbstractOOXMLExtractor.getXHTML(AbstractOOXMLExtractor.java:142) ~[?:?] at org.apache.tika.parser.microsoft.ooxml.OOXMLExtractorFactory.parse(OOXMLExtractorFactory.java:142) ~[?:?] at org.apache.tika.parser.microsoft.ooxml.OOXMLParser.parse(OOXMLParser.java:106) ~[?:?] at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:280) ~[?:?] at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:280) ~[?:?] at org.apache.tika.parser.AutoDetectParser.parse(AutoDetectParser.java:135) ~[?:?] at org.apache.manifoldcf.agents.transformation.tika.TikaParser.parse(TikaParser.java:74) ~[?:?] at org.apache.manifoldcf.agents.transformation.tika.TikaExtractor.addOrReplaceDocumentWithException(TikaExtractor.java:235) ~[?:?] at org.apache.manifoldcf.agents.incrementalingest.IncrementalIngester$PipelineAddEntryPoint.addOrReplaceDocumentWithException(IncrementalIngester.java:3226) ~[mcf-agents.jar:?] at org.apache.manifoldcf.agents.incrementalingest.IncrementalIngester$PipelineAddFanout.sendDocument(IncrementalIngester.java:3077) ~[mcf-agents.jar:?] at org.apache.manifoldcf.agents.incrementalingest.IncrementalIngester$PipelineObjectWithVersions.addOrReplaceDocumentWithException(IncrementalIngester.java:2708) ~[mcf-agents.jar:?] at org.apache.manifoldcf.agents.incrementalingest.IncrementalIngester.documentIngest(IncrementalIngester.java:756) ~[mcf-agents.jar:?] at org.apache.manifoldcf.crawler.system.WorkerThread$ProcessActivity.ingestDocumentWithException(WorkerThread.java:1583) ~[mcf-pull-agent.jar:?] at org.apache.manifoldcf.crawler.system.WorkerThread$ProcessActivity.ingestDocumentWithException(WorkerThread.java:1548) ~[mcf-pull-agent.jar:?] at org.apache.manifoldcf.crawler.connectors.sharedrive.SharedDriveConnector.processDocuments(SharedDriveConnector.java:939) ~[?:?] at org.apache.manifoldcf.crawler.system.WorkerThread.run(WorkerThread.java:399) [mcf-pull-agent.jar:?] I need to create an incident ticket? ---------- Thanks for your help. Cordialement,
