cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Full text indexing / search - SolidWorks drawings

mlockwood
19-Tanzanite

Full text indexing / search - SolidWorks drawings

I may have posted this earlier - don't remember.

We now have indexing in place (Solr - 10.2 M020), working perfectly for Creo Drawings.   Hoping to extend to SolidWorks drawings.

Tech support case C12917558 has been open for a bit.  In general tech support has been saying (not totally confidently) that indexing of SolidWorks drawings seems to be working on their end but no Solid Works drawings are retuned using keyword search for text on the drawings.

Looking for anyone who has experience with this.

thanks

12 REPLIES 12

Hi Mike,

When you ran the wt.content.DataFormatUtil utility which mime types did you specify?

I passed this along to the TSE handling your case, but I thought I'd also mention it here. I found this suggestion in another similar TS case in order to help troubleshoot this kind of issue:

Enable the following verbosity and collect the log (MS/BGMS as per hosted Solr):


log4j.logger.com.ptc.search.solr.contentReader.contentExtraction.BinaryExtractor=TRACE

You can perform this activity on TEST environment. User can add suspected objects (DRW/PRT) which will go for indexing and we can collect the logs i.e. actual content being extracted.

In addition to the logs please send list of keywords which you are looking to be search against those object. For, example, we would be interested in a string like “Extracted Content=” in the collected log.

mlockwood
19-Tanzanite
(To:LoriSood)

Today I received a definitive answer on Case C12917558 stating that SolidWorks drawings cannot be indexed by Solr.  Major, major disappointment.  I opened the case Jan 27 and got the answer Apr 5 - wow.

The answer does not seem to be based on any experimentation or consultation with R&D - or even on what would be good to provide to customers.

PTC or anyone: Is there any way to get SolidWorks drawings to index and allow full text search for drawing not content?

thanks

TomU
23-Emerald II
(To:mlockwood)

Mike,

Here are a few different articles you might want to review.  One of them provides a method to test the indexing on a particular file.  Since SOLR is unable to index compressed Creo drawings, I'm guessing it may be having the same issue with SolidWorks, but at least this should let you verify.  FWIW.

mlockwood
19-Tanzanite
(To:TomU)

Thanks for the links - nothing new here that I haven't already been thru.

on this one, I can't get the command to work no matter what I try.  It doesn't recognize >java org.

https://support.ptc.com/appserver/cs/view/solution.jsp?n=CS32720

java org.apache.tika.cli.TikaCLI “<filename>”

Discouraging that there seems to no real message from PTC on this (index SW drawing text).

Hello Mike,

I remember you have to set the full path to the jar, like this "set classpath=D:\PTC\Windchill_10.2\Windchill\solr-webapp\WEB-INF\lib\apache-tika-standalone.jar"  to run the command.

Thank you

Binesh

Thanks - I carefully did exactly what you state above.  Returns from set classpath statement w/no errors.  Entering just >java org. or anything after, it says that "org" not recognized.  Tried all possible variations that I could think of.  It correctly finds the folders and files (confirmed by using tab for each) but something small is missing in the writeup or my execution.

It looks like it is not recognizing the class path. Do you see apache-tika-standalone.jar file in your D:\PTC\Windchill_10.2\Windchill\solr-webapp\WEB-INF\lib\ directory? I was able to run it in our dev system, I also ran it for a slddrw, here is the screenshot.

tika.png

Thank you

Binesh Kumar

Thanks - may have been UPPER/lower case issue.  Output looks exactly like you posted.

Given this, how can one tell if in fact a SolidWorks drawing can be indexed?  The output lists all the parameters, etc. but I'm not seeing where it reports the needed info about being indexable.

TomU
23-Emerald II
(To:mlockwood)

If I understand correctly, the output of this command should show you everything in this document that is being indexed.  If you don't see the data from tables, notes, and dimensions, then those are not being indexed.  Maybe run this command against and Creo model and compare the results.

See this for more info:  https://support.ptc.com/appserver/cs/view/solution.jsp?n=CS101260

BineshKumar1
12-Amethyst
(To:TomU)

Hello Tom,

I tried that this evening and it doesn't appear to read any of the actual contents from a Solidworks file. It can read the basic document properties and meta data by using the microsoft/default parser. I couldn't find a parser in the code to read Solidworks content. So I guess we can index only metadata with the current version of Solr which comes with Windchill, I tried in WC 11 too

I couldn't see the Solidworks type (SLDWORKS) in the supported list of MIME types.

Thank you,

Binesh Kumar

Who is the current PTC Project Manager for indexing?  Please jump in and provide some info.

thanks very much

Vikram Sinha‌  - Are you able to comment on this, or do you know who could?

Announcements