Atom - Taxonomy Issue

[Sangeeta K]

We have had a few issues with our uploads  in regards to Physical Storage.

 

When uploading some records via csv on the user interface, one of the uploads was stopped and showed this error:

 Please see a.jpg

 

We checked the data in the CSV and it was all correct. I then checked the Physical Object Type taxonomy in Atom and found that there were two entries for Box. One was in under Container (the one we set up to use) as shown here:

 Please see b.jpg

 

 

And the second was in the top level of the physical object type as shown here:

 Please see c.jpg

I deleted the second Box version. We then tried the upload again and received the following message:

 Please see d.jpg

 

The objects imported, but in the taxonomy, the second instance of the Box had been created again but this time with a warning saying that the term box already exists as shown here:

Please see e.jpg

 

  

It seems to me that the narrow term of Box in the broad term Container is not being connected to and as a result Atom is creating additional physical object types. 

[Dan Gillean]

Hi Sangeeta, 

I've not seen that error before, and currently I'm not sure what is causing it. It may have something to do with your particular installation environment - you mentioned that you are installed on RHEL - do you know if you installed the MySQL distribution we recommend? Percona 5.6?

I ran a quick test in our 2.4 development envrionment - we have made a number of changes to the import code, so it's possible that if there is a bug, it has been resolved. I'm attaching the CSV I used for my quick test - it includes one parent fonds record, and several item-level children. Each of the children use different physicalStorageType values - I used Box twice, as well as Hollinger box and Cardboard box. I also used a test term, "Wackadoodle," that was not in the taxonomy, to ensure that AtoM could create it without issue on import. 

In my test, all existing types used the existing terms on import. Wackadoodle was created, added to the Physical Object Type taxonomy, and noted in the job details page without causing any errors or warnings. So, in my local 2.4 instance, everything appears to be okay...

We have a task that can merge duplicate terms in a taxonomy together, but unfortunately there is a bug in the task in the 2.3 release, which will be fixed in the 2.4 release. So, if you are not continuing to get the strange Serialization failure error anymore, one approach you could take is to continue with your imports and ignore the duplicate terms for now. Then when 2.4 is released, you can run the taxonomy:normalize task to merge the duplicates. 

• Here is the current documentation on the taxonomy:normalize task: https://www.accesstomemory.org/docs/2.3/admin-manual/maintenance/cli-tools/#taxonomy-normalization 
  
• Here is the bug ticket for the task not currently working in 2.3: https://projects.artefactual.com/issues/10698
  
• Here is the related pull request that fixed it: https://github.com/artefactual/atom/pull/545/files
  

This is obviously not an ideal solution, but merely a workaround until you can upgrade to a release that hopefully solves the issue. This also depends on the SQL error not re-occurring. If this error continues to happen, you may want to review your installation configuration and ensure that all recommended dependencies are installed, and the right versions of each tool/library/dependency are used. 

Regards, 

Dan Gillean, MAS, MLIS
AtoM Program Manager
Artefactual Systems, Inc.
604-527-2056

@accesstomemory

--
You received this message because you are subscribed to the Google Groups "AtoM Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ica-atom-users+unsubscribe@googlegroups.com.
To post to this group, send email to ica-atom-users@googlegroups.com.
Visit this group at https://groups.google.com/group/ica-atom-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/ica-atom-users/51c7ed3d-c391-462b-aeaf-7982d30f135e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Sangeeta K]

Hi Dan,

We are using MariaDB 5.5.52. As per the documentation, MariaDB is supported. Could it be a reason for this error?

# mysql --version

mysql  Ver 15.1 Distrib 5.5.52-MariaDB, for Linux (x86_64) using readline 5.1

[Sangeeta K]

We don't see this issue as of now. We will let you know if this issue reoccurs.

We see one more issue, details below:

I uploaded a CSVwith terms in the subject access point and place access point taxonomies. These terms are set up as use for terms of the broader terms in the relating atom taxonomies.

I understand that Atom should function to convert these use for terms into the approved terms within the taxonomy. Ie NSW should be replace with NEW SOUTH WALES in the place taxonomy and Typist and Branch Manager should be converted to human resources in the subjects taxonomy.

Instead, atom is created new terms for NSW in the places taxonomy and new terms for Typist and Branch Manager in the subject taxonomy

 

Interestingly, when entering an archival description into atom directly (not via CSV) the option of the use for terms does appear as shown below. 

Please see 1.jpg

but if entered via CSV, NSW is not converted properly as seen in L_7733.tmp snapshot.

Thanks and Regards

Sangeeta

On Thursday, July 20, 2017 at 8:25:02 PM UTC+10, Sangeeta K wrote:

[Dan Gillean]

Hi Sangeeta, 

Unfortunately, the "Use for" functionality in AtoM is curently very limited - it can be used for display, but it will not automatically swap deprecated terms for preferred ones. If you have a term called "NSW" that already exists, and you import a CSV that has a place access point whose name matches exactly, then AtoM *should* use the existing term instead of creating a new one, but it will not substitute the term for the preferred "use" term. 

If you select a term that has a "Use" relationship, AtoM will still use the original term, but it will display the preferred term in parentheses. We would love to enhance this functionality but the very permissive way it behaves currently was added early on, and the module has not been significantly developed since. It would require community support for Artefactual to be able to improve linking and importing behavior around preferred terms. 

Regards, 

Dan Gillean, MAS, MLIS
AtoM Program Manager
Artefactual Systems, Inc.
604-527-2056

@accesstomemory

--

You received this message because you are subscribed to the Google Groups "AtoM Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ica-atom-users+unsubscribe@googlegroups.com.
To post to this group, send email to ica-atom-users@googlegroups.com.
Visit this group at https://groups.google.com/group/ica-atom-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/ica-atom-users/15795e8c-4478-47d1-978d-75438b7f18d8%40googlegroups.com.