kitodo / kitodo-production

Kitodo.Production is a workflow management tool for mass digitization and is part of the Kitodo Digital Library Suite.
http://www.kitodo.org/software/kitodoproduction/
GNU General Public License v3.0
62 stars 63 forks source link

2.x Export Resources into Hotfolder #2614

Closed M3ssman closed 5 years ago

M3ssman commented 5 years ago

Hello,

I'm struggeling with Kitodo 2.x. Is it possible to export other types of Ressources than Images from the Metadata Folder into the Hotfolder within the Step "Export to DMS"?

Our Usecase: We have a Script attached to the final DMS-Export that creates custom XML- and Textfiles that must be processed in later Export Stages. We tried to create these Files inside the (tifpath)-Folder in the Metadata-Ressort of Kitodo with a custom Script. So far, this works well, with the Files being created. But the Files aren't included in the Kitodo-Hotfolder when the final Export takes place, and therefore missing at later Stages.

Going further down the Kitodo 2.x-Branch, there's a a Class de.sub.goobi.export.dms.AutomaticDmsExportWithoutHibernate which forwards filtering to Helper.dataFilter. At first, this seems to be an Image-Filter, but at second, it's a bit strange. Why are Files with Extensions .xsl , .ppt or .doc considered legal according to this list, but plain XML or .txt get skipped?

Kitodo 2.x

Greetings, M3ssman

henning-gerhardt commented 5 years ago

In general you can use the processDirs property in goobi_config.properties to create this kind of directories and this directories are exported too. In general inside the images directory only image like files should stored in this place and maybe through historical reasons other kind of files are stored there too. If you have other files like from OCR processing you can use the ocrdirectory on the same level like the images directory. As we at SLUB only have XML files from OCR in a subdirectory and no other kind of files I don't know if other files would be copied too. You don't need to define the ocr subdirectory inside the processDirs property as all subdirectories from ocr should be copied on export.

M3ssman commented 5 years ago

@henning-gerhardt Thanks for your Advices!

Are you referring to the processDirs Property or rather to the Data copy rules inside goobi_config.properties ?

henning-gerhardt commented 5 years ago

This is depending on what you want to do :-D

If you want to create and export directories then you should work with processDirs property. If you want to manipulate the meta data on export then you should work with the copyData.onExport property. You can use both if you need both. This properties are for independent features of Kitodo.Production.

M3ssman commented 5 years ago

@henning-gerhardt

Thanks! I'll give a try and report back tomorrow!

M3ssman commented 5 years ago

Hm, I didn't work it out.

To provide some more Context:

We need to pass additional Data to a DSpace-Repository, i.e. a collection-File (exactly this name, no extension). Before the DMS-Export actualles takes place, I create this File in a custom Subfolder <processtitle>/dspace . When DMS is triggered, File and Folder are not transferred into the Kitodo-content-Folder.

So I went for goobi_config.properties and added:

processDirs=dspace/(processtitle)_dspace

Hopefully, to export the metadata/<processid>/dspace- Folder into something like hotfolder/<processtitle>_dspace , but this had no effect.

henning-gerhardt commented 5 years ago

Is a XML file get copied if you put an XML file into the dspace/(processtitle)_dspace) directory?

frank-ulrich-weber commented 5 years ago

Dear Mr. Hartwig,

If your folder structure looks like this:

… …/metadata//image/… …/metadata//dspace/… …

You have to configure processDirs as in the following:

processDirs=dspace

It just copies the …/metadata//dspace folder to the configured “DMS import path“. There is, at the moment, no way to define a different folder-name (e.g. (processtitle)_dspace) for the export.

Best regards

Frank Ulrich Weber

From: Uwe Hartwig [mailto:notifications@github.com] Sent: Tuesday, June 4, 2019 3:27 PM To: kitodo/kitodo-production Cc: Subscribed Subject: Re: [kitodo/kitodo-production] 2.x Export Resources into Hotfolder (#2614)

Hm, I didn't work it out.

To provide some more Context:

We need to pass additional Data to a DSpace-Repository, i.e. a collection-File (exactly this name, no extension). Before the DMS-Export actualles takes place, I create this File in a custom Subfolder /dspace . When DMS is triggered, File and Folder are not transferred into the Kitodo-content-Folder.

So I went for goobi_config.properties and added:

processDirs=dspace/(processtitle)_dspace

Hopefully, to export the metadata//dspace- Folder into something like hotfolder/_dspace , but this had no effect.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/kitodo/kitodo-production/issues/2614?email_source=notifications&email_token=ABDRPAREMTNHC7UHLHDNLSDPYZUTPA5CNFSM4HRZXOZKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODW4R4MI#issuecomment-498671153, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ABDRPARFVKUBZ6MM7K2XLWTPYZUTPANCNFSM4HRZXOZA.

M3ssman commented 5 years ago

@henning-gerhardt No, the folder is completely missing and nothing is copied.

@frank-ulrich-weber Yes, it is transferred this way. Unfortunately, the folder is not prefixed alike the the Image Export. As I found out, when no Value for proceedDirs is present, for example:

processedDirs=

all Metadata directories (in my case: tif and dspace) are copied to the Hotfolder without restrictions (and, sadly, prefixations).

But, taken further, what happens, if there are more than one Kitodo-Exports running at the same time? I expect the worst, since the Hotfolder dirs are not distinguishable by a Prefix like the tif-Folders, so we might end up with inconsistent data.

The best approach seems to store additional data outside Kitodo into some arbitrary temp-Folder, that can be dismissed afterwards.

frank-ulrich-weber commented 5 years ago

Check the option "Create process folder" within the project settings and kitodo.production creates a superordinate folder for the export. No more problems with inconsistent data.

M3ssman commented 5 years ago

@frank-ulrich-weber Yes, this works as suggested. But our Problem is somehow more complex. We use an additional Module from Zeutschel called zedExporter and it looks this has some special requirements. It doesn't work out-of-the-box with these subordinate Process-Folders.

frank-ulrich-weber commented 5 years ago

OK, I can confirm this behaviour of zedExporter. I added this requirement to the list of new features for the next release. Thank you!