Auto-tag DEVONthink content after importing

Hello,
i am trying to clean up my standard DEVONthink database for articles at the moment, using quite an amount of tags and groups. What i have seen, is, that auto-tagging on import runs fine on html-files, and delivers some incredible results, but on nearly all other filetypes, like pdf and webarchive, no auto-tags are provided. Since i use a lot of webarchives and pdfs, i wanted to ask if there is some other way to auto-tag these files, or if this is a job i would have to really do manually.

Thx for the help,
Martin

Anyone a clue? Or should i create a feature request for adding a function to create auto-tags?

Is it possible that the “tags” you are seeing are displayed under the Keywords column in DT? DT does not have an “auto-tag” feature.

However, HTML files frequently have keywords, and DT will display those in the “Keywords” column. (Tags, on the other hand, are displayed in the tagging bar at the bottom on the document preview and in the Tags portion of the Information pane (Tools > Show Info).) Because webarchives are constructed from HTML pages, the keywords will also display. PDFs can have keywords, too.

Keywords in HTML/webarchives and PDFs are put there by the document’s author, not DT. You may see the raw keywords in HTML if, in Safari, you View > Source and look for in the element for a <meta name=“keywords”…/> element. Keywords are searchable in DT.

The “incredible results” you’re reporting probably are due to the authors of the web pages being careful to include a wide variety of keyword.

To add auto-tagging to DT would be a feature request, as you noted.

I’m trying out Devon coming from Evernote. Recently imported 3k notes, all with tags. I’ve been looking for an auto-tag function as well. In my exploration of the product… it appears there is an auto group function which is intriguing, but no auto tag. So I may need to create groups for all the tagged documents in order to use the auto-group. I have about 300 tags, so it’s not a small task.

This may be for a separate topic, but looking for any clarification for when to use group and when to use tag. One difference between the two is: an item can only exist in one group. But can have multiple tags.

DEVONthink will recognize OpenMeta tags applied to documents from other applications (I believe Evernote does not support OpenMeta), keywords applied to some documents as described above, and DEVONthink will also convert categories in RSS feeds to tags if the preference is set for this. Beyond this, there is no auto-tag function in DEVONthink. Having played with auto-tagging in a competing product, I’d say that this ‘feature’ is not a big loss. The only thing worse than having no tags applied is having the wrong tags applied.

Not exactly. An item can exist in many groups by replicating the document. There is no right or wrong answer on when to use a group and when to use a tag.

On one end of the spectrum, some users have gone away from a group structure almost entirely, choosing to dump all documents into one or a few groups, and applying tags to apply meta data to the documents. At the other end of the spectrum some users still retain a ridged group structure and apply tags sparingly, if at all. Group tags can be turned off on a group-by group basis in a group’s Info pane, or turned off for the entire database via the database properties window. I would speculate that most users are somewhere in the middle, still using a group structure while tagging documents as well.

Thanks for helping me out with understanding how Devon works. Do you know of any Scripts that convert all tags to groups? I’m interested in seeing how well the AI works with suggesting a group.

Figured out an easy way to convert tags to groups. This is kinda cool.

The thing I’m learning is Tags & Groups are practically the same thing in the program. I took one of my tags and dragged it to the top level of the database, and it was converted to a group.

I created a group called Reference and dragged all the tags to that group. The only problem I’m having is the ‘Auto & Classify’ pane is not showing all those newly created groups.

I’ve noticed this in the past when I’ve wanted to move an inbox item into a newly created group. The newly created group doesn’t appear in the ‘Auto & Classify’ pane, until I drag that first item into the group.

Tried closing the software and opening it back up.

Any trick for getting the software to recognize those newly created groups?

DEVONthink compares the content of a selected documents with the documents in the database, and suggest groups for classifying based on similar content. If a group is empty, there is no content for a comparison, thus it will never appear in the classify pane.

Edited to add: If you want a window to drag documents from the inbox to newly created groups, try the Groups & Tags window, available from the Tools menu or the control-command-G shortcut.

That’s how I thought it was working. But all the tags that I dragged over to create groups have content, and yet aren’t showing up.

The window is neat, but I’m really interested in the ‘auto & categorize’ pain. (viewed by clicking on the sorcerer’s hat). It suggests the group an item should be in with a button to ‘move to’.

It’s almost like I need to tell Devon to ‘reindex’ the groups.

I just realized that when I attempt to create a group… a yellow folder is created. Which I assumed was a group. However, if you look at the yellow folder in the Icon view… it has a tag on it.

One time, when I clicked on the gear icon and created a group… a blue folder appeared. But I haven’t gotten that to appear again.

Yellow groups indicate that tagging is enabled for that group-documents placed in it will inherit the name of the group as a group tag. This can be turned off by checking the exclude from tagging box in the info pane for the group. Group tagging can be turned off for the entire database in the database properties pane. Group tags and global tags (tags in the Tag group) work pretty much the same, although group tags will appear in gray while global tags are blue.

Ahh… that’s it. Thanks for the education.