Auto Classification Rules Based on Tags?


Does anyone have any method/script that can auto group documents based on tags?

For me, the auto classify feature has never really been effective. It would be good to be able to establish rules/logic for grouping. Effectively it would be great if you could take the rule-based nature of the smart group but allow it to actually physically group the documents as well.



Why don’t you just select the items in the Smart Group and Group them (noting this will move the files themselves fromtheir original locations)?

Note this will only work with a local Smart Group, not a Global Smart Group in your sidebar.

Why are smart groups not sufficient in this case? E.g. do you want to add additional items to the groups on your own? Or do you want to create subgroups? Would nested smart groups solve the problem?

Smart Smart Groups? :open_mouth: :exclamation: Yes, please. :smiley:

I think what he’s asking is this:

Instead of (1) tag the item (2) classify the item, you have a single process that classifies the item as soon as it’s tagged based on the tag applied.

It is not about how the item is displayed, it is about saving a step.

For example, the user tags an item “Letter” in the inbox and a script immediately moves it to the “Correspondence” group.

llcckk can you clarify your use case? It seems odd that you would need to use a tag if all items of that tag go to the same group anyway. In other words, is the tagging redundant?

For instance, in the above hypothetical a user could use a subgroup called Correspondence > Letters and move the item there, and just not use a tag at all.

When tags were first introduced to DEVONthink their visual representation was the same as the visual representation of groups (a folder), but a different color. Indeed, their logical similarity is very real; groups themselves can be treated as tags in DEVONthink.

By default now (but not when first introduced) groups are checked as excluded from tagging in Database Properties.

If you wish to apply identical group and tag properties to items (which could be viewed as redundant, although not necessarily so), unchecking the option to exclude groups from tagging does precisely that, in one step. Place documents into a group named, e.g., Physics, and that’s both their location and tag. Each document now has one group location and one tag.

Now suppose we replicate one of the documents to another group, e.g., Cosmology. Each instance (replicant) now has one group location. How many tags does it have? (This question can be answered by experiment.)

Groups, tags, aggregation by searches and smart groups (another form of aggregation by search) are ways of applying metadata to a collection of documents. We add metadata to collections of documents to help work with them for various purposes and workflows.

I could choose to create a database holding tens of thousands of documents without using group organization or tags. For my main database that I use for research and writing, my personal practice is to organize documents into groups, each of which defines a relationship (metadata) among documents placed within a group. I almost never tag new content when added (and do exclude groups from tagging almost always), but often find tags useful when I’m working on a project, to assign metadata to items for the purposes of that project. Those tags often result from aggregation of documents useful for the purpose of the project by searches or See Also suggestions–or perhaps because of my knowledge of the discipline and its literature. Because of the way I use such tags, which are related to the project of the moment, I usually remove most tags when a project has been completed. My experience is that they would probably be of little use or even be counterproductive for the purposes of the next project. Note that I don’t treat all my databases the same way, as they differ in types of content and the purposes they serve.

DEVONthink can accommodate a very wide range of workflows. If you find that for your purposes time and effort spent in assigning multiple tags to each item as it is added results in greater productivity in working with the database, do so.