Proposal for a tag manager

Hi there. I have been an on-and-off Zotero user for the past 6 or 7 years. In the beginning, I had never payed much attention to anything related to tags, such as item tagging, the tag selector, tag colors, automatic tags etc. However, I recently found myself having to organize/sort/delete a collection of over two thousand items in Zotero. I realized that using only collections would not cut it, so I started getting myself familiar with tags.

As a long time Zotero user, I felt that the whole tag management experience is very manual and clumsy compared to other parts of the software. For example, anything related to deleting or merging multiple tags had to be done on an individual basis.

As an example, a single three-word term such as "multi-agent systems" was actually represented by multiple tags (with and without the hyphen, different words capitalized) totaling 12 different tags which had no easy way to be merged. I had to choose a version and manually change the other 11 by copy-pasting.

I also had lots of tags that were used only a single time. However, I had no way of easily filtering these tags, such as a list or counter for how many times a tag is used in the library, in order to deal with them.

By searching the forums, I found instances of users mentioning they have libraries with thousands of tags that they would like to quickly delete (which may be automatic tags or not) and have to resort to editing the SQLite database, since Zotero does not offer an easy way to manage tags.

I believe that, in the same way that Zotero has a dedicated interface to manage saved searches, it would greatly benefit from an interface dedicated to tag management, with features such as selecting multiple tags for merging/removal and counting how many times each tag is used (i.e. in how many items it is employed). These few features would go a long way into solving the issues of many users, as well as making the use of tags less painful and more widespread.

Maybe a window with a list and two columns (Tag Name, N. of Uses), allowing users to sort by column, select multiple entries and then merge or delete the selected entries. The rest could be done through the functionalities Zotero already has.
  • Yes, something like this is planned.

    You may be aware, but note that you can merge tags by renaming one tag into another in the tag selector, and you can delete tags from multiple items by Cmd-dragging (macOS) or Shift-dragging (Windows/Linux) items to tags in the tag selector. Note the same as what you're asking for, but might help in certain situations.

    For your "multi-agent systems" example, the quickest option would probably be to enable "Display All Tags in This Library" in the tag selector menu, click on each of the 12 tags in turn and drag all matching items to the variant you wanted to keep (which would remain visible in gray), and then just right-click → Delete on all the other tags. (You could also do it without "Display All Tags in This Library" by using a temporary collection.)
  • I agree with the comments by douglasrizzom. A proper tag manager is needed for Zotero. I use Adobe Lightroom for cataloging my images and it has a rather sophisticated keywiording module. Keywords ( or tags) can be structured in a hierarchical format and edited etc. Something akin to the Lightroom approach would be of great benefit and enhancement to Zotero as I like to tag most of my entries.
    thomb
  • @dstillman thanks for the hints, I didn't know I could drag and Shift+drag items to tags. That trick partially solved my problem. Unfortunately, I have multiple instances of this problem, probably because I imported bib files at some point that contained many tags. That may be an issue with my bad practices and not with Zotero, but it would be good if Zotero could help me solve this issue ;-)

    I believe good ideas could come from thinking of ways that people with too many tags could simplify/summarize their tag collections without too much manual labor.

    And I second @thomb 's idea of hierarchical tags, although I admit that is a little more complicated than what I originally envisioned. I can see how some of my tags could be naturally contained inside others.
  • This would be great. The ability to merge multi-word tags would be useful; conversely, some translators don't handle delimiters well and it would be nice to be able to split some based on a particular string (e.g., my database has quite a few "A;;B;;C" tags because one site I frequently use puts double semicolons between its tags.
  • Please, please, PLEASE do this! It is the one thing that keeps me from wholeheartedly recommending Zotero to friends - the tags features are amazing, but the difficulty in batch renaming, removing, etc. is extremely frustrating!
  • edited June 16, 2020
    You cannot imagine how happy I am that you‘re developing such a feature! I think there’s lots of potential, especially for the group libraries and for data visualization.

    Here some ideas from my experience with an institutional media manager, hope it helps for the database:

    1. Tag ID

    2. Tag Color [Hexadecimal?]
    Everyone loves colors.

    3. Preferred Name [text or list/array of text]
    Actual Tag name; A text list would make easier to allow translations.

    4. Alternative Name [list of text]
    Can help to standardize tags or for translations. list because there can be more synonyms to one tag.
    Eg. "AI" is an alternative name for "Artificial Intelligence". When searching or tagging for “AI”, redirects to "Artificial Intelligence".

    5. Broader Concepts [list of tags] and
    6. Narrower Concepts [lsit of tags]
    This allows to create a Tree of Concepts.
    Every time we set A abroad B and C, we need to set B and C under A. (It's not possible that A is upon B and B upon A); Then searching can be "A" (only A) or "inside A" (A or B or C);

    7. Related Concepts [list of tags]
    Relate tags which don’t have a hierarchical (up-down) relation. "Internet Art" is part of "Art", but have a relation with "Internet".

    This can be useful for future integrations:

    8. External technical reference [text]
    For example, links to Subject Headings Authority File, maybe would work good for future integrations.

    9. Scope Note/Definition [text, little longer]
    A definition, explaining what the term means, or not: can be useful for groups.

    10. Hidden Terms [list of text]
    Make a term search-able but not visible (not sure I’ve got this on; not sure that we need it).

    11- Pseudo tags [boolean]
    If a tag is pseudo, then is not searchable or writable: it can be useful for organizing the tag-tree, but nothing else.

    I would like to help, but my coding skills are quite limited. But I'm sure there are more things to do! ;)
  • edited March 27, 2021
    Tag manager will be a great improvement, as for example, I have hundreds of tags and many are badly formatted, now global manipulation is very cumbersome. Tag manager could start with a standalone window or full height side pane, search or filter function, possibility to create/merge edited tags which user separate with delimiter (; or ,). Next features can follow as suggested before, e.g. I use tags for usability rating, it can be replaced with stars or hierarchical tag. Hierarchical tags can extend versatility significantly as number of visible collections is limited, we can use it for project labeling with categories, rating...
  • Agree with @"victor fancelli" with what is clearly a proposal to apply the SKOS ontology to this problem.

    There is only so long I can tolerate fixing up lexical variants without being able to consolidate semantic variants. Synonym control is needed! We don't have thousands of records yet in our research group but the tag cloud is dark and expanding faster than I can control it. Frustrating too because I would like our Zotero metadata to align with the tagging system we develop in Drupal. This will be difficult if we have a long tail of single-use tags in Zotero.

    Enabling hierarchical features would be a significant impact on the existing interface. I wonder if integration with third party RDF vocabulary applications might be required to achieve full thesaurus editing functionality.
  • Hello,
    There is at least another thread with similar concerns at:
    https://forums.zotero.org/discussion/84867/zotero-and-ontology-integration-scaling#latest

    This is obviously a new level of development for Zotero: "how to integrate controlled vocabularies, thesauruses and ultimately ontologies", to make Zotero even more powerful in terms of search capabilities and data structuration.
    Could Zotero become as well an ontology manager (which other open source solutions provide), or could Zotero simply be interfaced with such an ontology manager, via the "tags" indexing Zotero entries. In that case, Zotero entries would become instances of ontologies managed by the external software.
  • I second this request!
  • edited August 18, 2022
    > Could Zotero become as well an ontology manager (which other open source solutions provide)

    I realize that the message is a bit old, but could someone give specific examples? I'm curious.
  • I agree that a tag manager would be very helpful.
  • edited August 18, 2022
    If I can share this, I am considering applying the dewey decimal classification to manage my references in zotero.


    It is not about importing the whole classification, only the first main ten plus a couple of one related to my work. Could goes up to aprox. one hundred, if not more.

    It is grounded on the light-weight ability of zotero to be able to put one item in one or more than one collection, plus with tags. Note though that too many collection could alter performance. First important question then is from how many collection performance start to reduce? I have currently no answer.

    As for the tags lexicon, remains the choice whether to put a dewey number at the beginning or a the end of the tag, with '.' ';' or space separation. For example 00 Generalities or 00.generalities or generalities 00. Most consistent way would be at the beginning in my opinion. Other option is to insert a 'dd.' prefix, so to mark separation to other type of tag system you might have (though I think it add unnecessary clutter). For example work management related tags (to-do ; to-read ; etc.).

    As is it very unlikely that one could learn all the dewey classification, the workflow would be approximately the following : to tag as much as you can as you go, then put it in the corresponding collection when you clean your library (spring cleaning).

    From that being said, what I am looking for is a tagging system enabling autocompletion and permutation from the web conector. As it is the main point of entry of my item, I would like to be able to start typing let say "cooking" and have the corresponding one or more dewey number I already have. If the dewey decimal does not exist, I would add the word anyway using (tema_cooking or subject_cooking, or even just "cooking") and later on look up for a dewey decimal. Then I would create the decimal.

    Current workaround is to perform such search from the search tag field in the left down corner of the general interface. It does do auto-completion and permutation (propose academic writing if I start to write writing). But sadly, I have to write again in the add tag to item field (right side of the pane), as there is no permutation in this field. And as I cannot copy paste neither drag and drop from the left pane toward the selected item.

    Alternative to a dewey decimal system, one option would be to tag using a facet-like classification, with tags like subject_a ; place_b ; date_c ; discipline_d ; project_e and so on...

    PS: the value added for me is in particular from the 00 generality category, where I can order all my sources and perform literature and press review. Also, it comes with being able to "connect" with libraries that use dewey decimal over the world. And, you could extend you researches knowing the dewey decimals corresponding to your field search (search with dd: in some platform). Hospitality of the decimal system and guidelines has responded (partially) to the biais toward white anglo-saxon centrism and baconian overall classification of knowledge.
  • Could we have an update on the estimated delivery of a Tag Manager?
  • Is there a page/url where we can see what the priority list for development.
    Not having a decent tagging system/manager is killing me, and apparently other too, by the looks at comments over the years.
  • @bolejj

    see webpage below for progress on a component of tag management
    https://github.com/zotero/zotero/pull/2716
  • @waynerroper: No, that's something different from what's being discussed here.

    But we don't give ETAs, sorry.
  • I guess all I'm asking is... Is tag management actually being worked on in earnest, or is it a low priority item?
  • Recent update enable to find such tags (tagged anotation) from the overall left pane :) It is not yet fully searchable as we are wishing for but it is more than fair! For example I could retrieve all the text I've highlighted as "definition".

    See announcement from 30 sept 2022: https://www.zotero.org/support/changelog#changes_in_6015_september_30_2022
Sign In or Register to comment.