Explode with action code

Is there way to explode with the action code, or can it only be done through the popup?

Thus far, Explode is a manually invoked action. The general view has been that an action-code invoked Explode offers undue scope for unintended outcomes and/or starting large efforts (exploding 1,000s of items can take a while).

Still, there are lots of other ways to plug that gap: Keyboard Maestro action (or some similar automation utility app), or an AppleScript to essentially push the buttons for you. That ought to work well , especially if you don’t need to configure the Explode parameters (i.e. the pop-up’s controls) each time you run the task.

1 Like

Keyboard Maestro would be fine but I am unsure how I would do that for a set of Notes within a Tbx document. Automating the process for one note would be workable, but I am not sure how to do that for multiple notes except via what would amount to fairly fragile GUI scripting.

The same applies to Applescript. I do not see any entry for Explode in the Applescript Dictionary for Tinderbox. So presumably it would require GUI scripting, which is not ideal.

Yes, that is correct.

Why not merge the notes and explode them all. What is the background process driving this?

As I mentioned in another thread recently Explode via Action Code? my use case involves working with medical literature citations and the concept would likely apply to most other academic disciplines as well.

Merging the notes and then exploding them would lose the context of parent-child relationships. I would love to explode an academic citation into child notes which represent “cited by” references. Then perhaps do that recursively for a few generations depending on the volume of notes that creates.

Yes I understand your point that this can easily get out of hand and overwhelm the capability of Tinderbox by creating huge numbers of notes. But that’s really not a big deal - just backup your tbx file before you do the programmatic explode (something which should be routine practice when testing any new Action code); then if it hangs up, do a Force quit and revert to the backup. No harm in the end.

It seems to me this feature could be of use to lots of users with minimal downside. If it is challenging technically to implement it, then I understand it may not be a priority. But if implementing it is straightforward and the major reason to not do it is simply the risk of overloading Tinderbox, that seems to me to be an easily managed risk with much overriding benefit.

I’d suggest mailing in a feature request with the detail of your needs, as this is just user to user chat.

Personally, given the generally appalling quality of most citation detail, I’m not sure I’d bother with automating reference extraction, as the so the time spend cleaning the data one might more quickly create clear data by hand. That said, maybe medical references are atypically accurate.

Separately. as you acknowledge, collecting all citations can lead to bloat (without necessarily adding to the utility of the TBX’s content. From practical experience, I suspect most sources will need different explode handling which adds complexity and likely undermines the ROI of automating this. Again, I’m quite happy to be wrong. My perspective is drawn partly from my own academic research work and partly a previous career doing data management. The latter taught me that most of us are way too trusting of input data quality (i.e. as a self-excuse for not checking input in detail!).

All good thoughts.

@eastgate responded to the prior post so he seems to be aware of the request.

Medical references are not particularly accurate compared with other academic fields, but if my goal is to write a curated “roundup” or review article on a given topic, then a large number of “cited by” references is a clue that the cited article may be particularly pertinent. Moreover checking “cited by” references is good practice when using the medical literature for any purpose because it is one way to check whether there are newer articles which render the original obsolete or present rebuttals to the original article.

As for the practical details of handling the exploding, almost all relevant medical journals are included in PubMed; PubMed lets you add a parameter to a citation URL which results in a list of the cited-by references; the format used is predictable and easily allows for import into Tinderbox via $URL and subsequent processing by Regex action code.

Google Scholar could serve a similar function to PubMed for other academic disciplines. The output format is not quite as structured as with PubMed so the data might require some additional manual curating, but it certainly would be useful for the same reasons it would be useful in medicine.

1 Like

I certainly use Google scholar a lot as my work isn’t single-domain. Things like Web Of Science and other aggregations are—for me—a consistent disappointment.

One thing you might be able to do is pull in the counts without all the attendant detail which is still there at the end of the URL for when/if it it needed.

Storing extra references in-doc for easy ad hoc access means to want to take care to identify them as a discrete strand of data, so all agents aren’t searching 000s/0000s of reference notes when not needed as at scale all those un-deeded tests mount up. If they all use the Reference prototype, for example, then starting your query:

$Prototype!="Reference" & ...

Will then exclude reference notes from the subsequent parts of the query.

Storing extra references in-doc for easy ad hoc access means to want to take care to identify them as a discrete strand of data, so all agents aren’t searching 000s/0000s of reference notes when not needed as at scale all those un-deeded tests mount up. If they all use the Reference prototype, for example, then starting your query:

Agreed - obviously different users may have different workflows where this could be helpful. As for me, I am viewing Tinderbox as an “authoring” system or programmable word processor. I would not plan to keep the exploded data lurking around for future use but rather as an easy way for me to import the set of data which I am trying to curate and likely efficiently delete the extraneous entries.

The big-picture project I am working on is figuring out a way to manage the volume of medical citations which is frankly so large that it is hard for even an experienced physician or researcher or a literate patient to go to PubMed and get a credible answer or point/counterpoint analysis on almost any topic. I often get questions from friends/colleagues/patients trying to sort out conflicting recommendations in medicine; the problem is you can go to PubMed on just about any topic in medicine and get a dozen articles supporting some treatment or theory and a dozen refuting it.

So what I am trying to build is a system that works well to sort through citations on any given topic and from that write a response which is a curated set of hyperlinks showing the most credible views on all sides of a given question. I believe Tinderbox is ideal for this purpose.

A totally separate question is how to search/archive such summaries or roundup/review articles if I accumulate a number of them on various topics or if other authors contribute similar content. That becomes more of a database question and I don’t think Tinderbox scales well for that. My gut feeling so far is that MadCap Flare would work well in that phase but first I want to get the Tinderbox authoring part working.

1 Like