I have a 1.2 MB XML file -- how can I get it into TBX?

[mwra (Mark Anderson)]

Thanks for the clarification. I always worry it seems snarky to ask, but it’s so easy to hammer harder on the thing in front of one rather than step back and review assumptions: I’m as prone as anyone.

Abstracts. Tinderbox parses these into the $Abstract attribute, though that’s not ideal place to read an attribute. The action code (agent, stamp, rule, whatever) to copy the abstract to note $Text is:

$Text = $Abstract;

More efficient is to ‘move’ the abstract data so you don’t have redundant storage:

$Text = $Abstract; $Abstract =;

That voids the $Abstract attribute data after copying it to text. Perhaps you have text already? You can move the abstract to the top of the taxt and keep your existing notes:

$Text = $Abstract + "\n~~~~\n" +$Text; $Abstract =;

The "\n~~~~\n" part is just to give a clear indication of the end of the abstract.

The problem of ‘full metadata’. I don’t know (pretty sure it’s not documented) as to what Bookends sticks as payload on the Cmd+Opt+drag to Tinderbox. The data Tinderbox reads is in RIS format, so it may be the RIS format as in the Bookends format manager. Jon, the Bookends dev is very responsive and there’s also a user forum where you can ask. IOW, it might be that you can alter the payload going onto the drag/drop by altering formats at the Bookends end - or it might be a feature they can implement. I say that as Tinderbox can only read what’s there.

Tinderbox can, IIRC, read a dropped ‘.ris’ file so you could experiment by making an enriched (in terms of fields populated) RIS format and then report a file and drop that into TBX.

Or, I suspect AppleScript is another route. Both apps have AppleScript support so this might be a more efficient vector for passing info.

OK, neither Bookends format customisation syntax nor AppleScript script are that easy but with help available at both ends it might be a do-able solution. Thus if the MESH data gets to Bookends I’m pretty confident it’s possible to find a way to bring it to Tinderbox.

I trust that gives some usable ideas to move you forward.

[rkaplan (Richard Kaplan)]

Those are excellent suggestions - thanks. I will follow up.

[rkaplan (Richard Kaplan)]

Well it turns out that Bookends has configurable Import Filter setttings; by tweaking those, it is possible to get all of the MESH fields do appear - they are referred to as Keywords:

That is a bug plus. However, the PMID (PubMed ID) and the DOI (Digital Object Identifier) are two important pieces of metadata which are in the “Additional Fields” section of Bookends; those fields do not import in a Drag/Drop to Tinderbox. Can I configure that or is that fixed in the Tinderbox software?

[mwra (Mark Anderson)]

Nice when a hunch pays off. Looking at Bookends Preferences and the Formats Manager, DOI data is stored in Bookends database field User17, and PMID in User18. In the Formats Manager, the macro codes are ‘u17’ and ‘u18’ respectively. I’m not confident to suggest the exact edits - I suggest you contact Bookends’s support or user forum and get more reliable help on adding DOI and PMIS to your RIS export filter.

Whilst doing so you can check with Bookends whether editing your local apps RIS filter will load that data onto the drag payload. Don’t simply assume editing the export filter does so.

Having ascertained that the question becomes does Tinderbox handle that info. Tinderbox does have a system attribute $DOI but not PMIS data. so, my starting hunch is RIS data for DOI would be mapped but probably not PMIS. Before hitting up Eastgate on the latter, I’d first check that even if only via per-user customisation, DOI & PMIS data cab be placed on the dragged link data payload created by Bookends (Tinderbox can’t use what isn’t there!). Then if the PMIS data is there but not mapped, that is perhaps the point to ask Eastgate. Both apps’ developers liaise so this may be fixable, even if not in today’s version (of either/both app).

If the need is more pressing, AppleScript should be able to bridge the gap, albeit with “some assembly required”, e.g. writing such a script!

[rkaplan (Richard Kaplan)]

Thanks for all your help @mwra and @eastgate

Mark Bernstein pointed out that the RIS data is in $RefererenceRIS and even though it seems to be only 1 line in the Displayed Attributes, it actually is longer due to carriage returns. By adding a Stamp to add that to $Text a lot more data appears.

Right now the data is not in perfect format all in Attributes, but everything essential is in fact there from the Bookends Drag. I should be able fairly easily to use Regular Expressions and and Agent to clean it up and move everything into its own Attribute. There is certainly no shortage of ways to customize a Tinderbox project.

So it appears I am good to go to proceed to the next steps - thanks for the help.

[mwra (Mark Anderson)]

Great! Come back if you get stuck.

[rkaplan (Richard Kaplan)]

To clarify by the way- Tinderbox imports RIS data per the RIS Bibliography as set up in the Formats Manager in Bookends. This means you can take any field you wish and configure the Bibliography in whatever format you wish to make it friendly to then extract the data in Tinderbox using an Agent and put each field as its own attribute:

[rkaplan (Richard Kaplan)]

So this is the revised RIS import format (now in Tinderbox) using the Bookends Format manager - note PMID and DOI at the bottom. But it is quite flexible and can be used to add whatever other fields in the future with whatever delimiters are desired to make it easy to move the data into a Tinderbox Attribute with a Regular Expression.