Nice to meet you all. I am Jacco, a newbie user(bought the software last monday) of the great Tinderbox software. Working as an IT-Business Analist, I will be using the software with my analyses. Have searched on the forum, but havent found a solution of the thing i want to achieve with tinderbox. I will elaborate more on this in the following sections. Would appreciate your thoughts and solutions
Context:
I am exploring the subject of knowledge graphs. Specifically the Labeled Property Knowledge Graph subject. I can do a lot already with the current capabilities, but not all. Which i will explain in the following sections.
Goal:
Ability to export (part of) the knowledge graphs in PDF for further analyses. Preferably including all metadata visualized with it(option to show/not show). This for discussing the graphs with my co-workers.
Ability to export a part or entire knowledge graph in JSON including the relationships/edges (including properties) and nodes (including properties). This for using it in Information Systems(importing in Neo4J database) ->using tinderbox as a design tool.
Current supposed match between my goal and the software (perhaps I am wrong :
It is possible to print the entire graph area to PDF with a maximum range of A3. Especially with larger maps, there will be more pages and there will be a risk of loosing information because of the overlap. Could be mitigated by moving the nodes and do trial and error, but this will be cumbersome, because you canât see the borders of the A3 area. Furthermore you canât print the properties of it at the moment?
It is not possible to export to JSON. It is possible to export to Outline (in the format of RTF, Text or Word) or Text(in the form of RTF, Word, Text, OPML or scrivener) but these lack the relationships (and their properties) and the nodes and their properties?
I am looking forward to the possibilities which you see to accomplish my goals
JSON export is straightforward, actually, though youâll need to know what JSON you want to export. Once that is done, simply create an HTML export template. What Tinderbox calls âHTML Exportâ can handle just about any export format you want.
Copy View As Image will put a pdf version of map view on the clipboard, which can then be pasted into the editor of your choice. (PDF doesnât handle translucent objects well, so this creates some minor discrepancies.). There has been some discussion of that here on the forum in a current thread.
Very interesting! Thatâs an aspect of Tinderbox Iâm trying to use as well. But one must consider: Tinderbox is not a vector database. To utilize the significance of the edges (links) between the nodes (notes), Iâve chosen an approach that I explain in more detail here:
As @eastgate already mentioned, you can export to multiple formats using the HTML export. Although itâs called HTML export, in reality, you decide what and how Tinderbox exports. You just have to do everything yourself. You could just as well export the data as CSV and import it into Gephi, for example, to analyze it more thoroughly there.
Thanks for your swift reply :-). I will tinker some more. Can you direct me how to find the sources of the edges/relationship metadata like (labelname, direction etc etc)? for the node i can reverse engineer this by the properties button to the right of the nodes, but how do i do this for the labels/edges?
OK, I think youâre inadvertantly missing the broader picture. You come from a concept map perspective so assume Tinderbox (presumably due to map view) is the same. I suspect not. Under everything, Tinderbox is a tool for making hypertextâinterlinked nodes. Only nodes are first-class objects so if you want link properties they arenât there. also , it isnât âjustâ a matter of âadding some propertiesâ. No software is un-opinionated, and Tinderbox isnâtâby design intent a niche native concept mapping tool. If you need properties upon links you are likely using the wrong toolâat least as Tinderbox exists nowâbecause it is more than adding properties. the whole app needs re-jigging to 'think â in such terms. It could be done, but is there an obvious ROI in this? I donât know (only the dev can tell that).
Taking a more productive tack, what is Tinderbox offering that you current concept mapping tool canât do? this might make it easier to triage why the app isnât doing what you want and how/if that can be resolved.
Thereâs no negative intent in the last. In short, just because something looks like [thing] doesnât mean it is [thing] or that we should therefore assume it must be designed to do [thing]. the real question is how much of [thing] can it do.
Lasty thought. for the model you describe, Iâd ignore Map view and start with Hyperbolic view.
Iâm not sure if I understand you correctly, but itâs as @mwra writes: links in Tinderbox are indeed not much more than links and essentially cannot store or have additional information or attributes. Yes, you can save a comment and for HTML export, you can also include information about class, title, target, and the link description, and the link line can be influenced as well â but thatâs essentially it. Notes, on the other hand, offer many more possibilities, and thatâs the reason why I ultimately store information about the link in a separate note (I call these notes: link-note).
Thanks for your clarification , my question was how to unlock the current metadata of the link(not properly formulated by me, sorry for that). This was explained in your linked article. Like the concept of using a linked note to address the properties of the link. This way it can be possible to design the labeled property graph and use it in target systems like Neo4J. Still need to learn tinderbox basics some more, but will get there eventually
Summarized, it works like this:
You can store data in a limited way âonâ a link (e.g., a comment or the âTargetâ field) and read these with the command âeachLink()â. The data read can be stored in an attribute of the data type âDictionaryâ. Once stored there, you can access the values.
Thatâs the principle. You can read about the specific implementation behind the link above.
However, I should not fail to mention that the values âTargetâ, âTitleâ, and âClassâ actually serve a function in HTML export, and their improper use could cause problems if you need these fields for HTML export. @mwra has also pointed this out proactively.
@mwra´s page acrobatfaq.comA Tinderbox Reference File is a very good documentation for the functionality of the individual tools in Tinderbox. It describes a lot in great detail, but also very abstractly. However, once youâve worked through it, the knot unravels.
Thanks for this clarification . Using the tool for labeled property graphs design is one of the ways i will use the software.(forefront of innovative way of generating and using knowledge as a data fabric) Will use the tool also for more âregularâ business analysis and using concept mapping . Thanks for mentioning the hyperbolic view.
Iâve used Tinderboxâs export to make Gephi projects. I generally 2 exports, one for a Gephi nodes table and one for the edges table: both as tab-delim text. I find tab-delim less error prone if using free-text fields (stray commas are more likely than stray tabs in the text).
For example, this ACM Hypertext in-conference Citations 1987-2020 page was generated from Gephi (using the sigma.js export plugin) using data imported from Tinderbox-generated export files.
FWIW, all 3 demos as listed here use data from the same source TBX, albeit using different exports (only one of the demos used Gephi). This shows Tinderboxâs power in generating a range of outputs. The user makes all the export templates and necessary attributes for the data. It might seem like extra work, but more choice demands more configuration and at some point the app canât offer every possible choice.
:
, my question was how to unlock the current metadata of the link(not properly formulated by me, sorry for that). This was explained in your linked article. Like the concept of using a linked note to address the properties of the link. This way it can be possible to design the labeled property graph and use it in target systems like Neo4J. Still need to learn tinderbox basics some more, but will get there eventually 