Post Post (not Post Script :) Using the standard TOC method, you could build a hierarchy of primitives, new here for a sub primitive, If you pick any leaf in this hierarchy and use Bimlas's kin operator (Coming for TW 5.1.18) you could use the strait line of tiddlers from the root tiddler to the current leaf, as the list of primitives for a given query. I would use a TOC Tabbed Internal for this. Imagine if you clicked on each node/leaf in the hierarchy you would see the list of tiddlers (or not) that that meet your compound search defined as from the root.
Noting here that each node would be named according to its "filter nature" with fields containing the translation to tiddlywiki filter code eg "tagged fred" means +[tag[fred]] Inspiring. Tony On Friday, November 30, 2018 at 5:05:51 PM UTC+11, TonyM wrote: > > h0p3/Swicky > > It sounds like a great challenge, but I am sure I do not understand all > the details. A few acronyms have me a little confused. Never the less I > think I get the Gist. > > To be clear are you searching a large collection of tiddlers containing > text, tags and fields that you want to list as the result of a > sophisticated compound search consisting of multiple filters? > > I do feel that TiddlyWikis filters are in someways specific to the > TiddlyWiki's platform requirements. Rather than let this complexity > overwhelm, If I understand what you are trying to do, I would build a set > of search primitives from TiddlyWiki filters that generate a set of > tiddlers, then add these abstracted primitives together rather than drop > down into the TiddlyWiki filters syntax, in your large compound "search > criteria". > > Designed carefully every search primitive could be invoked regardless of > what proceeds or follows it. In which case you could specify the primitives > and their order by tagging a tiddler lets call it a "Query tiddler" with > the set of primitives, and the desired order (using the tag pill drag and > drop ordering) and have the result "drop out at the end" or Not. This would > allow you to look at the intermediate results ie Primative1 generates list > A, Primative2 to uses this list to generate list B etc... > > I feel a form of loop could be built that takes as its input the previous > list generated by the previous primitive. If constructed correctly each > step will result in a much smaller subset to test/filter and as a result > the performance should be good. > > In keeping with the original post I would then use the resulting list to > transclude the content. > > Perhaps I am way off here, but this is my brave attempt to respond to what > looks like a deeply conceptual experiment, the type I love, but are best > constructed in my own mind, rather than trying to understand prior work. > > > Regards > tony > > > On Friday, November 30, 2018 at 3:28:59 PM UTC+11, h0p3 wrote: >> >> Hey TonyM, >> >> You are always gracious. Thank you. My 12yo daughter didn't see your post >> until it was bedtime, and I don't want to leave you hanging. She is trying >> to build a tool for me as part of her CS subject in school. She's trying to >> make a GUI search tool which can construct arbitrarily long filters. We're >> still brainstorming and just trying to make a prototype. The gist of it can >> be found here: >> >> >> https://philosopher.life/#Search%20of%20h0p3:%5B%5BSearch%20of%20h0p3%5D%5D >> >> The goal is to have a graphical tool for incrementally constructing an >> arbitrary number of atomic and complex sentences (built from atomic and >> other complex sentences using drop-down menus that include the previously >> constructed sentences). The final complex sentence will contain the main >> connective and will be the final filter for the search. Writing several >> complex filters back-to-back is labor intensive (some people who use my >> wiki cannot write them at all, but they know enough First Order Logic they >> could use a graphical tool), and I want to abstract away as much as I can. >> >> Performance matters to me here too. My wiki is fairly large (now pushing >> ~22MB of pure text in ~7k tiddlers), and I think it will only continue to >> grow year after year. There are certain kinds of filter-based tools which >> do not work on my wiki; it's simply too big. Presumably, this tool will >> also be used to wrap other search tools which I otherwise couldn't afford >> to use without complex filters first narrowing down the search space. >> >> There may be multiple ways to do this. We may even be wrong about what is >> possible or worthwhile here. We could certainly benefit from your >> expertise. Please advise us. What are your thoughts and what direction do >> you think we should go? >> >> Sincerely, >> >> h0p3 >> >> >> On Thursday, November 29, 2018 at 5:54:44 PM UTC-5, TonyM wrote: >>> >>> Swicky, >>> >>> I am sure it would be possible to apply an unknown number of filters >>> in a filter search. >>> >>> I think there would be more value in explaining what you want to do, >>> than asking questions halfway through your current attempt to answer these >>> questions. >>> >>> There are a number of ways to concatenate text of any type into a >>> variable which you can then use in a filter and other places. Of course the >>> order is important in filters. >>> >>> Similarly you can use nested <$list widgets to separate the "problem" >>> into a number of filter statements rather than trying to force many into >>> one. The trick is to use the variable=variablename parameter of one list >>> widget as the "input" "[<variablename>... to the next List widget, unless >>> you always refer to current Tiddler and do not use the variable parameter. >>> >>> The reality however is, what are these filter meant to achieve?, because >>> once experienced with filter's there is a lot that can be done with "smart" >>> filters". Perhaps you do not need to concatenate these filters as you >>> think. >>> >>> Why are your filters unknown in number and what are they trying to find. >>> >>> Regards >>> Tony >>> >>> >>> >>> >>> On Friday, November 30, 2018 at 6:19:44 AM UTC+11, Swicky wrote: >>>> >>>> I'm trying to use an unknown number of filters in a filter search. I >>>> can, so far, hardcode transclusions of two or more tiddlers. So, if >>>> tiddler >>>> A has [all[]] in it, and tiddler B has +[tag[asdf]] in it, >>>> transcluding into filter search shows everything that's tagged asdf.. >>>> but I want to do this for any number of filters. >>>> >>>> At the moment, I'm thinking of two ideas, but I'm not quite sure about >>>> them: >>>> >>>> - Use tagged tiddlers - I could make something like a "new journal >>>> here" button, which would make a new tiddler tagged "filter" (or >>>> something). Then, just transclude the text field of every tiddler >>>> tagged >>>> "filter". >>>> - Use a field - I could also make a completely different "new" >>>> button, that adds a filter to a field, and transclude the entire field. >>>> >>>> Of course, I might be thinking about this entirely wrong. Maybe >>>> Tiddlywiki is the wrong tool for this job. If it is, then I need something >>>> else that can work in the wiki. Any ideas? >>>> >>> -- You received this message because you are subscribed to the Google Groups "TiddlyWiki" group. To unsubscribe from this group and stop receiving emails from it, send an email to tiddlywiki+unsubscr...@googlegroups.com. To post to this group, send email to tiddlywiki@googlegroups.com. Visit this group at https://groups.google.com/group/tiddlywiki. To view this discussion on the web visit https://groups.google.com/d/msgid/tiddlywiki/5e556450-ede1-4809-aade-8a99a471961f%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
