New function to sort possible world visualizations by number of overlapping MIR #10
Description
New issue, related to dealing with under-specified input datasets. Suppose the user is exploring a consistent alignment, gradually. Early stages of providing articulations are under-specified, resulting in many possible worlds.
To avoid running Euler/X for hours at a time, the user employs commands such as:
euler2 align [filename] -n 1 [showing only 1 possible world)
euler2 align [filename] -n 10
euler2 align [filename] -n 100
euler2 align [filename] -n 1000
..sequentially, to assess the "subset MIR" and identify concept pairs in the resulting (incomplete, but informative) MIR need of (more specific) articulations.
At some point, 21 possible worlds are produced (see attached example file). That is a manageable number to look at every PW, but frequently the number of PWs is in the hundreds or thousands - even though the user is "close" to having very few or one PW(s). The "right alignment" is already entailed among the hundreds of PWs.
Typically, among a larger set of PWs, those with fewest overlapping articulations are closest to the desired outcome. In the attached example, that number of overlapping (><) articulations in the MIR ranges from 0 to 16! "0" is the desired PW. But the PWs are not ordered in any user-friendly way, and if there are hundreds or thousands - this becomes an issue.
There should therefore be an option to do the following.
- euler2 align [filename](optional: -n 200) => generates (e.g.) 150 PWs for a dataset.
- euler2 show pw => generates the 150 visualizations
- euler2 sort pw => evaluates the 150 PWs, and re-sorts and re-numbers these files by number of overlapping MIR, in ascending order. This way the user knows to look at the very early set of PWs in the output folder to find those PWs with likely the highest level of congruence, lowest number of >< articulations, and highest chance of being close to the desired, well-specified outcome.