A stand-alone, web or FIS Viewer & Modeler

[joewheaton]

My hope is that we eventually release a simple, easy-to-use command line, web-app, and stand-alone that runs either a *.fis or *.py specification of an FIS and runs it on our own C# library. See #174 for explanation of difference.
The tool would allow:

• User to specify in a form their own FIS (i.e. membership functions for inputs, outputs and rule tables)
• Load and or modify their own *.fis or *.py version of these FIS models
• Visualize the output (similar to Fuzzy Logic Toolbox's Rule Viewer ruleview)
• Visualize the inputs (e.g. plotmf
• Evaluate an FIS for a specific set of inputs (i.e. evalfis
• Allow a user to load an FIS, and then do a vector-based or raster-based implementation:
  • In Vector Based:
    • User would load one shapefile (could be point, polygon or polyline), in which the required FIS inputs all exist as attribute fields in shapefile.
    • When the user runs the FIS, a new shapefile, with same geometry and an additional FIS-Output attribute field get added to the table and the model has been run on every feature (i.e. row) in the feature class.
  • In Raster Based
    • User would be prompted to load concurrent rasters (using same checks we have built into GCD's load survey) for each input (first input would define the 'master raster').
    • FIS would be run for every non nodata overlapping cell (i.e. in which all input rasters have a value for that cell), and spit out a new raster that represents that output.
• The whole thing would be packaged in a simple riverscapes project, that packages up inputs, the FIS models (as both scikit *.py and as Matlab *.fis), the parameters and the output.

This is a rainy-day ticket.

[wally-mac]

This seems really useful. If Braden works full time this summer this might be something we task him with.

[MattReimer]

I may have to create something like this just to figure out why FIS isn't working. If I don't get to it I'll let you know.