ModelInference / texada

Efficient techniques for mining LTL formulae over linear input traces

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Remove dependency on SPOT code in spot-2.x

carolemieux opened this issue · comments

Issue #31 was solved in ed0f430, adding a version of Texada that supports SPOT-2.X.

Two key issues:

  • The code in ed0f430 includes code from SPOT, making that code GPL-3
  • SPOT-2.X removed the capacity to print formulas in native SPOT format, so there is a change to the output of Texada (G will be written as [], for instance)

Ideally, we should remove this dependency on code pulled from SPOT, either via:

  1. Refactoring the code in texada-src/src/formula to remove code pulled from SPOT. We should also update texada::ltl to remove all the types of formulas that are not supported by Texada (BunOp, AutomatOp, etc.)
  2. Entirely removing the SPOT dependency (i.e., building out own LTL parser).

In the end, 2. may be the better option. As far as I can tell the only real functional dependency on SPOT beyond parsing/printing is the negative_normal_form() function used in map_checker.