steynvl / inferrer Goto Github PK

Hi, I was using your RPNI implementation in a project I'm working on, and I notice that it learns an incomplete DFA. Here's an example:

In a complete DFA, every transition should be defined in every state. But the last state is undefined for each transition. I understand that, implicitly, this state transitions to a sink state on each input. That's all well and good, but for my use case, it would be really nice if I could have the explicit, complete DFA. Is there a simple switch I can flip to make this happen? Thanks!

For the sample below :
%22+
Math
.ran
dom(
)+%2
233e
call: python /usr/local/bin/inferrer/cli.py /tmp/params/rnd /tmp/no rpni
Gives the result of:
(t|r|o|n|m|h|e|d|a|M|3|2|.|+|)|(|%)*

While for the sample below, only extends the last item:
%22+
Math
.ran
dom(
)+%2
233eab
call: python /usr/local/bin/inferrer/cli.py /tmp/params/rnd /tmp/no rpni
---->go into endless processing

From top, the process of "python" takes 100% cpu.

Hi, I'm trying to run the nlstar algorithm on the examples on the /resources folder but the process won't terminate. I don't know if I'm doing it incorrectly or if the process is just meant to take a while. Support will be greatly appreciated.

python cli.py resources/positive_01.txt resources/negative_01.txt nlstar --show-dfa