Adding -prompt flag to WebCCG to allow tools like Python's pexpect to parse it more easily.
Specifying the prompt flag wccg -prompt ">>>" GRAMMAR will case wccg to show the prompt expecting input like this:
>>>
Entering a sentence looks thus:
>>> she sings a song
By default, the old behavior is kept stable (no prompt sign at all).
I use this feature to parse the output using replwrap, to keep a process alive for a longer time (before I used Popen.communicate, which is a one-off command and forced me to create new wccg processes for each parse).
To keep the diff small, I tried to keep the whitespace as it is in the surroundings, but the file mixes tabs and spaces a lot and it seems in lines 85 and 120/121 I somehow missed the conventions a little bit, as I just realized.
If you want, I can update my lines or reindent the file completely.
Adding -prompt flag to WebCCG to allow tools like Python's pexpect to parse it more easily.
Specifying the prompt flag
wccg -prompt ">>>" GRAMMARwill case wccg to show the prompt expecting input like this:Entering a sentence looks thus:
By default, the old behavior is kept stable (no prompt sign at all).
I use this feature to parse the output using replwrap, to keep a process alive for a longer time (before I used Popen.communicate, which is a one-off command and forced me to create new wccg processes for each parse).
To keep the diff small, I tried to keep the whitespace as it is in the surroundings, but the file mixes tabs and spaces a lot and it seems in lines 85 and 120/121 I somehow missed the conventions a little bit, as I just realized. If you want, I can update my lines or reindent the file completely.