Re: [eclipse-clp-users] PyCLP inconsistency in treating 0 (zero)

From: Thorsten Winterer <thorsten_winterer_at_...126...>
Date: Fri, 13 Dec 2013 17:11:07 +0100
Hi,

I think I found a fix for the 0 conversion problem in PyCLP.

In class Term, the line

    if init_arg:

needs to be changed into

    if init_arg is not None:


As for the booleans, I added code to class Term that I copied from class
Atom:

            #Booleans
            elif isinstance(init_arg,bool):
                if init_arg:
                    string = "true"
                else:
                    string = "false"
                #Convert to byte array
                py_byte_string = tobytes(string)
                c_string = py_byte_string
                # Create dictionary entry
                ptr=pyclp.ec_did(c_string,0)
                temp=pyclp.ec_atom(ptr)
                self.ref.set(temp)

A cdef definition needs to be added to the class Term before the
definition of function __init__():

    cdef char* c_string

Also, when the ECLiPSe boolean is read out in Python, it should be
transformed into a Python Boolean. For this, I changed the handling of
atoms in pword2object():

    elif pyclp.ec_get_atom(in_pword,&dummy_dident)== pyclp.PSUCCEED:
        result=Atom(None)
        (<Atom>result).set_pword(in_pword)
        if str(result) == "true":
            result = True
        elif str(result) == "false":
            result = False


With these changes, I was able to pass zeros and Python booleans to
ECLiPSe and back. I'm not sure, though, whether the handling of Booleans
is the best way - it was rather quick&dirty to see whether I can get
this to work.


Regards,
Thorsten


Am 13.12.2013 11:56, schrieb Thorsten Winterer:
> Hi,
>
> as I was fighting the very same problem this morning, I would like to
> point to another related problem:
> the Python booleans True and False are not transformed into ECLiPSe
> booleans true and false, but instead treated as integers, resulting in
> 1 and _ values, respectively.
>
> Unfortunately, these two issues are showstoppers for me, so if anyone
> has a suggestion how to fix PyCLP's Term class, I'd be very grateful...
>
> Thorsten
>
>
> Am 13.12.2013 11:45, schrieb Federico Ferri:
>> I found a problem in how pyclp builds terms for eclipse.
>>
>> Let's say I want to post the goal X is 2+3:
>>
>>   import pyclp
>>   pyclp.init()
>>   X=pyclp.Var()
>>   goal=pyclp.Compound('is', X, pyclp.Compound('+', 2, 3))
>>   print('goal: %s' % goal)
>>   goal.post_goal()
>>   ret,arg=pyclp.resume()
>>   print('X = %s' % X.value() if ret else 'goal failed')
>>
>> gives the expected output:
>>
>> goal: is(_,+(2,3))
>> X = 5
>>
>>
>> Now if I want to post the goal X is 0+3, I use the very same program
>> as above, just replacing 2 with 0, and here's the output:
>>
>> goal: is(_,+(_,3))
>> X = None
>>
>> Apparently, 0 becomes a variable, which does not look correct.
>>
>> I thought that this was my mistake, and I should have been explicitly
>> creating numbers with a pyclp constructor, but building the goal as:
>>
>>   goal=pyclp.Compound('is', X, pyclp.Compound('+', pyclp.Term(0),
>> pyclp.Term(3)))
>>
>> or as:
>>
>>   goal=pyclp.Compound('is', X, pyclp.Compound('+', pyclp.Atom('0'),
>> pyclp.Atom('3')))
>>
>> yields to the same output:
>>
>> goal: is(_,+(_,3))
>> X = None
>>
>>
>> Additionally, building the goal as:
>>
>>   goal=pyclp.Compound('is', X, pyclp.Compound('+', '0', '3'))
>>
>> yields to a different stringification:
>>
>> goal: is(_,+("0","3"))
>> X = None
>>
>> I'm not sure if "0" and "3" are truly strings, because in an
>> interactive eclipse interpreter, I get those results instead:
>>
>> [eclipse 1]: X is 0 + 3.
>> X = 3
>> Yes (0.00s cpu)
>>
>> [eclipse 2]: X is '0' + '3'.
>> calling an undefined procedure '0'(_209) in module eclipse
>> Abort
>>
>> [eclipse 3]: X is "0" + "3".
>> number expected in +("0", "3", _281)
>> Abort
>>
>>
>>
>> Some interesting facts from the python console:
>>
>> >>> str(pyclp.Atom('0'))
>> '0'
>>
>> >>> str(pyclp.PList([1,0,1,0]))
>> '[1,_,1,_]'
>>
>> >>> str(pyclp.PList(list(pyclp.Atom(str(x)) for x in [1,0,1,0])))
>> '[1,0,1,0]'
>>
>>
>> I'm a bit puzzled, I think I will file a bug report to PyCLP
>> Has anyone encountered this before?
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Rapidly troubleshoot problems before they affect your business. Most IT 
>> organizations don't have a clear picture of how application performance 
>> affects their revenue. With AppDynamics, you get 100% visibility into your 
>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
>>
>>
>> _______________________________________________
>> ECLiPSe-CLP-Users mailing list
>> ECLiPSe-CLP-Users_at_lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/eclipse-clp-users
>
>
>
> ------------------------------------------------------------------------------
> Rapidly troubleshoot problems before they affect your business. Most IT 
> organizations don't have a clear picture of how application performance 
> affects their revenue. With AppDynamics, you get 100% visibility into your 
> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
>
>
> _______________________________________________
> ECLiPSe-CLP-Users mailing list
> ECLiPSe-CLP-Users_at_lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/eclipse-clp-users
Received on Fri Dec 13 2013 - 16:11:17 CET

This archive was generated by hypermail 2.2.0 : Mon Jul 09 2018 - 02:05:30 CEST