Question

See below matrix data:

    A   B   C   D   E   F   G
1   89  92  18  7   90  35  60
2   62  60  90  91  38  30  50
3   59  91  98  81  67  88  70
4   20  28  31  9   91  6   18
5   80  27  66  1   33  91  18
6   82  30  47  8   39  22  32
7   14  11  70  39  18  10  56
8   98  95  84  47  28  62  99

I need to define “rule” function can return “true” or “false” for below asserts string for each row data:

A=B and B=C
A>C
B>C
C>D and D<E or D>F
A+B<30
A+B<=30                # this may using "A+B<30 or A+B=30" as alternative
str(A) march regex"[2-5][0-2]" 
myfoo(A) > 100 
A in myfoo(B)
A not_in $listname
A in $listname

Take “A=B and B=C” for example: If I pass row 1 into this rule, I want the rule return false because it is not right in this case.

My questions are:

1. How can I define a DSL parser to translater those “rules string” into a workable lambda function, then I can invoke this lambda with data row as parameter to return the assert result?

2. I noticed the module operation has a lot of math functions I can reuse to define the rule, can I create a “mapper” for those keywords for DSL parser use?
   It may looks like:

   keywords = {“+”: operation.add(),
   “/”: operation.div(),
   “and”: my_and_define()
   }

3. if above 2 are possible, how can I process the “A in $listname” in the lambda and in the mapper?

Thanks for your help.

Rgs,

KC

Answer 1

Like this.

class Rule( object ):
    def __init__( self, text ):
        self.text= text
    def test( self, A, B, C, D, E, F, G ):
        return eval( self.text )

r1= Rule( "A==B" )
r2= Rule( "A==B and B==C" )
r3= Rule( "A in {listname!s}".format( listname=someList ) )

etc.

>>> r1.test( 89,  92,  18,  7,   90,  35, 60 )
False

---

Edit.

• str(A) march regex”[2-5][0-2]”
• myfoo(A) > 100
• A is in myfoo(B)

These are all trivial Python code. I’m not sure why the comment is even included as being interesting or difficult.

r4= Rule( "re.match( r'[2-5][0-2]', str(A) )" )
r5= Rule( "myfoo(A) > 100" )
r6= Rule( "A in myfoo(B)" )

There’s a trick to this. The trick is to write the Python code; and then enclose the code in quotes. Any Python code is legal.

If the Python aspect of these rules is confusion, a Python tutorial may be helpful.