PayPal notification email parsing

[Werner F. Bruhin]

I need to parse a PayPal email, got it mostly but have a problem with the following.

The postal address with this email can be either:

NameOfSomeone's UNCONFIRMED Address
-----------------------------------

NameOfSomeone
addressline 1
addressline 2
Postal, city line
Country

or:
NameOfSomeone's UNCONFIRMED Address
-----------------------------------

NameOfSomeone
addressline 1
Postal, city line
Country

Also note that "UNCONFIRMED" my be "CONFIRMED".

I can't make it work for both formats, fragments of my code are:

        alphanums = pyp.Word(pyp.alphanums)
        addLine = pyp.Combine((pyp.OneOrMore(alphanums) + pyp.restOfLine))
        whiteLine = pyp.FollowedBy(pyp.Suppress(pyp.White()))

        # address
        str_address = (pyp.Suppress("UNCONFIRMED Address") | pyp.Suppress("CONFIRMED Address")) + pyp.Suppress(pyp.Word("-")) +\                 addLine + addLine + addLine + addLine + whiteLine
               
        address = str_address.setParseAction(self.str_addressAction)

        self.grammar = str_buyer | str_transid | str_itemnum | str_address

If I add another "addLine" then I get garbage on the 4 line version.

Any hints on how I could do this would be very appreciated.

Werner

[Werner F. Bruhin]

I think I figured it out, with lots of try and error.

As the address is followed by at least one blank line I do this now.

        addLine = pyp.Combine((pyp.OneOrMore(alphanums) + pyp.restOfLine))
        emptyLine = pyp.LineStart() + pyp.LineEnd()

        # address
        str_address = ((pyp.Suppress("UNCONFIRMED Address") | pyp.Suppress("CONFIRMED Address")) + pyp.Suppress(pyp.Word("-")) +\                 addLine + addLine + addLine + (addLine + emptyLine.suppress() ^ (addLine + addLine + emptyLine.suppress()) ))

Werner

[Paul McGuire]

Werner -

This is certainly one way to do this.  I wonder if you would try an experiment for me though, since there is a new feature that was released in pyparsing 1.4.9, the ability to multiply an expression by an exact number of repetitions, or by a min-max tuple of repetitions.

Using the min-max tuple, I think your str_address will simplify from:

str_address = ((pyp.Suppress("UNCONFIRMED Address") | pyp.Suppress("CONFIRMED Address")) + pyp.Suppress(pyp.Word("-")) +\
addLine + addLine + addLine + (addLine + emptyLine.suppress() ^ (addLine + addLine + emptyLine.suppress()) ))

to:

str_address = ((pyp.Suppress("UNCONFIRMED Address") | pyp.Suppress("CONFIRMED Address")) + pyp.Suppress(pyp.Word("-")) +\
addLine*(4,5) + emptyLine.suppress()

(not sure, I might have a mismatched paren in there or two)

Also, please check out the Examples page of the pyparsing wiki, you may find some help in parsing street addresses.

Cheers!
-- Paul

[Werner F. Bruhin]

Paul,

Thanks for the tip, will try it out tomorrow or Monday.

I have seen the screen parsing code, but as I store the address in one block I don't need this - at least not yet.

Werner

[Werner F. Bruhin]

Paul,

Just upgraded to 1.4.11.

addLine*(4,5) doesn't work for me.

Please note that I also had to change the following code:
        emptyLine = pyp.LineStart() + pyp.LineEnd()
to:
        emptyLine = pyp.LineEnd()

To make my code work again.

If you want me to try something else, or can provide you more info just let me know.

Werner