Use Case: XML Parsing With Python

One of the common tasks I am given in my day job is to take some data format input and parse it to create a report or some other document. Today we'll look at taking some XML input, parsing it with the Python programming language and then creating a letter in PDF format using Reportlab, a 3rd party package for Python. Let's say my company receives an order for three items that I need to fulfill. The XML for that could look like the following code:

<?xml version="1.0"?>
<invoice>
    <order_number>456789</order_number>
	<customer_id>789654</customer_id>
	<address1>John Doe</address1>
	<address2>123 Dickens Road</address2>
	<address3>Johnston, IA 55555</address3>
	<address4/>
	<order_items>
		<item>
			<id>11123</id>
			<name>Expo Dry Erase Pen</name>
			<price>1.99</price>
			<quantity>5</quantity>
		</item>
		<item>
			<id>22245</id>
			<name>Cisco IP Phone 7942</name>
			<price>300</price>
			<quantity>1</quantity>
		</item>
		<item>
			<id>33378</id>
			<name>Waste Basket</name>
			<price>9.99</price>
			<quantity>1</quantity>
		</item>
	</order_items>
</invoice>

Save the code above as order.xml. Now I just need to write a parser and PDF generator script in Python. You can use Python builtin XML parsing libraries which include SAX, minidom or ElementTree or you can go out and download one of the many external packages for XML parsing. My favorite is lxml which includes a version of ElementTree as well as a really nice piece of code that they call "objectify". This latter piece will basically take XML and turn it into a dot notation Python object. I'll be using it to do our parsing because it is so straight-forward, easy to implement and understand. As stated earlier, I'll be using Reportlab to do the PDF creation piece.  Here's a simple script that will do everything we need:

from decimal import Decimal
from lxml import etree, objectify

from reportlab.lib import colors
from reportlab.lib.pagesizes import letter
from reportlab.lib.styles import getSampleStyleSheet
from reportlab.lib.units import inch, mm
from reportlab.pdfgen import canvas
from reportlab.platypus import Paragraph, Table, TableStyle

########################################################################
class PDFOrder(object):
    """"""

    #----------------------------------------------------------------------
    def __init__(self, xml_file, pdf_file):
        """Constructor"""
        self.xml_file = xml_file
        self.pdf_file = pdf_file
        
        self.xml_obj = self.getXMLObject()
        
    #----------------------------------------------------------------------
    def coord(self, x, y, unit=1):
        """
        # http://stackoverflow.com/questions/4726011/wrap-text-in-a-table-reportlab
        Helper class to help position flowables in Canvas objects
        """
        x, y = x * unit, self.height -  y * unit
        return x, y  
        
    #----------------------------------------------------------------------
    def createPDF(self):
        """
        Create a PDF based on the XML data
        """
        self.canvas = canvas.Canvas(self.pdf_file, pagesize=letter)
        width, self.height = letter
        styles = getSampleStyleSheet()
        xml = self.xml_obj
        
        address = """ <font size="9">
        SHIP TO:<br/>
        <br/>
        %s<br/>
        %s<br/>
        %s<br/>
        %s<br/>
        </font>
        """ % (xml.address1, xml.address2, xml.address3, xml.address4)
        p = Paragraph(address, styles["Normal"])
        p.wrapOn(self.canvas, width, self.height)
        p.drawOn(self.canvas, *self.coord(18, 40, mm))
        
        order_number = '<font size="14"><b>Order #%s </b></font>' % xml.order_number
        p = Paragraph(order_number, styles["Normal"])
        p.wrapOn(self.canvas, width, self.height)
        p.drawOn(self.canvas, *self.coord(18, 50, mm))
        
        data = []
        data.append(["Item ID", "Name", "Price", "Quantity", "Total"])
        grand_total = 0
        for item in xml.order_items.iterchildren():
            row = []
            row.append(item.id)
            row.append(item.name)
            row.append(item.price)
            row.append(item.quantity)
            total = Decimal(str(item.price)) * Decimal(str(item.quantity))
            row.append(str(total))
            grand_total += total
            data.append(row)
        data.append(["", "", "", "Grand Total:", grand_total])
        t = Table(data, 1.5 * inch)
        t.setStyle(TableStyle([
            ('INNERGRID', (0,0), (-1,-1), 0.25, colors.black),
            ('BOX', (0,0), (-1,-1), 0.25, colors.black)
        ]))
        t.wrapOn(self.canvas, width, self.height)
        t.drawOn(self.canvas, *self.coord(18, 85, mm))
        
        txt = "Thank you for your business!"
        p = Paragraph(txt, styles["Normal"])
        p.wrapOn(self.canvas, width, self.height)
        p.drawOn(self.canvas, *self.coord(18, 95, mm))
        
    #----------------------------------------------------------------------
    def getXMLObject(self):
        """
        Open the XML document and return an lxml XML document
        """
        with open(self.xml_file) as f:
            xml = f.read()
        return objectify.fromstring(xml)
    
    #----------------------------------------------------------------------
    def savePDF(self):
        """
        Save the PDF to disk
        """
        self.canvas.save()
    
#----------------------------------------------------------------------
if __name__ == "__main__":
    xml = "order.xml"
    pdf = "letter.pdf"
    doc = PDFOrder(xml, pdf)
    doc.createPDF()
    doc.savePDF()
	

 Let's take a couple minutes to go over this code. First off is a bunch fo imports. This just sets up our environment with the needed compents from Reportlab and lxml. I also import the decimal module as I will be adding amounts and it is much more accurate for float mathematics than just using normal Python math. Next we create our PDFOrder class which accepts two arguments: an xml file and a pdf file path. In our initialization method, we create a couple class properties, read the XML file and return an XML object. The coord method is for positioning Reportlab flowables, which are dynamic objects with the ability to split across pages and accept various styles. The createPDF method is the meat of the program. The canvas object is used to create our PDF and "draw" on it. I set it up to be letter sized and I also grab a default stylesheet.
Next I create a shipping address and position it near the top of the page, 18mm from the left and 40mm from the top. After that, I create and place the Order Number. Finally, I iterate over the items in the order and place them in a nested list, which is then placed in Reportlab's Table flowable. Finally, I position the table and pass it some styles to give it a border and an inner grid. Lastly, we save the file to disk.  The document is created and I've now got a nice prototype to show my colleagues. At this point, all I need to do is tweak the look and feel of the document by passing in different styles for the text (i.e. bold, italic, font size) or changing the layout a bit. This is usually up to management or the client, so you'll have to wait and see what they want.  Now you know how to parse an XML document in Python and create a PDF from the parsed data.