Example of using the rlab helper classes

the rlab package provides a simple way to create complex pdfs. It provides the following pieces.

components: provide wrappers to pdf components so they can dynamically scale themselves based on available space
table: numpy indexing to provide formatting
builder: manage the building of a pdf and defining grid based templates

Below I will show the following simplified example. For each security (MSFT, CSCO, INTC), we want to generate 2 pdf pages, one page has images and the other has tabular data.



In [1]:

    
import tia.rlab as rlab
import matplotlib.pyplot as plt
from tia.rlab.table import PercentFormatter, MillionsFormatter, mdYFormatter

Get some financial data



In [2]:

    
# get some real data to plot and build tables for
from pandas.io.data import get_data_yahoo
pxs = get_data_yahoo(['MSFT', 'INTC', 'CSCO'])

Define the builder and templates



In [3]:

    
# - Generate a pdf path
# - define a cover page
# - Create a Pdf Builder
# - Define the templates and register with the builder
#
#        TEMPLATE_1
#  |-------------------------|
#  |        HEADER           | 
#  |-------------------------|
#  |            |            |
#  |            |            |
#  |            |            |
#  |   IMG_1    |    IMG_2   |
#  |            |            |
#  |            |            |
#  |            |            |
#  |-------------------------|
#
#        TEMPLATE_2
#  |-------------------------|
#  |        HEADER           | 
#  |-------------------------|
#  |            |            |
#  |            |    TBL_2   |
#  |            |            |
#  |   TBL_1    |------------|
#  |            |            |
#  |            |    TBL_3   |
#  |            |            |
#  |-------------------------|

pdfpath = 'rlab_usage.pdf'

coverpage = rlab.CoverPage('SecurityOverview', 'Reported on Jan-20-2015')
pdf = rlab.PdfBuilder(pdfpath, coverpage=coverpage, showBoundary=0)

# Define TEMPLATE_1
template = rlab.GridTemplate('TEMPLATE_1', nrows=100, ncols=100)
# uses numpy style slicing to define the dimensions
template.define_frames({
    'HEADER': template[:10, :],
    'IMG_1': template[10:, :50],
    'IMG_2': template[10:, 50:]
})
template.register(pdf)

# Define TEMPLATE_2
template = rlab.GridTemplate('TEMPLATE_2', nrows=100, ncols=100)
template.define_frames({
    'HEADER': template[:10, :],
    'TBL_1': template[10:, :50],
    'TBL_2': template[10:55, 50:],
    'TBL_3': template[55:, 50:]
})
template.register(pdf)

Define a style and method for building a header



In [4]:

    
from reportlab.platypus.flowables import HRFlowable
from reportlab.lib.colors import HexColor
from reportlab.lib.styles import ParagraphStyle, TA_CENTER

# Add a stylesheet to the pdf
tb = ParagraphStyle('TitleBar', parent=pdf.stylesheet['Normal'], fontName='Helvetica-Bold', fontSize=14, 
                    leading=14, alignment=TA_CENTER)


'TitleBar' not in pdf.stylesheet and pdf.stylesheet.add(tb)


def title_bar(pdf, title):
    # Build a title bar for top of page
    w, t, c = '100%', 2, HexColor('#404040')
    title = '<b>{0}</b>'.format(title)    
    return [HRFlowable(width=w, thickness=t, color=c, spaceAfter=5, vAlign='MIDDLE', lineCap='square'),
            pdf.new_paragraph(title, 'TitleBar'),
            HRFlowable(width=w, thickness=t, color=c, spaceBefore=5, vAlign='MIDDLE', lineCap='square')]

Define method to convert a dataframe to a formatted pdf table object



In [6]:

    
def to_pdf_table(frame):
    table =  pdf.table_formatter(frame, inc_header=1, inc_index=1)
    # use the default style to add a bit of color
    table.apply_basic_style()    
    # apply a percent formatter to the return column
    table.cells.match_column_labels('return').apply_number_format(PercentFormatter)
    # apply a millions formatter to volumn column
    table.cells.match_column_labels('Volume').apply_number_format(MillionsFormatter)
    table.index.apply_format(mdYFormatter)
    return table.build(vAlign='MIDDLE')

Create generic method which builds a graph and tables for the specified security



In [7]:

    
# Define a matplotlib helper to store images by key
from tia.util.mplot import FigureHelper
figures = FigureHelper()

def add_security_to_report(sid, pxframe):
    # build the images
    with plt.style.context('fivethirtyeight'):
        plt.rcParams['lines.linewidth'] = 1.5
        img1_key = '{0}_open_pxs'.format(sid)
        img2_key = '{0}_close_pxs'.format(sid)
        pxframe['Open'].plot(title='{0} Open Price'.format(sid))    
        figures.savefig(key=img1_key)
        pxframe['Close'].plot(title='{0} Close Price'.format(sid))
        figures.savefig(key=img2_key)
        # build the tables
        pxframe['return'] = pxframe.Close.pct_change()
        tbl1 = to_pdf_table(pxframe[['Open', 'High', 'Low', 'Close']].tail(40))
        tbl2 = to_pdf_table(pxframe.iloc[:10])
        tbl3 = to_pdf_table(pxframe.iloc[-10:])
    
        # Marry the template with the components
        pdf.build_page('TEMPLATE_1', {
            'HEADER': title_bar(pdf, '{0} Images'.format(sid)),
            'IMG_1': rlab.DynamicPdfImage(figures[img1_key]),
            'IMG_2': rlab.DynamicPdfImage(figures[img2_key]),        
        })
    
        pdf.build_page('TEMPLATE_2', {
            'HEADER': title_bar(pdf, '{0} Tables'.format(sid)),
            'TBL_1': tbl1,
            'TBL_2': tbl2,
            'TBL_3': tbl3,
        })

For each security, add to the pdf



In [8]:

    
for sid in pxs.minor_axis:
    add_security_to_report(sid, pxs.minor_xs(sid))
pdf.save()

Focusing on function and not form (working on better default style, check table example for nicer one)

Click ME.



In [ ]: