git filter to strip the thumnails from a tableau workbook.

sample_workbook.twb

<?xml version='1.0' encoding='utf-8' ?>

<!-- build 9100.15.1013.2200                                -->
<workbook source-platform='mac' version='9.1' xmlns:user='http://www.tableausoftware.com/xml/user'>
  <preferences>
    <preference name='ui.encoding.shelf.height' value='24' />
    <preference name='ui.shelf.height' value='26' />
  </preferences>
  <datasources>
    <datasource caption='Earnings - ICT (ICT_Vetted)' inline='true' name='excel-direct.42297.702443067130' version='9.1'>
      <connection class='excel-direct' cleaning='no' compat='no' dataRefreshTime='' filename='/Users/sechilds/Dropbox/Work/Space/Tableau Tax/ICT_Vetted.xlsx' interpretationMode='0' password='' server='' username='' validate='no'>
        <relation name='&apos;Earnings - ICT$&apos;' table='[&apos;Earnings - ICT$&apos;]' type='table'>
          <columns gridOrigin='A1:H183:no:A1:H183' header='yes' outcome='6'>
            <column datatype='string' name='Graduation Year' ordinal='0' />
            <column datatype='integer' name='Years Since Graduation' ordinal='1' />
            <column datatype='string' name='ICT Flag' ordinal='2' />
            <column datatype='integer' name='Mean' ordinal='3' />
            <column datatype='integer' name='Mean Low' ordinal='4' />
            <column datatype='integer' name='Mean High' ordinal='5' />
            <column datatype='integer' name='Median' ordinal='6' />
            <column datatype='integer' name='Rounded Count' ordinal='7' />
          </columns>
        </relation>
        <metadata-records>
          <metadata-record class='column'>

script_output.txt

The root of the XML tree is a workbook tag
with the attribute {'version': '9.1', 'source-platform': 'mac'}

The children of the root are:

preferences
datasources
worksheets
windows
thumbnails

Looking at each individual thubnail.
Printing first three lines only.

thumbnail {'width': '192', 'name': 'ICT Graduates', 'height': '192'}
      iVBORw0KGgoAAAANSUhEUgAAAMAAAADACAYAAABS3GwHAAAACXBIWXMAAAsTAAALEwEAmpwY
      AAAgAElEQVR4nO2deZycVZX3v7XvWy/Ve3d6S9LZSCchEBYRgXHEGQERB0HAZWTGEXl1RmZG
      35Hx1XF0dMAR0XGBQRBl3wZEERCQkASydfalk/TeXb1Wde1Vz3LfP550gJClkt6qU883n/50
      ...

thumbnail {'width': '192', 'name': 'Faculty Groups', 'height': '192'}
      iVBORw0KGgoAAAANSUhEUgAAAMAAAADACAYAAABS3GwHAAAACXBIWXMAAAsTAAALEwEAmpwY
      AAAgAElEQVR4nOy9d3xcV5n//57eR9M00ox6b7Ysdye2Yyex00MCSVhI2MAXFlgg/HbZpZcN
      9Uv7wrLAUneBhYRkgUASElKdOLbjbsu2rN6lURlpem/33t8fsiYxTpG7Q/R+ve5LI2lm7p07
      ...

tableau_clean_thumbnails.py

#! /usr/bin/python

"""
    git filter to strip the thumnails from a tableau workbook.
    
    To install, store a copy of this file into your $PATH.
    Then you need to tell git about the filter
    
        git config --global filter.tableau.smudge "tableau_clean_thumnails.py"
    
    Then you need to specify which files to filter.
    In this case, you want to filter Tableau workbooks (.twb)
    
        echo '*.twb filter=tableau' >> .gitattributes
        
    You can learn more about git filters and .gitattributes here:
    https://git-scm.com/book/en/v2/Customizing-Git-Git-Attributes

"""

import sys
import os
from xml.etree import ElementTree

def smudge():
    #to_parse = sys.stdin.read()

    tree = ElementTree.parse(sys.stdin)

    root = tree.getroot()
    thumbnails = list(root.findall('thumbnails'))[0]
    for i in thumbnails.findall('thumbnail'):
        thumbnails.remove(i)

    tree.write(sys.stdout, encoding="unicode")

if __name__ == "__main__":
    smudge()

Tableau_Filter.md

Filter Tableau Workbook Thumbnails

This is a git filter to strip the thumnails from a tableau workbook.

To install, store a copy of this file into your $PATH. Git looks for a script to run that takes a file in as standard input and outputs a text file.

Then you need to tell git about the filter

git config --global filter.tableau.smudge "tableau_clean_thumnails.py"

Then you need to specify which files to filter. In this case, you want to filter Tableau workbooks (.twb)

echo '*.twb filter=tableau' >> .gitattributes

You can learn more about git filters and .gitattributes in the git book

The script uses the xml module from Python's standard library. It parses the XML tree of a tableau workbook and removes all the thumbnails. Then it writes the remaining XML tree to the standard output.

tableau_xml_struct.py

"""
    Simple python script to look at the top level of a
    XML file.
"""

from xml.etree import ElementTree
import sys

def xml_hierarchy(filename):
    with open(filename, 'r') as f:
        tree = ElementTree.parse(f)
    root = tree.getroot()
    print('The root of the XML tree is a {tag} tag\nwith the attribute {attrib}\n'.format(tag=root.tag, attrib=root.attrib))
    print('The children of the root are:\n')
    for child in root:
        print(child.tag)
    print('\nLooking at each individual thubnail.')
    print('Printing first three lines only.\n')
    thumbnails = list(root.findall('thumbnails'))[0]
    for tb in thumbnails.findall('thumbnail'):
        print(tb.tag, tb.attrib)
        for line in tb.text.split('\n', 5)[1:4]:
            print(line)
        print('      ...\n')

if __name__=='__main__':
    xml_hierarchy(sys.argv[1])