kicad/eeschema/plugins/python_scripts/bom_html_with_advanced_grou...

132 lines
4.3 KiB
Python

#
# Example python script to generate a BOM from a KiCad generic netlist
#
# Example: Sorted and Grouped HTML BOM with advanced grouping
#
"""
@package
Output: HTML
Grouped By: Value, Part, Footprint, Tolerance, Manufacturer, Voltage
Sorted By: Ref
Fields: Ref, Qnty, Value, Part, Footprint, Description, Vendor
Command line:
python "pathToFile/bom_with_advanced_grouping.py" "%I" "%O.html"
"""
from __future__ import print_function
# Import the KiCad python helper module and the csv formatter
import kicad_netlist_reader
import sys
# Start with a basic html template
html = """
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<title>KiCad BOM Example 5</title>
</head>
<body>
<h1><!--SOURCE--></h1>
<p><!--DATE--></p>
<p><!--TOOL--></p>
<p><!--COMPCOUNT--></p>
<table>
<!--TABLEROW-->
</table>
</body>
</html>
"""
def myEqu(self, other):
"""myEqu is a more advanced equivalence function for components which is
used by component grouping. Normal operation is to group components based
on their Value and Footprint.
In this example of a more advanced equivalency operator we also compare the
custom fields Voltage, Tolerance and Manufacturer as well as the assigned
footprint. If these fields are not used in some parts they will simply be
ignored (they will match as both will be empty strings).
"""
result = True
if self.getValue() != other.getValue():
result = False
elif self.getPartName() != other.getPartName():
result = False
elif self.getFootprint() != other.getFootprint():
result = False
elif self.getField("Tolerance") != other.getField("Tolerance"):
result = False
elif self.getField("Manufacturer") != other.getField("Manufacturer"):
result = False
elif self.getField("Voltage") != other.getField("Voltage"):
result = False
return result
# Override the component equivalence operator - it is important to do this
# before loading the netlist, otherwise all components will have the original
# equivalency operator.
kicad_netlist_reader.comp.__eq__ = myEqu
# Generate an instance of a generic netlist, and load the netlist tree from
# <file>.tmp. If the file doesn't exist, execution will stop
net = kicad_netlist_reader.netlist(sys.argv[1])
# Open a file to write to, if the file cannot be opened output to stdout
# instead
try:
f = open(sys.argv[2], 'w')
except IOError:
e = "Can't open output file for writing: " + sys.argv[2]
print(__file__, ":", e, file=sys.stderr)
f = sys.stdout
# Output a set of rows for a header providing general information
html = html.replace('<!--SOURCE-->', net.getSource())
html = html.replace('<!--DATE-->', net.getDate())
html = html.replace('<!--TOOL-->', net.getTool())
html = html.replace('<!--COMPCOUNT-->', "<b>Component Count:</b>" + \
str(len(net.components)))
row = "<tr><th style='width:640px'>Ref</th>" + "<th>Qnty</th>"
row += "<th>Value</th>" + "<th>Part</th>" + "<th>Footprint</th>"
row += "<th>Description</th>" + "<th>Vendor</th></tr>"
html = html.replace('<!--TABLEROW-->', row + "<!--TABLEROW-->")
components = net.getInterestingComponents()
# Get all of the components in groups of matching parts + values
# (see kicad_netlist_reader.py)
grouped = net.groupComponents(components)
# Output all of the component information
for group in grouped:
refs = ""
# Add the reference of every component in the group and keep a reference
# to the component so that the other data can be filled in once per group
for component in group:
if len(refs) > 0:
refs += ", "
refs += component.getRef()
c = component
row = "\n "
row += "<tr><td>" + refs +"</td><td>" + str(len(group))
row += "</td><td>" + c.getValue() + "</td><td>" + c.getLibName() + ":"
row += c.getPartName() + "</td><td>" + c.getFootprint() + "</td><td>"
row += c.getDescription() + "</td><td>" + c.getField("Vendor")
row += "</td></tr>"
html = html.replace('<!--TABLEROW-->', row + "<!--TABLEROW-->")
# Print the formatted html to output file
print(html, file=f)