gistfile1.py

#!/usr/bin/python

# This program reads a directory containing Fink package descriptions
# and generates a graph of the dependencies among the packages. The
# graph is in the DOT language and can be displayed by Graphviz.

# The pydot and pyparsing modules must be installed.

# Version 1.0
# Written by Trevor Harmon <trevor@vocaro.com>
# License: GPL - http://www.gnu.org/copyleft/gpl.html

import sys, os
from string import *
from pydot import *

packages = 0
multiline_deps = 0
graph = Dot()

# Reads the given file (which is assumed to be a valid Fink package description)
# and returns a dictionary containing the package name ('package') and an array
# of dependencies ('dependencies') as defined in the file.
def get_package_info(filename):

	global multiline_deps

	# Read the file
	infile = open(filename, 'r')
	inlines = infile.readlines()

	package = None;
	dependencies = None;

	# Walk through the file line by line
	for line in inlines:
	
		# Check for package name definition
		if line.startswith('Package:'):
			# Strip out everything but the name
			package = line[len('Package:'):].lower().strip()
			# If this package contains %n, then it must be a split-off. Skip it.
			if package.find('%n') == -1:
				# Strip out everything after (and including) the % symbol. This is to handle
				# packages defined like "svn-swig-pm%type_pkg[perl]"
				package = package.split('%')[0]

		# Check for dependency definitions
		if line.startswith('Depends:'):
			# We don't handle multi-line dependencies (yet). Abort.
			if line.find('<<') >= 0:
				multiline_deps += 1
				return None
			# Convert the dependency list into an array of dependencies
			deps = line[len('Depends:'):].strip().split(',')
			dependencies = []
			for dep in deps:
				# Ignore everything after (and including) the first whitespace
				dep = dep.lower().split()
				# Make sure the dependency actually was defined and is not just whitespace
				if len(dep) > 0:
					# Do percent expansion
					dep[0] = dep[0].replace('%n', package)
					# Strip out everything after (and including) the % symbol. This is to handle
					# packages defined like "svn-swig-pm%type_pkg[perl]"
					dep[0] = dep[0].split('%')[0]
					# Add the dependency to the array
					dependencies.append(dep[0])
			# We're done; quit now to avoid picking up splitoffs
			return {'package' : package, 'dependencies' : dependencies}

	# No dependencies were found
	return {'package' : package, 'dependencies' : dependencies}

# Retrieves the name and dependency information in the given Fink package
# description and adds it to the PyDOT graph instance.
def add_package_to_graph(filename):

	info = get_package_info(filename)

	if info != None:

		global packages;
		packages += 1

		# Add the package name as a node in the graph
		graph.add_node(Node(info['package']))
	
		# Add each dependency as an edge in the graph
		deps = info['dependencies']
		if deps != None:
			for dep in deps:
				graph.add_edge(Edge(info['package'], dep))

# Searches for all .info files in the given directory and adds them to the graph.
def add_directory_to_graph(directory):
	if os.path.isdir(sys.argv[1]):
		# Remove the trailing slash
		if directory.endswith('/'):
			directory = rstrip(directory, '/')
		print 'Parsing', directory
		# Search for all files in the given directory and clean each one
		for root, dirs, files in os.walk(directory):
			for filename in files:
				if filename.endswith('.info'):
					add_package_to_graph(root + '/' + filename)
	else:
		print 'Error:', directory, 'is not a directory.'

if len(sys.argv) < 2:
	print 'Usage: fink-dep-graph <directory> [directory2, directory3, ...]'
else:
	print 'Parsing package descriptions...'
	outfile = 'fink-dependencies.dot';
	for directory in sys.argv[1:]:
		add_directory_to_graph(directory)
	print 'Found', packages, 'packages. (Split-offs are ignored.)'
	if multiline_deps > 0:
		print 'Skipped', multiline_deps, 'packages due to multi-line dependency declarations, which are not supported in this version.'
	graph.write(outfile)
	print 'Wrote dependency graph to', outfile