1#! /usr/bin/python
2
3#this is a script to extract given named nodes from a dot file, with
4#the associated edges.  An edge is kept iff for edge x -> y
5# x and y are both nodes specified to be kept.
6
7#known issues: if a line contains '->' and is not an edge line
8#problems will occur.  If node labels do not begin with
9#Node this also will not work.  Since this is designed to work
10#on DSA dot output and not general dot files this is ok.
11#If you want to use this on other files rename the node labels
12#to Node[.*] with a script or something.  This also relies on
13#the length of a node name being 13 characters (as it is in all
14#DSA dot output files)
15
16#Note that the name of the node can be any substring of the actual
17#name in the dot file.  Thus if you say specify COLLAPSED
18#as a parameter this script will pull out all COLLAPSED
19#nodes in the file
20
21#Specifying escape characters in the name like \n also will not work,
22#as Python
23#will make it \\n, I'm not really sure how to fix this
24
25#currently the script prints the names it is searching for
26#to STDOUT, so you can check to see if they are what you intend
27
28import re
29import string
30import sys
31
32
33if len(sys.argv) < 3:
34	print 'usage is ./DSAextract <dot_file_to_modify> \
35			<output_file> [list of nodes to extract]'
36
37#open the input file
38input = open(sys.argv[1], 'r')
39
40#construct a set of node names
41node_name_set = set()
42for name in sys.argv[3:]:
43	node_name_set |= set([name])
44
45#construct a list of compiled regular expressions from the
46#node_name_set
47regexp_list = []
48for name in node_name_set:
49	regexp_list.append(re.compile(name))
50
51#used to see what kind of line we are on
52nodeexp = re.compile('Node')
53#used to check to see if the current line is an edge line
54arrowexp = re.compile('->')
55
56node_set = set()
57
58#read the file one line at a time
59buffer = input.readline()
60while buffer != '':
61	#filter out the unnecessary checks on all the edge lines
62	if not arrowexp.search(buffer):
63		#check to see if this is a node we are looking for
64		for regexp in regexp_list:
65			#if this name is for the current node, add the dot variable name
66			#for the node (it will be Node(hex number)) to our set of nodes
67			if regexp.search(buffer):
68				node_set |= set([re.split('\s+',buffer,2)[1]])
69				break
70	buffer = input.readline()
71
72
73#test code
74#print '\n'
75
76print node_name_set
77
78#print node_set
79
80
81#open the output file
82output = open(sys.argv[2], 'w')
83#start the second pass over the file
84input = open(sys.argv[1], 'r')
85
86buffer = input.readline()
87while buffer != '':
88	#there are three types of lines we are looking for
89	#1) node lines, 2) edge lines 3) support lines (like page size, etc)
90
91	#is this an edge line?
92	#note that this is no completely robust, if a none edge line
93	#for some reason contains -> it will be missidentified
94	#hand edit the file if this happens
95	if arrowexp.search(buffer):
96		#check to make sure that both nodes are in the node list
97		#if they are print this to output
98		nodes = arrowexp.split(buffer)
99		nodes[0] = string.strip(nodes[0])
100		nodes[1] = string.strip(nodes[1])
101		if nodes[0][:13] in node_set and \
102				nodes[1][:13] in node_set:
103					output.write(buffer)
104	elif nodeexp.search(buffer): #this is a node line
105		node = re.split('\s+', buffer,2)[1]
106		if node in node_set:
107			output.write(buffer)
108	else: #this is a support line
109		output.write(buffer)
110	buffer = input.readline()
111
112