Mercurial > repos > nick > dna_visualizer

comparison visualize.py @ 8:534516105ca9

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
visualize.py: The second tool, for a more aesthetic, abstract visualization.
author Nick Stoler <nstoler@psu.edu>
date Sun, 02 Mar 2014 23:47:41 -0500
parents
children
comparison
equal deleted inserted replaced
7:2d99feb4a8e3 8:534516105ca9
1 #!/usr/bin/python
2 import sys
3 import random
4 import hashlib
5 import Image
6 import ImageDraw
7
8 def main():
9
10 if len(sys.argv) == 1:
11 print """USAGE:
12 $ ./visualize.py genome.fa outfile.png pxsize
13 The last two arguments are optional, but must be in those positions.
14 If no outfile name is given, it will attempt to display the image directly.
15 If no pxsize is given, the default is 512. A power of 2 is highly
16 recommended as the resulting image will be much better laid out."""
17 sys.exit(0)
18
19 DEFAULT_SIZE = (512,512)
20 size = DEFAULT_SIZE
21 if len(sys.argv) > 3:
22 user_size = int(sys.argv[3])
23 size = (user_size,user_size)
24
25 # can skip hashing and specify the seed as an option (for testing)
26 if len(sys.argv) > 4:
27 seed = sys.argv[4]
28 else:
29 seed = get_hash(sys.argv[1])
30 print "seed: "+seed
31 random.seed(seed)
32 level = 1
33 layers = []
34 image = draw_layer(size, level)
35 while 1:
36 level+=1
37 layers.append(draw_layer(size, level))
38 if layers[-1] == False:
39 break
40 # opacity = 256/2**level # opacity #1
41 opacity = 256/2**(level-1) # opacity #2
42 # opacity = 256*(level/2)/2**(level-1) # opacity #3
43 # opacity = 256*(level/float(2**level)) # opacity #4
44 print "opacity: "+str(int(opacity/2.56))+"%"
45 mask = Image.new("L", size, color=opacity)
46 image.paste(layers[-1], (0,0), mask)
47
48 if len(sys.argv) > 2:
49 image.save(sys.argv[2], 'PNG')
50 else:
51 image.show()
52
53 def draw_layer(image_size, level):
54 """Draw every block for a particular layer
55 (blocks of a particular pixel size).
56 Returns an image of the finished layer."""
57 (image_width,image_height) = image_size
58 (width, height) = block_size(image_size, level)
59 if width < 1 or height < 1:
60 return False
61 print "width, height: "+str(width)+", "+str(height)
62 layer = Image.new("RGB", image_size, (0,0,0))
63 draw = ImageDraw.Draw(layer)
64 (x,y) = (0,0)
65 while y < image_height:
66 while x < image_width:
67 draw.rectangle([(x,y), (x+width-1, y+height-1)], fill=randcolor())
68 x += width
69 y += height
70 x = 0
71 return layer
72
73 def get_hash(filepath):
74 """Compute hash of the file"""
75 hashed = hashlib.sha256()
76 with open(filepath, 'rb') as filehandle:
77 for chunk in iter(lambda: filehandle.read(65536), b''):
78 hashed.update(chunk)
79 return hashed.hexdigest()
80
81 def randcolor():
82 """Return a tuple of random color values"""
83 color = []
84 for i in range(3):
85 color.append(random.randrange(0,255))
86 return tuple(color)
87
88 def block_size(image_size, level):
89 """Compute the block width and height for this layer."""
90 (width,height) = image_size
91 while level > 0:
92 width = width/2
93 height = height/4
94 level-=1
95 if level < 1:
96 break
97 width = width/4
98 height = height/2
99 level-=1
100 return (width,height)
101
102 def format_genome():
103 """Eventually I'd like to attempt to process the genome file into a standard
104 format, stripping out details of representation that can change the output,
105 such as:
106 - upper/lowercase
107 - line endings
108 - chromosome naming
109 - chromosome order
110 - noncanonical chromosomes"""
111
112 def check_genome():
113 """Eventually I'd like this to check assumptions I might make about the
114 genome file. A first check would be that it contains all the chromosomes, in
115 the format ''>chr1' or ''>1'. If these assumptions (which format_genome()
116 depends on) aren't met, and the format isn't recognized, then default to a
117 straight digest of the unmodified file."""
118
119 if __name__ == "__main__":
120 main()