source: mds-and-trees/tree-genealogy.py @ 571

Last change on this file since 571 was 571, checked in by konrad, 8 years ago

Additional options in tree-genealogy.py, dots and jitter
File size: 14.8 KB
Line 
1# Draws a genealogical tree (generates a SVG file) based on parent-child relationship information.
2
3import json
4import random
5import math
6import argparse
7
8TIME = "" # BIRTHS / GENERATIONAL / REAL
9BALANCE = "" # MIN / DENSITY
10
11DOT_STYLE = "" # NONE / NORMAL / CLEAR
12
13JITTER = "" #
14
15# ------SVG---------
16svg_file = 0
17
18svg_line_style = 'stroke="rgb(90%,10%,16%)" stroke-width="1" stroke-opacity="0.8"'
19svg_dot_style = 'r="2" stroke="black" stroke-width="0.2" fill="red"'
20svg_clear_dot_style = 'r="2" stroke="black" stroke-width="0.4" fill="none"'
21
22svg_spine_line_style = 'stroke="rgb(0%,0%,80%)" stroke-width="2" stroke-opacity="1"'
23svg_spine_dot_style = 'r="1" stroke="black" stroke-width="0.2" fill="rgb(50%,50%,100%)"'
24
25def svg_add_line(from_pos, to_pos, style=svg_line_style):
26    svg_file.write('<line ' + style + ' x1="' + str(from_pos[0]) + '" x2="' + str(to_pos[0]) + '" y1="' + str(from_pos[1]) + '" y2="' + str(to_pos[1]) + '" />')
27
28def svg_add_dot(pos, style=svg_dot_style):
29    svg_file.write('<circle ' + style + ' cx="' + str(pos[0]) + '" cy="' + str(pos[1]) + '" />')
30
31def svg_generate_line_style(percent):
32    # hotdog
33    from_col = [100, 70, 0]
34    to_col = [60, 0, 0]
35    # lava
36    # from_col = [100, 80, 0]
37    # to_col = [100, 0, 0]
38    # neon
39    # from_col = [30, 200, 255]
40    # to_col = [240, 0, 220]
41
42    from_opa = 0.2
43    to_opa = 1.0
44    from_stroke = 1
45    to_stroke = 3
46
47    opa = from_opa*(1-percent) + to_opa*percent
48    stroke = from_stroke*(1-percent) + to_stroke*percent
49
50    percent = 1 - ((1-percent)**20)
51
52    return 'stroke="rgb(' + str(from_col[0]*(1-percent) + to_col[0]*percent) + '%,' \
53           + str(from_col[1]*(1-percent) + to_col[1]*percent) + '%,' \
54           + str(from_col[2]*(1-percent) + to_col[2]*percent) + '%)" stroke-width="' + str(stroke) + '" stroke-opacity="' + str(opa) + '"'
55
56def svg_generate_dot_style(percent):
57    from_col = [100, 70, 0]
58    to_col = [60, 0, 0]
59    # neon
60    # from_col = [30, 200, 255]
61    # to_col = [240, 0, 220]
62
63    from_opa = 0.2
64    to_opa = 1.0
65
66    opa = from_opa*(1-percent) + to_opa*percent
67
68    percent = 1 - ((1-percent)**20)
69
70    return 'fill="rgb(' + str(from_col[0]*(1-percent) + to_col[0]*percent) + '%,' \
71           + str(from_col[1]*(1-percent) + to_col[1]*percent) + '%,' \
72           + str(from_col[2]*(1-percent) + to_col[2]*percent) + '%)" r="1.5" stroke="black" stroke-width="0.2" fill-opacity="' + str(opa) + '" ' \
73           'stroke-opacity="' + str(opa) + '"'
74
75# -------------------
76
77def load_data(dir):
78    global firstnode, nodes, inv_nodes, time
79    f = open(dir)
80    for line in f:
81        sline = line.split(' ', 1)
82        if len(sline) == 2:
83            if sline[0] == "[OFFSPRING]":
84                creature = json.loads(sline[1])
85                #print("B" +str(creature))
86                if "FromIDs" in creature:
87                    #assert(len(creature["FromIDs"]) == 1)
88                    nodes[creature["ID"]] = creature["FromIDs"][0]
89                    if not creature["FromIDs"][0] in nodes:
90                        firstnode = creature["FromIDs"][0]
91                if "Time" in creature:
92                    time[creature["ID"]] = creature["Time"]
93
94    for k, v in sorted(nodes.items()):
95        inv_nodes[v] = inv_nodes.get(v, [])
96        inv_nodes[v].append(k)
97
98
99def load_simple_data(dir):
100    global firstnode, nodes, inv_nodes
101    f = open(dir)
102    for line in f:
103        sline = line.split()
104        if len(sline) > 1:
105            #if int(sline[0]) > 15000:
106            #    break
107            if sline[0] == firstnode:
108                continue
109            nodes[sline[0]] = str(max(int(sline[1]), int(firstnode)))
110        else:
111            firstnode = sline[0]
112
113    for k, v in sorted(nodes.items()):
114        inv_nodes[v] = inv_nodes.get(v, [])
115        inv_nodes[v].append(k)
116
117    #print(str(inv_nodes))
118    #quit()
119
120def compute_depth(node):
121    my_depth = 0
122    if node in inv_nodes:
123        for c in inv_nodes[node]:
124            my_depth = max(my_depth, compute_depth(c)+1)
125    depth[node] = my_depth
126    return my_depth
127
128# ------------------------------------
129
130def xmin_crowd(x1, x2, y):
131    if BALANCE == "RANDOM":
132        return (x1 if random.randrange(2) == 0 else x2)
133    elif BALANCE == "MIN":
134        x1_closest = 999999
135        x2_closest = 999999
136        for pos in positions:
137            pos = positions[pos]
138            if pos[1] == y:
139                x1_closest = min(x1_closest, abs(x1-pos[0]))
140                x2_closest = min(x2_closest, abs(x2-pos[0]))
141        return (x1 if x1_closest > x2_closest else x2)
142    elif BALANCE == "DENSITY":
143        x1_dist = 0
144        x2_dist = 0
145        for pos in positions:
146            pos = positions[pos]
147            if pos[1] > y-10 or pos[1] < y+10:
148                dy = pos[1]-y
149                dx1 = pos[0]-x1
150                dx2 = pos[0]-x2
151
152                x1_dist += math.sqrt(dy**2 + dx1**2)
153                x2_dist += math.sqrt(dy**2 + dx2**2)
154        return (x1 if x1_dist > x2_dist else x2)
155
156# ------------------------------------
157
158def prepos_children_reccurent(node):
159    for c in inv_nodes[node]:
160        #print(node + "->" + c)
161        if JITTER == True:
162            dissimilarity = random.gauss(0,1)
163        else:
164            dissimilarity = 1
165            #TODO take this info from proper fields
166
167        if TIME == "BIRTHS":
168            id = ""
169            if c[0] == "c":
170                id = int(c[1:])
171            else:
172                id = int(c)
173            positions[c] = [xmin_crowd(positions[node][0]-dissimilarity, positions[node][0]+dissimilarity, id), id]
174        elif TIME == "GENERATIONAL":
175            positions[c] = [xmin_crowd(positions[node][0]-dissimilarity, positions[node][0]+dissimilarity, positions[node][1]+1), positions[node][1]+1]
176        elif TIME == "REAL":
177            positions[c] = [xmin_crowd(positions[node][0]-dissimilarity, positions[node][0]+dissimilarity, time[c]), time[c]]
178
179    for c in inv_nodes[node]:
180        if c in inv_nodes:
181            prepos_children_reccurent(c)
182
183def prepos_children():
184    global max_height, max_width, min_width
185
186    if not bool(time):
187        print("REAL time requested, but no real time data provided. Assuming BIRTHS time instead.")
188        TIME = "BIRTHS"
189
190    positions[firstnode] = [0, 0]
191
192    prepos_children_reccurent(firstnode)
193
194    for pos in positions:
195        max_height = max(max_height, positions[pos][1])
196        max_width = max(max_width, positions[pos][0])
197        min_width = min(min_width, positions[pos][0])
198
199# ------------------------------------
200
201def draw_children_recurrent(node, max_depth):
202    global max_height, max_width, min_width
203    for c in inv_nodes[node]:
204        if c in inv_nodes:
205            draw_children_recurrent(c, max_depth)
206
207        line_style = (svg_line_style if args.mono_tree else svg_generate_line_style(depth[c]/max_depth))
208        svg_add_line( (w_margin+w_no_margs*(positions[node][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[node][1]/max_height),
209            (w_margin+w_no_margs*(positions[c][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[c][1]/max_height), line_style)
210
211        if DOT_STYLE == "NONE":
212            continue
213        elif DOT_STYLE == "CLEAR":
214            dot_style = svg_clear_dot_style
215        else: # NORMAL, default
216            dot_style = svg_generate_dot_style(depth[c]/max_depth)
217        svg_add_dot( (w_margin+w_no_margs*(positions[c][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[c][1]/max_height), dot_style)
218def draw_children():
219    max_depth = 0
220    for k, v in depth.items():
221            max_depth = max(max_depth, v)
222    draw_children_recurrent(firstnode, max_depth)
223
224    if DOT_STYLE == "NONE":
225        return
226    elif DOT_STYLE == "CLEAR":
227        dot_style = svg_clear_dot_style
228    else: # NORMAL, default
229        dot_style = svg_generate_dot_style(depth[firstnode]/max_depth)
230    svg_add_dot( (w_margin+w_no_margs*(positions[firstnode][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[firstnode][1]/max_height), dot_style)
231
232def draw_spine_recurrent(node):
233    global max_height, max_width, min_width
234    for c in inv_nodes[node]:
235        if depth[c] == depth[node] - 1:
236            if c in inv_nodes:
237                draw_spine_recurrent(c)
238
239            line_style = svg_spine_line_style
240            svg_add_line( (w_margin+w_no_margs*(positions[node][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[node][1]/max_height),
241                (w_margin+w_no_margs*(positions[c][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[c][1]/max_height), line_style)
242            #svg_add_dot( (w_margin+w_no_margs*(positions[c][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[c][1]/max_height), svg_spine_dot_style)
243def draw_spine():
244    draw_spine_recurrent(firstnode)
245    #svg_add_dot( (w_margin+w_no_margs*(positions[firstnode][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[firstnode][1]/max_height), svg_spine_dot_style)
246
247def draw_skeleton_reccurent(node, max_depth):
248    global max_height, max_width, min_width
249    for c in inv_nodes[node]:
250        if depth[c] >= min_skeleton_depth or depth[c] == max([depth[q] for q in inv_nodes[node]]):
251            if c in inv_nodes:
252                draw_skeleton_reccurent(c, max_depth)
253
254            line_style = svg_spine_line_style
255            svg_add_line( (w_margin+w_no_margs*(positions[node][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[node][1]/max_height),
256                (w_margin+w_no_margs*(positions[c][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[c][1]/max_height), line_style)
257            #svg_add_dot( (w_margin+w_no_margs*(positions[c][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[c][1]/max_height),
258            #             svg_spine_dot_style)
259def draw_skeleton():
260    max_depth = 0
261    for k, v in depth.items():
262            max_depth = max(max_depth, v)
263
264    draw_skeleton_reccurent(firstnode, max_depth)
265    #svg_add_dot( (w_margin+w_no_margs*(positions[firstnode][0]-min_width)/(max_width-min_width), h_margin+h_no_margs*positions[firstnode][1]/max_height),
266    #             svg_spine_dot_style)
267
268
269
270##################################################### main #####################################################
271
272args = 0
273
274h = 800
275w = 600
276h_margin = 10
277w_margin = 10
278h_no_margs = h - 2* h_margin
279w_no_margs = w - 2* w_margin
280
281max_height = 0
282max_width = 0
283min_width = 9999999999
284
285min_skeleton_depth = 0
286
287firstnode = ""
288nodes = {}
289inv_nodes = {}
290positions = {}
291depth = {}
292time = {}
293
294def main():
295    global svg_file, min_skeleton_depth, args, TIME, BALANCE, DOT_STYLE, JITTER
296
297    parser = argparse.ArgumentParser(description='Process some integers.')
298    parser.add_argument('--in', dest='input', required=True, help='input file with stuctured evolutionary data')
299    parser.add_argument('--out', dest='output', required=True, help='output file for the evolutionary tree')
300    draw_tree_parser = parser.add_mutually_exclusive_group(required=False)
301    draw_tree_parser.add_argument('--draw-tree', dest='draw_tree', action='store_true', help='whether drawing the full tree should be skipped')
302    draw_tree_parser.add_argument('--no-draw-tree', dest='draw_tree', action='store_false')
303
304    draw_skeleton_parser = parser.add_mutually_exclusive_group(required=False)
305    draw_skeleton_parser.add_argument('--draw-skeleton', dest='draw_skeleton', action='store_true', help='whether the skeleton of the tree should be drawn')
306    draw_skeleton_parser.add_argument('--no-draw-skeleton', dest='draw_skeleton', action='store_false')
307
308    draw_spine_parser = parser.add_mutually_exclusive_group(required=False)
309    draw_spine_parser.add_argument('--draw-spine', dest='draw_spine', action='store_true', help='whether the spine of the tree should be drawn')
310    draw_spine_parser.add_argument('--no-draw-spine', dest='draw_spine', action='store_false')
311
312    #TODO: better names for those parameters
313    parser.add_argument('--time', default='BIRTHS', dest='time', help='values on vertical axis (BIRTHS/GENERATIONAL/REAL); '
314                                                                      'BIRTHS: time measured as the number of births since the beggining; '
315                                                                      'GENERATIONAL: time measured as number of ancestors; '
316                                                                      'REAL: real time of the simulation')
317    parser.add_argument('--balance', default='MIN', dest='balance', help='method of placing node in the tree (RANDOM/MIN/DENSITY)')
318
319    parser.add_argument('--dots', default='NORMAL', dest='dots', help='method of drawing dots (individuals) (NONE/NORMAL/CLEAR)')
320
321    parser.add_argument('-j', '--jitter', dest="jitter", action='store_true', help='draw horizontal positions of children from the normal distribution')
322
323    mono_tree_parser = parser.add_mutually_exclusive_group(required=False)
324    mono_tree_parser.add_argument('--mono-tree', dest='mono_tree', action='store_true', help='whether the tree should be drawn with a single color')
325    mono_tree_parser.add_argument('--no-mono-tree', dest='mono_tree', action='store_false')
326
327    parser.add_argument('--min-skeleton-depth', type=int, default=2, dest='min_skeleton_depth', help='minimal distance from the leafs for the nodes in the skeleton')
328    parser.add_argument('--seed', type=int, dest='seed', help='seed for the random number generator (-1 for random)')
329
330    parser.add_argument('--simple-data', type=bool, dest='simple_data', help='input data are given in a simple format (#child #parent)')
331
332    parser.set_defaults(mono_tree=False)
333    parser.set_defaults(draw_tree=True)
334    parser.set_defaults(draw_skeleton=False)
335    parser.set_defaults(draw_spine=False)
336
337    parser.set_defaults(seed=-1)
338
339    args = parser.parse_args()
340
341    TIME = args.time
342    BALANCE = args.balance
343    DOT_STYLE = args.dots
344    JITTER = args.jitter
345
346    dir = args.input
347    min_skeleton_depth = args.min_skeleton_depth
348    seed = args.seed
349    if seed == -1:
350        seed = random.randint(0, 10000)
351    random.seed(seed)
352    print("seed:", seed)
353
354    if args.simple_data:
355        load_simple_data(dir)
356    else:
357        load_data(dir)
358
359    compute_depth(firstnode)
360
361    svg_file = open(args.output, "w")
362    svg_file.write('<svg xmlns:svg="http://www.w3.org/2000/svg" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" version="1.0" '
363                   'width="' + str(w) + '" height="' + str(h) + '">')
364
365    prepos_children()
366
367    if args.draw_tree:
368        draw_children()
369    if args.draw_skeleton:
370        draw_skeleton()
371    if args.draw_spine:
372        draw_spine()
373
374    svg_file.write("</svg>")
375    svg_file.close()
376
377main()
378
Note: See TracBrowser for help on using the repository browser.