source: mds-and-trees/tree-genealogy.py @ 568

# Draws a genealogical tree (generates a SVG file) based on parent-child relationship information.

import json
import random
import math
import argparse

TIME = "GENERATIONAL"
BALANCE = "MIN"

# ------SVG---------
svg_file = 0

svg_line_style = 'stroke="rgb(90%,10%,16%)" stroke-width="1" stroke-opacity="0.8"'
svg_dot_style = 'r="2" stroke="black" stroke-width="0.2" fill="red"'
svg_spine_line_style = 'stroke="rgb(0%,0%,80%)" stroke-width="2" stroke-opacity="1"'
#svg_spine_dot_style = 'r="3" stroke="black" stroke-width="0.4" fill="rgb(50%,50%,100%)"'
svg_spine_dot_style = 'r="1" stroke="black" stroke-width="0.2" fill="rgb(50%,50%,100%)"'

def svg_add_line(from_pos, to_pos, style=svg_line_style):
    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]) + '" />')

def svg_add_dot(pos, style=svg_dot_style):
    svg_file.write('<circle ' + style + ' cx="' + str(pos[0]) + '" cy="' + str(pos[1]) + '" />')

def svg_generate_line_style(percent):
    # hotdog
    from_col = [100, 70, 0]
    to_col = [60, 0, 0]
    # neon
    # from_col = [30, 200, 255]
    # to_col = [240, 0, 220]

    from_opa = 0.2
    to_opa = 1.0
    from_stroke = 1
    to_stroke = 3

    opa = from_opa*(1-percent) + to_opa*percent
    stroke = from_stroke*(1-percent) + to_stroke*percent

    percent = 1 - ((1-percent)**20)

    return 'stroke="rgb(' + str(from_col[0]*(1-percent) + to_col[0]*percent) + '%,' \
           + str(from_col[1]*(1-percent) + to_col[1]*percent) + '%,' \
           + str(from_col[2]*(1-percent) + to_col[2]*percent) + '%)" stroke-width="' + str(stroke) + '" stroke-opacity="' + str(opa) + '"'

def svg_generate_dot_style(percent):
    from_col = [100, 90, 0]
    to_col = [60, 0, 0]
    # neon
    # from_col = [30, 200, 255]
    # to_col = [240, 0, 220]

    from_opa = 0.2
    to_opa = 1.0

    opa = from_opa*(1-percent) + to_opa*percent

    percent = 1 - ((1-percent)**20)

    return 'fill="rgb(' + str(from_col[0]*(1-percent) + to_col[0]*percent) + '%,' \
           + str(from_col[1]*(1-percent) + to_col[1]*percent) + '%,' \
           + str(from_col[2]*(1-percent) + to_col[2]*percent) + '%)" r="1.5" stroke="black" stroke-width="0.2" fill-opacity="' + str(opa) + '" ' \
           'stroke-opacity="' + str(opa) + '"'

# -------------------

def load_data(dir):
    global firstnode, nodes, inv_nodes
    f = open(dir)
    for line in f:
        sline = line.split(' ', 2)
        if len(sline) == 3:
            if sline[1] == "[OFFSPRING]":
                creature = json.loads(sline[2])
                #print("B" +str(creature))
                if "FromIDs" in creature:
                    assert(len(creature["FromIDs"]) == 1)
                    nodes[creature["ID"]] = creature["FromIDs"][0]
                    if not creature["FromIDs"][0] in nodes:
                        firstnode = creature["FromIDs"][0]
                if "Time" in creature:
                    time[creature["ID"]] = creature["Time"]

    for k, v in sorted(nodes.items()):
        inv_nodes[v] = inv_nodes.get(v, [])
        inv_nodes[v].append(k)


def load_simple_data(dir):
    global firstnode, nodes, inv_nodes
    f = open(dir)
    for line in f:
        sline = line.split()
        if len(sline) > 1:
            #if int(sline[0]) > 15000:
            #    break
            if sline[0] == firstnode:
                continue
            nodes[sline[0]] = str(max(int(sline[1]), int(firstnode)))
        else:
            firstnode = sline[0]

    for k, v in sorted(nodes.items()):
        inv_nodes[v] = inv_nodes.get(v, [])
        inv_nodes[v].append(k)

    #print(str(inv_nodes))
    #quit()

def compute_depth(node):
    my_depth = 0
    if node in inv_nodes:
        for c in inv_nodes[node]:
            my_depth = max(my_depth, compute_depth(c)+1)
    depth[node] = my_depth
    return my_depth

# ------------------------------------

def xmin_crowd(x1, x2, y):
    if BALANCE == "RANDOM":
        return (x1 if random.randrange(2) == 0 else x2)
    elif BALANCE == "MIN":
        x1_closest = 999999
        x2_closest = 999999
        for pos in positions:
            pos = positions[pos]
            if pos[1] == y:
                x1_closest = min(x1_closest, abs(x1-pos[0]))
                x2_closest = min(x2_closest, abs(x2-pos[0]))
        return (x1 if x1_closest > x2_closest else x2)
    elif BALANCE == "DENSITY":
        x1_dist = 0
        x2_dist = 0
        for pos in positions:
            pos = positions[pos]
            if pos[1] > y-10 or pos[1] < y+10:
                dy = pos[1]-y
                dx1 = pos[0]-x1
                dx2 = pos[0]-x2

                x1_dist += math.sqrt(dy**2 + dx1**2)
                x2_dist += math.sqrt(dy**2 + dx2**2)
        return (x1 if x1_dist > x2_dist else x2)

# ------------------------------------

def prepos_children_reccurent(node):
    for c in inv_nodes[node]:
        #print(node + "->" + c)
        dissimilarity = 0.5 #random.gauss(0,0.3)
        if TIME == "BIRTHS":
            id = ""
            if c[0] == "c":
                id = int(c[1:])
            else:
                id = int(c)
            positions[c] = [xmin_crowd(positions[node][0]-dissimilarity, positions[node][0]+dissimilarity, id), id]
        elif TIME == "GENERATIONAL":
            positions[c] = [xmin_crowd(positions[node][0]-dissimilarity, positions[node][0]+dissimilarity, positions[node][1]+1), positions[node][1]+1]
        elif TIME == "REAL":
            positions[c] = [xmin_crowd(positions[node][0]-dissimilarity, positions[node][0]+dissimilarity, time[c]), time[c]]

    for c in inv_nodes[node]:
        if c in inv_nodes:
            prepos_children_reccurent(c)

def prepos_children():
    global max_height, max_width, min_width

    if not bool(time):
        print("REAL time requested, but no real time data provided. Assuming BIRTHS time instead.")
        TIME = "BIRTHS"

    positions[firstnode] = [0, 0]

    prepos_children_reccurent(firstnode)

    for pos in positions:
        max_height = max(max_height, positions[pos][1])
        max_width = max(max_width, positions[pos][0])
        min_width = min(min_width, positions[pos][0])

# ------------------------------------

def draw_children_recurrent(node, max_depth):
    global max_height, max_width, min_width
    for c in inv_nodes[node]:
        if c in inv_nodes:
            draw_children_recurrent(c, max_depth)

        line_style = (svg_line_style if args.mono_tree else svg_generate_line_style(depth[c]/max_depth))
        dot_style = (svg_dot_style if args.mono_tree else svg_generate_dot_style(depth[c]/max_depth))

        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),
            (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)
        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)
def draw_children():
    max_depth = 0
    for k, v in depth.items():
            max_depth = max(max_depth, v)
    draw_children_recurrent(firstnode, max_depth)
    dot_style = (svg_dot_style if args.mono_tree else svg_generate_dot_style(depth[firstnode]/max_depth))
    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)

def draw_spine_recurrent(node):
    global max_height, max_width, min_width
    for c in inv_nodes[node]:
        if depth[c] == depth[node] - 1:
            if c in inv_nodes:
                draw_spine_recurrent(c)

            line_style = svg_spine_line_style
            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),
                (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)
            #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)
def draw_spine():
    draw_spine_recurrent(firstnode)
    #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)

def draw_skeleton_reccurent(node, max_depth):
    global max_height, max_width, min_width
    for c in inv_nodes[node]:
        if depth[c] >= min_skeleton_depth or depth[c] == max([depth[q] for q in inv_nodes[node]]):
            if c in inv_nodes:
                draw_skeleton_reccurent(c, max_depth)

            line_style = svg_spine_line_style
            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),
                (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)
            #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)
def draw_skeleton():
    max_depth = 0
    for k, v in depth.items():
            max_depth = max(max_depth, v)

    draw_skeleton_reccurent(firstnode, max_depth)
    #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)



##################################################### main #####################################################

args = 0

h = 800
w = 600
h_margin = 10
w_margin = 10
h_no_margs = h - 2* h_margin
w_no_margs = w - 2* w_margin

max_height = 0
max_width = 0
min_width = 9999999999

min_skeleton_depth = 0

firstnode = ""
nodes = {}
inv_nodes = {}
positions = {}
depth = {}
time = {}

def main():
    global svg_file, min_skeleton_depth, args, TIME, BALANCE

    parser = argparse.ArgumentParser(description='Process some integers.')
    parser.add_argument('--in', dest='input', required=True, help='input file with stuctured evolutionary data')
    parser.add_argument('--out', dest='output', required=True, help='output file for the evolutionary tree')
    draw_tree_parser = parser.add_mutually_exclusive_group(required=False)
    draw_tree_parser.add_argument('--draw-tree', dest='draw_tree', action='store_true', help='whether drawing the full tree should be skipped')
    draw_tree_parser.add_argument('--no-draw-tree', dest='draw_tree', action='store_false')

    draw_skeleton_parser = parser.add_mutually_exclusive_group(required=False)
    draw_skeleton_parser.add_argument('--draw-skeleton', dest='draw_skeleton', action='store_true', help='whether the skeleton of the tree should be drawn')
    draw_skeleton_parser.add_argument('--no-draw-skeleton', dest='draw_skeleton', action='store_false')

    draw_spine_parser = parser.add_mutually_exclusive_group(required=False)
    draw_spine_parser.add_argument('--draw-spine', dest='draw_spine', action='store_true', help='whether the spine of the tree should be drawn')
    draw_spine_parser.add_argument('--no-draw-spine', dest='draw_spine', action='store_false')

    #TODO: better names for those parameters
    parser.add_argument('--time', default='BIRTHS', dest='time', help='values on vertical axis (REAL/GENERATIONAL)')
    parser.add_argument('--balance', default='MIN',dest='balance', help='method of placing node in the tree (RANDOM/MIN/DENSITY)')

    mono_tree_parser = parser.add_mutually_exclusive_group(required=False)
    mono_tree_parser.add_argument('--mono-tree', dest='mono_tree', action='store_true', help='whether the tree should be drawn with a single color')
    mono_tree_parser.add_argument('--no-mono-tree', dest='mono_tree', action='store_false')

    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')
    parser.add_argument('--seed', type=int, dest='seed', help='seed for the random number generator (-1 for random)')

    parser.add_argument('--simple-data', type=bool, dest='simple_data', help='input data are given in a simple format (#child #parent)')

    parser.set_defaults(mono_tree=False)
    parser.set_defaults(draw_tree=True)
    parser.set_defaults(draw_skeleton=False)
    parser.set_defaults(draw_spine=False)

    parser.set_defaults(seed=-1)

    args = parser.parse_args()

    TIME = args.time
    BALANCE = args.balance

    dir = args.input
    min_skeleton_depth = args.min_skeleton_depth
    seed = args.seed
    if seed == -1:
        seed = random.randint(0, 10000)
    random.seed(seed)
    print("seed:", seed)

    if args.simple_data:
        load_simple_data(dir)
    else:
        load_data(dir)

    compute_depth(firstnode)

    svg_file = open(args.output, "w")
    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" '
                   'width="' + str(w) + '" height="' + str(h) + '">')

    prepos_children()

    if args.draw_tree:
        draw_children()
    if args.draw_skeleton:
        draw_skeleton()
    if args.draw_spine:
        draw_spine()

    svg_file.write("</svg>")
    svg_file.close()

main()