gistfile1.py

#!/usr/bin/python
import sys
sys.stdout = sys.stderr
from cgi import parse_qsl
import os
os.environ[ 'HOME' ] = '/tmp/'
import matplotlib
matplotlib.interactive(0)
matplotlib.rcParams['path.simplify'] = True

from matplotlib.figure import Figure
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
from cStringIO import StringIO
import numpy as np
from numpy.lib import format
import simplejson

path = os.path.dirname(__file__)
seqids = {}
posns = {}
for ichr in range(1, 6):
    schr = str(ichr)
    # memmap these
    seqids[schr] = format.open_memmap(os.path.join(path, \
                               'at.tome.data.%s.npy' % (schr,)), mode='r')
    # read these into memory.
    posns[schr] = np.load(os.path.join(path, 'at.tome.posn.%s.npy' % schr))


COLORS = [(102, 102, 0), (255, 51, 0), (204, 153, 204), (51, 51, 255),
          (102, 102, 51), (102, 204, 153), (0, 102, 255), (255, 51, 51),
          (255, 102, 0)]


COLORS = [map(lambda x: x/255., list(c)) for c in COLORS]

COLS = 5

def application(env, start_response):

    p = dict(parse_qsl(env['QUERY_STRING']))

    xmin = max(1, int(p['xmin']))
    xmax = int(p['xmax'])
    if xmax < 0: return ['']

    seqid = p['seqid']

    tissue = p.get('tissue')

    aspng = p.get('width')

    minidx = posns[seqid].searchsorted(xmin)
    maxidx = posns[seqid].searchsorted(xmax)



    data = seqids[seqid][minidx: maxidx]
    data_idx = posns[seqid][minidx: maxidx]


    if aspng:
        start_response("200 OK", [("Content-type", "image/png")])
        io = StringIO()

        height = int(p.get('height', 300))
        f = Figure(frameon=False)
        f.canvas = FigureCanvas(f)
        dpi = 128.
        f.set_size_inches(int(p['width'])/dpi, height/dpi)

        ax = f.add_axes((0, 0, 1, 1), frameon=False, xticks=(), yticks=())
        ax.set_autoscale_on(0)
        ax.set_xlim(xmin - 1,xmax)
        ax.set_ylim(0, 1)
    else:
        start_response("200 OK", [("Content-type", "text/plain")])


    if tissue is None:
        if not aspng:
            if not p.get('summary'):
                d = {'+': zip(data_idx.tolist(), data[:,:COLS].tolist()),
                     '-': zip(data_idx.tolist(), data[:,COLS:].tolist())
                    }
            else:
                d = {'+': data[:,:COLS].mean(axis=0).tolist(), '-':
                     data[:,COLS:].mean(axis=0).tolist()}


            return [simplejson.dumps(d)]


        div = 1.
        plus_tot = np.zeros((data.shape[0], )) + 0.5
        minus_tot = plus_tot.copy()
        for i in range(COLS):
            ax.vlines(data_idx, plus_tot, plus_tot + data[:, i + COLS]/div, color=COLORS[i])
            ax.vlines(data_idx, minus_tot, minus_tot - data[:, i]/div, color=COLORS[i])
            plus_tot += data[:, i + COLS]/div
            minus_tot -= data[:, i]/div
    else:
        tissue = int(tissue)
        if not aspng:
            if not p.get('summary'):
                d = {'+': zip(data_idx.tolist(), data[:,tissue].tolist()),
                     '-': zip(data_idx.tolist(), data[:,tissue + COLS].tolist())
                    }
            else:
                d = {'+': data[:,tissue].mean(), '-': data[:,tissue + COLS].mean()}
                if p.get('summary') == 'both':
                    return [str(d['+']) + "," + str(d['-'])]
            return [simplejson.dumps(d)]

        ax.vlines(data_idx, 0.5, 0.5 + data[:, tissue] * COLS, color=COLORS[tissue])
        ax.vlines(data_idx, 0.5, 0.5 - data[:, tissue + COLS] * COLS, color=COLORS[tissue])


    ax.axhline(y=0.5)

    f.savefig(io, format='png', dpi=dpi)
    io.seek(0)
    return [io.read()]