testing multiprocessing and rpy2 (ape)

test_mp_rpy2.ipynb.json

{
 "metadata": {
  "name": "",
  "signature": "sha256:04f733fe38a12e7bbe8ee8d3f2900d3fecf0ae7a83ba728cd84650110e26b216"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import os\n",
      "os.environ['R_HOME'] = '/home/cfriedline/lib64/R'\n",
      "import rpy2.robjects as robjects\n",
      "import random\n",
      "import string\n",
      "import tempfile\n",
      "import dendropy\n",
      "from multiprocessing import Pool"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 84
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "r = robjects.r\n",
      "ape = r('library(ape)')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 85
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def create_tree(num_tips, type):\n",
      "    \"\"\"\n",
      "    creates the taxa tree in R\n",
      "    @param num_tips: number of taxa to create\n",
      "    @param type: type for naming (e.g., 'taxa')\n",
      "    @return: a dendropy Tree\n",
      "    @rtype: dendropy.Tree\n",
      "    \"\"\"\n",
      "    r = robjects.r\n",
      "    robjects.globalenv['numtips'] = num_tips\n",
      "    robjects.globalenv['treetype'] = type\n",
      "    name = _get_random_string(20)\n",
      "    if type == \"T\":\n",
      "        r(\"%s = rtree(numtips, rooted=T, tip.label=paste(treetype, seq(1:(numtips)), sep=''))\" % name)\n",
      "    else:\n",
      "        r(\"%s = rtree(numtips, rooted=F, tip.label=paste(treetype, seq(1:(numtips)), sep=''))\" % name)\n",
      "    tree = r[name]\n",
      "    return ape_to_dendropy(tree)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 86
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def ape_to_dendropy(phylo):\n",
      "    \"\"\"\n",
      "    converts an ape tree to dendropy tree\n",
      "    @param phylo: ape instance from rpy2\n",
      "    @return: a dendropy tree\n",
      "    @rtype: dendropy.Tree\n",
      "    \"\"\"\n",
      "    f = tempfile.NamedTemporaryFile()\n",
      "    robjects.r['write.nexus'](phylo, file=f.name)\n",
      "    tree = dendropy.Tree.get_from_path(f.name, \"nexus\")\n",
      "    f.close()\n",
      "    return tree"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 87
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def _get_random_string(length):\n",
      "    \"\"\"\n",
      "    gets a random string of letters/numbers, ensuring that it does not start with a\n",
      "    number\n",
      "    @param length: length of the string\n",
      "    @return: the random string\n",
      "    @rtype: string\n",
      "    \"\"\"\n",
      "    s = ''.join(random.choice(string.letters + string.digits) for i in xrange(length))\n",
      "    if not s[0] in string.letters:\n",
      "        return _get_random_string(length)\n",
      "    return s"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 88
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "tree = create_tree(100, \"T\")\n",
      "tree.as_newick_string()[0:80]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 89,
       "text": [
        "'((((((((T82:0.6231970771,(T94:0.4972914483,T65:0.1885578725):0.3080936125):0.260'"
       ]
      }
     ],
     "prompt_number": 89
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def create_tree_mp(num_taxa):\n",
      "    t = create_tree(num_taxa, \"T\")\n",
      "    return t\n",
      "\n",
      "def get_taxa_trees(num_trees, num_taxa):\n",
      "    jobs_mp = []\n",
      "    jobs = []\n",
      "    res = []\n",
      "    pool = Pool(num_trees)\n",
      "    for i in xrange(num_trees):\n",
      "        jobs_mp.append(pool.apply_async(create_tree_mp, (num_taxa,)))\n",
      "        jobs.append(create_tree_mp(num_taxa))\n",
      "    pool.close()\n",
      "    pool.join()\n",
      "    res.append(jobs)\n",
      "    res.append([x.get() for x in jobs_mp])\n",
      "    return res"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 90
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "trees = get_taxa_trees(5, 10)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 91
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for t in trees[0]:\n",
      "    print t.as_newick_string()[0:80]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "(((T3:0.2269473344,((T2:0.1200612716,T5:0.0573881655):0.9391158754,T9:0.94033852\n",
        "((T9:0.7055083751,(((T1:0.8150142939,T7:0.2128749883):0.5196571311,T4:0.07766074\n",
        "((T6:0.8410235606,(T5:0.01358497259,T2:0.4641908032):0.8032400277):0.6838729084,\n",
        "(((T2:0.8350919618,T4:0.3222864345):0.7693860317,T6:0.08913701377):0.1574694319,\n",
        "((((T9:0.8485502249,T2:0.6824201567):0.5791750827,(T8:0.4604402625,T5:0.14715224\n"
       ]
      }
     ],
     "prompt_number": 92
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for t in trees[1]:\n",
      "    print t.as_newick_string()[0:80]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "(((T3:0.2269473344,((T2:0.1200612716,T5:0.0573881655):0.9391158754,T9:0.94033852\n",
        "(((T3:0.2269473344,((T2:0.1200612716,T5:0.0573881655):0.9391158754,T9:0.94033852\n",
        "(((T3:0.2269473344,((T2:0.1200612716,T5:0.0573881655):0.9391158754,T9:0.94033852\n",
        "(((T3:0.2269473344,((T2:0.1200612716,T5:0.0573881655):0.9391158754,T9:0.94033852\n",
        "(((T3:0.2269473344,((T2:0.1200612716,T5:0.0573881655):0.9391158754,T9:0.94033852\n"
       ]
      }
     ],
     "prompt_number": 93
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}