test_mp_rpy2.ipynb-2.json

{
 "metadata": {
  "name": "",
  "signature": "sha256:24625e0326d219f8b77f0317e421df165d0adfb7925fac44e4f9cc5c35a28bd6"
 },
 "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\n",
      "import random\n",
      "import string\n",
      "import tempfile\n",
      "import dendropy\n",
      "from multiprocessing import Pool\n",
      "from IPython.parallel import Client"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "rc = Client(profile='sge')\n",
      "dview = rc[:]\n",
      "lview = rc.load_balanced_view()\n",
      "len(rc)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 2,
       "text": [
        "10"
       ]
      }
     ],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "with dview.sync_imports():\n",
      "    import os\n",
      "    import rpy2\n",
      "    import rpy2.robjects\n",
      "    import random\n",
      "    import string\n",
      "    import tempfile\n",
      "    import dendropy\n",
      "    import socket\n",
      "    from multiprocessing import Pool\n",
      "    \n",
      "def setup_cluster_engines():\n",
      "    os.environ['R_HOME'] = '/home/cfriedline/lib64/R'\n",
      "    r = rpy2.robjects.r\n",
      "    ape = r('library(ape)')\n",
      "    return socket.gethostname(), os.getpid(), os.environ['R_HOME'], rpy2.__version__\n",
      "dview['setup_cluster_engines'] = setup_cluster_engines\n",
      "dview.apply(setup_cluster_engines).get()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "importing os on engine(s)\n",
        "importing rpy2 on engine(s)\n",
        "importing rpy2.robjects on engine(s)\n",
        "importing random on engine(s)\n",
        "importing string on engine(s)"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "importing tempfile on engine(s)\n",
        "importing dendropy on engine(s)\n",
        "importing socket on engine(s)\n",
        "importing Pool from multiprocessing on engine(s)"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n"
       ]
      },
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 3,
       "text": [
        "[('godel97', 26811, '/home/cfriedline/lib64/R', '2.4.0'),\n",
        " ('godel97', 26870, '/home/cfriedline/lib64/R', '2.4.0'),\n",
        " ('godel97', 26819, '/home/cfriedline/lib64/R', '2.4.0'),\n",
        " ('godel97', 26866, '/home/cfriedline/lib64/R', '2.4.0'),\n",
        " ('godel97', 26873, '/home/cfriedline/lib64/R', '2.4.0'),\n",
        " ('godel199', 17510, '/home/cfriedline/lib64/R', '2.4.0'),\n",
        " ('godel199', 17464, '/home/cfriedline/lib64/R', '2.4.0'),\n",
        " ('godel199', 17512, '/home/cfriedline/lib64/R', '2.4.0'),\n",
        " ('godel199', 17511, '/home/cfriedline/lib64/R', '2.4.0'),\n",
        " ('godel199', 17507, '/home/cfriedline/lib64/R', '2.4.0')]"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "r = rpy2.robjects.r\n",
      "ape = r('library(ape)')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "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 = rpy2.robjects.r\n",
      "    rpy2.robjects.globalenv['numtips'] = num_tips\n",
      "    rpy2.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": 5
    },
    {
     "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",
      "    rpy2.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": 6
    },
    {
     "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": 7
    },
    {
     "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": 8,
       "text": [
        "'(((T6:0.3194573249,T88:0.7137461954):0.426507879,T42:0.09966852632):0.2807099563'"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "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_ip = []\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",
      "        jobs_ip.append(lview.apply_async(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",
      "    res.append([x.get() for x in jobs_ip])\n",
      "    return res"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "dview['create_tree'] = create_tree\n",
      "dview['_get_random_string'] = _get_random_string\n",
      "dview['ape_to_dendropy'] = ape_to_dendropy\n",
      "dview['create_tree_mp'] = create_tree_mp\n",
      "dview['get_taxa_trees'] = get_taxa_trees"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 10
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "trees = get_taxa_trees(5, 10)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 11
    },
    {
     "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": [
        "(((T2:0.7014786459,(((T10:0.1535879213,(T6:0.9837731475,T8:0.2049651504):0.79080\n",
        "((T7:0.5806423163,(T5:0.007633442292,T2:0.2177157113):0.09601430735):0.305033218\n",
        "((((((T6:0.2959137154,T7:0.4578026251):0.8627177938,(T8:0.08809919422,T3:0.10175\n",
        "(T7:0.2577645453,((T10:0.1396945077,T4:0.6536828352):0.7271465769,(T2:0.26139505\n",
        "((T10:0.01782502211,T5:0.9478464343):0.4384911573,((T1:0.201613334,T9:0.52947913\n"
       ]
      }
     ],
     "prompt_number": 12
    },
    {
     "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": [
        "(((T2:0.7014786459,(((T10:0.1535879213,(T6:0.9837731475,T8:0.2049651504):0.79080\n",
        "(((T2:0.7014786459,(((T10:0.1535879213,(T6:0.9837731475,T8:0.2049651504):0.79080\n",
        "(((T2:0.7014786459,(((T10:0.1535879213,(T6:0.9837731475,T8:0.2049651504):0.79080\n",
        "(((T2:0.7014786459,(((T10:0.1535879213,(T6:0.9837731475,T8:0.2049651504):0.79080\n",
        "(((T2:0.7014786459,(((T10:0.1535879213,(T6:0.9837731475,T8:0.2049651504):0.79080\n"
       ]
      }
     ],
     "prompt_number": 13
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for t in trees[2]:\n",
      "    print t.as_newick_string()[0:80]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "(((T1:0.1959578383,(T4:0.4092330183,T5:0.1314637179):0.6633579263):0.07092879713\n",
        "((((T10:0.5492198297,(T9:0.8054336836,T8:0.9056934002):0.6771697227):0.465427928\n",
        "(((T9:0.06959938002,((T10:0.05167591712,T1:0.7737517059):0.6082972996,T7:0.63657\n",
        "(T6:0.5223825315,((((T10:0.3305313124,T3:0.1639568484):0.5333597104,(T1:0.155836\n",
        "((T4:0.9040003123,(T1:0.7009222899,(T9:0.5908546504,T7:0.3244189187):0.872680337\n"
       ]
      }
     ],
     "prompt_number": 14
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}