graph_io.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import json\n",
    "import networkx as nx\n",
    "from networkx.readwrite import json_graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from lib.graph_enumerator import generate_graphs\n",
    "from lib.node_semantics import Node_Name_Rule, Edge_Semantics_Rule\n",
    "from lib import config, result_config\n",
    "from lib.likelihood_calculations import Inference\n",
    "from lib.utils import filename_utility"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "graph_iter = generate_graphs(**config.generator_dictionary)\n",
    "graphs = list(graph_iter)\n",
    "\n",
    "for graph in graphs:    \n",
    "    Node_Name_Rule.graph_semantics_apply(graph,config.node_semantics)\n",
    "    Edge_Semantics_Rule.graph_semantics_apply(graph,config.edge_semantics)\n",
    "\n",
    "test_graph = graphs[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'adjacency': [[{'edge_type': 'hidden_sample', 'id': 'C_★'},\n",
       "   {'edge_type': 'hidden_sample', 'id': 'D_★'},\n",
       "   {'edge_type': 'observed', 'id': 'A_obs'},\n",
       "   {'edge_type': 'hidden_sample', 'id': 'B_★'}],\n",
       "  [{'edge_type': 'hidden_sample', 'id': 'D_★'},\n",
       "   {'edge_type': 'observed', 'id': 'C_obs'},\n",
       "   {'edge_type': 'hidden_sample', 'id': 'B_★'}],\n",
       "  [{'edge_type': 'hidden_sample', 'id': 'C_★'},\n",
       "   {'edge_type': 'observed', 'id': 'D_obs'},\n",
       "   {'edge_type': 'hidden_sample', 'id': 'B_★'}],\n",
       "  [],\n",
       "  [],\n",
       "  [],\n",
       "  [],\n",
       "  [{'edge_type': 'intervention', 'id': 'A_★'},\n",
       "   {'edge_type': 'intervention', 'id': 'A_obs'}],\n",
       "  [{'edge_type': 'hidden_sample', 'id': 'C_★'},\n",
       "   {'edge_type': 'observed', 'id': 'B_obs'},\n",
       "   {'edge_type': 'hidden_sample', 'id': 'D_★'}]],\n",
       " 'directed': True,\n",
       " 'graph': [],\n",
       " 'multigraph': False,\n",
       " 'nodes': [{'id': 'A_★', 'node_type': 'hidden'},\n",
       "  {'id': 'C_★', 'node_type': 'hidden'},\n",
       "  {'id': 'D_★', 'node_type': 'hidden'},\n",
       "  {'id': 'A_obs', 'node_type': 'observed'},\n",
       "  {'id': 'B_obs', 'node_type': 'observed'},\n",
       "  {'id': 'D_obs', 'node_type': 'observed'},\n",
       "  {'id': 'C_obs', 'node_type': 'observed'},\n",
       "  {'id': 'A_int', 'node_type': 'intervener'},\n",
       "  {'id': 'B_★', 'node_type': 'hidden'}]}"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "json_graph.adjacency_data(test_graph)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('A_★', 'C_★', {'edge_type': 'hidden_sample', 'id': 'C_★'}),\n",
       " ('A_★', 'D_★', {'edge_type': 'hidden_sample', 'id': 'D_★'}),\n",
       " ('A_★', 'A_obs', {'edge_type': 'observed', 'id': 'A_obs'}),\n",
       " ('A_★', 'B_★', {'edge_type': 'hidden_sample', 'id': 'B_★'}),\n",
       " ('D_★', 'C_★', {'edge_type': 'hidden_sample', 'id': 'C_★'}),\n",
       " ('D_★', 'D_obs', {'edge_type': 'observed', 'id': 'D_obs'}),\n",
       " ('D_★', 'B_★', {'edge_type': 'hidden_sample', 'id': 'B_★'}),\n",
       " ('C_★', 'D_★', {'edge_type': 'hidden_sample', 'id': 'D_★'}),\n",
       " ('C_★', 'C_obs', {'edge_type': 'observed', 'id': 'C_obs'}),\n",
       " ('C_★', 'B_★', {'edge_type': 'hidden_sample', 'id': 'B_★'}),\n",
       " ('A_int', 'A_★', {'edge_type': 'intervention', 'id': 'A_★'}),\n",
       " ('A_int', 'A_obs', {'edge_type': 'intervention', 'id': 'A_obs'}),\n",
       " ('B_★', 'C_★', {'edge_type': 'hidden_sample', 'id': 'C_★'}),\n",
       " ('B_★', 'B_obs', {'edge_type': 'observed', 'id': 'B_obs'}),\n",
       " ('B_★', 'D_★', {'edge_type': 'hidden_sample', 'id': 'D_★'})]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_2= json_graph.adjacency_graph(json_graph.adjacency_data(test_graph))\n",
    "test_2.edges(data=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('A_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_★', 'D_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_★', 'A_obs', {'edge_type': 'observed'}),\n",
       " ('A_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('D_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('D_★', 'D_obs', {'edge_type': 'observed'}),\n",
       " ('D_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('C_★', 'D_★', {'edge_type': 'hidden_sample'}),\n",
       " ('C_★', 'C_obs', {'edge_type': 'observed'}),\n",
       " ('C_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_int', 'A_★', {'edge_type': 'intervention'}),\n",
       " ('A_int', 'A_obs', {'edge_type': 'intervention'}),\n",
       " ('B_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('B_★', 'B_obs', {'edge_type': 'observed'}),\n",
       " ('B_★', 'D_★', {'edge_type': 'hidden_sample'})]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for edge in test_2.edges():\n",
    "    attr = test_2[edge[0]][edge[1]]\n",
    "    if 'id' in attr:\n",
    "        del attr['id']\n",
    "test_2.edges(data=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "new_graphs = [json_graph.adjacency_data(graph) for graph in graphs]\n",
    "s = json.dumps(new_graphs)\n",
    "test_graph_back = json.loads(s)\n",
    "newer_graphs = [json_graph.adjacency_graph(js_graph) for js_graph in test_graph_back]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('A_★', 'C_★', {'edge_type': 'hidden_sample', 'id': 'C_★'}),\n",
       " ('A_★', 'D_★', {'edge_type': 'hidden_sample', 'id': 'D_★'}),\n",
       " ('A_★', 'A_obs', {'edge_type': 'observed', 'id': 'A_obs'}),\n",
       " ('A_★', 'B_★', {'edge_type': 'hidden_sample', 'id': 'B_★'}),\n",
       " ('D_★', 'C_★', {'edge_type': 'hidden_sample', 'id': 'C_★'}),\n",
       " ('D_★', 'D_obs', {'edge_type': 'observed', 'id': 'D_obs'}),\n",
       " ('D_★', 'B_★', {'edge_type': 'hidden_sample', 'id': 'B_★'}),\n",
       " ('C_★', 'D_★', {'edge_type': 'hidden_sample', 'id': 'D_★'}),\n",
       " ('C_★', 'C_obs', {'edge_type': 'observed', 'id': 'C_obs'}),\n",
       " ('C_★', 'B_★', {'edge_type': 'hidden_sample', 'id': 'B_★'}),\n",
       " ('A_int', 'A_★', {'edge_type': 'intervention', 'id': 'A_★'}),\n",
       " ('A_int', 'A_obs', {'edge_type': 'intervention', 'id': 'A_obs'}),\n",
       " ('B_★', 'C_★', {'edge_type': 'hidden_sample', 'id': 'C_★'}),\n",
       " ('B_★', 'B_obs', {'edge_type': 'observed', 'id': 'B_obs'}),\n",
       " ('B_★', 'D_★', {'edge_type': 'hidden_sample', 'id': 'D_★'})]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "newer_graphs[0].edges(data=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import json\n",
    "from networkx.readwrite import json_graph\n",
    "from lib.utils import open_filename\n",
    "\n",
    "\n",
    "def json_graph_list_dumps(graph_list):\n",
    "    return json.dumps([json_graph.adjacency_data(graph) for graph in graph_list])\n",
    "    \n",
    "\n",
    "def json_graph_list_dump(graph_list,file):\n",
    "    with open_filename(file,'w') as f:\n",
    "        json.dump([json_graph.adjacency_data(graph) for graph in graph_list],f)\n",
    "    pass\n",
    "\n",
    "def json_graph_list_loads(json_string):\n",
    "    js_graph_list = json.loads(json_string)\n",
    "    return [remove_id_from_json_graphs(json_graph.adjacency_graph(js_graph)) for js_graph in js_graph_list]\n",
    "    \n",
    "def json_graph_list_load(file):\n",
    "    with open_filename(file,'r') as f:\n",
    "        js_graph_list = json.load(f)\n",
    "    return [remove_id_from_json_graphs(json_graph.adjacency_graph(js_graph)) for js_graph in js_graph_list]\n",
    "\n",
    "def remove_id_from_json_graphs(G):\n",
    "    graph = G.copy()\n",
    "    for edge in sorted(graph.edges()):\n",
    "        attr = graph[edge[0]][edge[1]]\n",
    "        if 'id' in attr:\n",
    "            del attr['id']\n",
    "    return graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import os\n",
    "from lib.utils import filename_utility\n",
    "\n",
    "\n",
    "filename_base = \"json_test\"\n",
    "filename = filename_utility(filename_base)\n",
    "filename = os.path.join(\"results\",filename)\n",
    "\n",
    "json_graph_list_dump(graphs,filename)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('A_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_★', 'D_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_★', 'A_obs', {'edge_type': 'observed'}),\n",
       " ('A_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('C_★', 'D_★', {'edge_type': 'hidden_sample'}),\n",
       " ('C_★', 'C_obs', {'edge_type': 'observed'}),\n",
       " ('C_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('D_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('D_★', 'D_obs', {'edge_type': 'observed'}),\n",
       " ('D_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_int', 'A_★', {'edge_type': 'intervention'}),\n",
       " ('A_int', 'A_obs', {'edge_type': 'intervention'}),\n",
       " ('B_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('B_★', 'B_obs', {'edge_type': 'observed'}),\n",
       " ('B_★', 'D_★', {'edge_type': 'hidden_sample'})]"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_read = json_graph_list_load(filename)\n",
    "test_read[0].edges(data=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "test_write_npz = json_graph_list_dumps(graphs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "f_base = \"json_npz_test\"\n",
    "fname = filename_utility(f_base)\n",
    "fname = os.path.join(\"results\",fname)\n",
    "\n",
    "with open(fname,'wb') as f:\n",
    "    np.savez(f,graph_list=[test_write_npz]*4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('A_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_★', 'D_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_★', 'A_obs', {'edge_type': 'observed'}),\n",
       " ('A_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('C_★', 'D_★', {'edge_type': 'hidden_sample'}),\n",
       " ('C_★', 'C_obs', {'edge_type': 'observed'}),\n",
       " ('C_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('D_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('D_★', 'D_obs', {'edge_type': 'observed'}),\n",
       " ('D_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_int', 'A_★', {'edge_type': 'intervention'}),\n",
       " ('A_int', 'A_obs', {'edge_type': 'intervention'}),\n",
       " ('B_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('B_★', 'B_obs', {'edge_type': 'observed'}),\n",
       " ('B_★', 'D_★', {'edge_type': 'hidden_sample'})]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_glist_load = np.load(fname)\n",
    "test_glist = test_glist_load['graph_list'][0]\n",
    "test_nx_list_load = json_graph_list_loads(test_glist)\n",
    "test_nx_list_load[0].edges(data=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import lib.graph_json_io as g_json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "test_module_write_npz = g_json.json_graph_list_dumps(graphs)\n",
    "test_mod_w_npz_list = [test_module_write_npz]*4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('A_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_★', 'D_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_★', 'A_obs', {'edge_type': 'observed'}),\n",
       " ('A_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('C_★', 'D_★', {'edge_type': 'hidden_sample'}),\n",
       " ('C_★', 'C_obs', {'edge_type': 'observed'}),\n",
       " ('C_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('D_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('D_★', 'D_obs', {'edge_type': 'observed'}),\n",
       " ('D_★', 'B_★', {'edge_type': 'hidden_sample'}),\n",
       " ('A_int', 'A_★', {'edge_type': 'intervention'}),\n",
       " ('A_int', 'A_obs', {'edge_type': 'intervention'}),\n",
       " ('B_★', 'C_★', {'edge_type': 'hidden_sample'}),\n",
       " ('B_★', 'B_obs', {'edge_type': 'observed'}),\n",
       " ('B_★', 'D_★', {'edge_type': 'hidden_sample'})]"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f_base = \"g_json_module_npz_test\"\n",
    "fname = filename_utility(f_base)\n",
    "fname = os.path.join(\"results\",fname)\n",
    "\n",
    "with open(fname,'wb') as f:\n",
    "    np.savez(f,module_graph_list=test_mod_w_npz_list)\n",
    "    \n",
    "test_mod_glist_load = np.load(fname)\n",
    "test_mod_glist = test_mod_glist_load['module_graph_list'][0]\n",
    "test_mod_nx_list_load = g_json.json_graph_list_loads(test_mod_glist)\n",
    "test_mod_nx_list_load[0].edges(data=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}