01_Virtual_data_setup.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load Packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:59:38.339549Z",
     "start_time": "2020-07-22T21:59:35.561799Z"
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline \n",
    "from ifis_tools import database_tools as db \n",
    "from ifis_tools import asynch_manager as am \n",
    "from ifis_tools import auxiliar as aux\n",
    "from wmf import wmf\n",
    "import pandas as pd \n",
    "import numpy as np \n",
    "import os \n",
    "import pylab as pl\n",
    "from string import Template\n",
    "from param_ident import core \n",
    "from aclayer import core as ac"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:59:38.424172Z",
     "start_time": "2020-07-22T21:59:38.421850Z"
    }
   },
   "outputs": [],
   "source": [
    "names = {'south': {'path': 'data/for_hlm/south_skunk/'},\n",
    "        'turkey':{'path':'data/for_hlm/turkey/'}}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define Watershed setups\n",
    "\n",
    "In this step we will produce the setups of the watersheds making variations of $v_0$ and $\\lambda_1$ for different sizes of sub-watersheds."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:59:41.010885Z",
     "start_time": "2020-07-22T21:59:38.522995Z"
    }
   },
   "outputs": [],
   "source": [
    "#Get the watershed\n",
    "south = am.hlmModel(linkid=272678, ExtraParams=['h_order'])\n",
    "\n",
    "#Temporal rvr to get the topology of the watershed\n",
    "south.write_rvr('data/for_hlm/south_skunk/south_skunk.rvr')\n",
    "topo = south.topo.copy()\n",
    "topo.set_index('link_id', inplace = True)\n",
    "idx = south.Table.index.intersection(topo.index)\n",
    "south.Table['dest'] = topo['id'].loc[idx]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:59:43.549744Z",
     "start_time": "2020-07-22T21:59:41.133366Z"
    }
   },
   "outputs": [],
   "source": [
    "#Get the watershed\n",
    "turkey = am.hlmModel(linkid=434514, ExtraParams=['h_order'])\n",
    "\n",
    "#Temporal rvr to get the topology of the watershed\n",
    "turkey.write_rvr('data/for_hlm/turkey/turkey.rvr')\n",
    "topo = turkey.topo.copy()\n",
    "topo.set_index('link_id', inplace = True)\n",
    "idx = turkey.Table.index.intersection(topo.index)\n",
    "turkey.Table['dest'] = topo['id'].loc[idx]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:59:43.731610Z",
     "start_time": "2020-07-22T21:59:43.728904Z"
    }
   },
   "outputs": [],
   "source": [
    "wat = {'south': south, 'turkey': turkey}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Obtain sub-watersheds at different prun-levels\n",
    "\n",
    "Here we include levels from 7 to 4, then we will produce random values of $v_0$ and $\\lambda_1$ using a cascade approach."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-07T16:17:22.681258Z",
     "start_time": "2020-07-07T16:11:27.682259Z"
    }
   },
   "outputs": [],
   "source": [
    "# obtain the sub-watersheds for each one (this takes time try not to run it, instead read the data)\n",
    "core.Get_sub_watersheds(wat['south'], 7, 3)\n",
    "core.Get_sub_watersheds(wat['turkey'], 7, 3)\n",
    "# Save to csv files to avoid doing this again.\n",
    "wat['south'].Table.to_csv('data/for_hlm/south_skunk/watershed_table.csv')\n",
    "wat['turkey'].Table.to_csv('data/for_hlm/turkey/watershed_table.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:59:44.018910Z",
     "start_time": "2020-07-22T21:59:43.919909Z"
    }
   },
   "outputs": [],
   "source": [
    "wat['south'].Table = pd.read_csv('data/for_hlm/south_skunk/watershed_table.csv', index_col=0)\n",
    "wat['turkey'].Table = pd.read_csv('data/for_hlm/turkey/watershed_table.csv', index_col=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Distribute $v_0$ and $\\lambda_1$ using the cascade\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Synchronic production of the virtual variables $v_0$ and $\\lambda_1$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:59:44.530295Z",
     "start_time": "2020-07-22T21:59:44.521844Z"
    }
   },
   "outputs": [],
   "source": [
    "def cascade_values_sync(Table, param, levels, seed_values, norm_dev = 0.15):\n",
    "    '''Generates a virtual setup of a watershed for a given set of values (seed_values)\n",
    "    the result is updates on the cu.Table DataFrame. Contrasted with the async version, this \n",
    "    version creates values that depends on the previous level.\n",
    "    Params:\n",
    "        - level: the minimum orton order to distribute the seed_values.\n",
    "        - param: name of the parameter in the table.\n",
    "        - seed_values: a list with the initial values for the biggest horton order.\n",
    "        - norm_dev: deviation aprameter to produce the random cascade.\n",
    "    Resutls:\n",
    "        - cu.Table[param]'''\n",
    "    p = param + '7'\n",
    "    Table[p] = 0.0    \n",
    "    for sub, value in zip(Table['prun_7'].unique(), seed_values):\n",
    "        Table.at[Table['prun_7'] == sub, p] = value\n",
    "\n",
    "    Factors = {}\n",
    "    #Iterate for each level\n",
    "    for level in np.arange(6,levels-1,-1):\n",
    "        \n",
    "        #Initial value\n",
    "        name = 'prun_'+str(level)\n",
    "        p = param + str(level)\n",
    "        Table[p] = Table[param + str(level+1)].copy()\n",
    "        fac = []\n",
    "        \n",
    "        #Iterate through the sub-watersheds of each level `\n",
    "        for subw in Table[name].unique():\n",
    "            factor = np.random.normal(1, norm_dev)            \n",
    "            Table.loc[Table[name] == subw, p] *= factor            \n",
    "            fac.append(factor)\n",
    "        Factors.update({str(level): fac})\n",
    "    return Factors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:52:11.957977Z",
     "start_time": "2020-07-22T21:52:01.192146Z"
    }
   },
   "outputs": [],
   "source": [
    "for name, seed in zip(['turkey','south'], [[0.15,0.2,0.18,0.25,0.3,0.35,0.4], [0.15,0.3,0.22]]):\n",
    "    f_vo = cascade_values_sync(wat[name].Table, 'vo', 4,  [0.15,0.3,0.22], 0.2)\n",
    "for name, seed in zip(['turkey','south'], [[0.19,0.22,0.2,0.22,0.2,0.23,0.3], [0.18,0.23,0.25]]):\n",
    "    f_l1 = cascade_values_sync(wat[name].Table, 'la', 4, seed, 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-09T02:44:33.669188Z",
     "start_time": "2020-07-09T02:44:33.664229Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.22068185877495802"
      ]
     },
     "execution_count": 237,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = a['vo4'] * (a['Acum'] / a['Acum'].sum())\n",
    "x.sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-09T02:45:02.429246Z",
     "start_time": "2020-07-09T02:45:02.424150Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.18913619460298808"
      ]
     },
     "execution_count": 239,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = a['la4'] * (a['Acum'] / a['Acum'].sum())\n",
    "x.sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-09T02:45:19.572996Z",
     "start_time": "2020-07-09T02:45:19.568611Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.16132851376984242"
      ]
     },
     "execution_count": 243,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a['la4'].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot with the obtained distribution at each case."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-08T20:58:43.035112Z",
     "start_time": "2020-07-08T20:58:42.943424Z"
    }
   },
   "outputs": [],
   "source": [
    "wat['south'].Table = pd.read_csv('data/for_hlm/south_skunk/water_virtual.csv', index_col = 0)\n",
    "wat['turkey'].Table = pd.read_csv('data/for_hlm/turkey/water_virtual.csv', index_col = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-08T20:58:52.834640Z",
     "start_time": "2020-07-08T20:58:46.222321Z"
    },
    "hide_input": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x864 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig  = pl.figure(figsize=(20,12))\n",
    "bins = np.arange(0.1,0.5,0.05)\n",
    "\n",
    "###################### Vo\n",
    "\n",
    "ax = fig.add_subplot(221)\n",
    "for n,alpha in zip(['vo6','vo5','vo4'],[0.85,0.6,0.4]):\n",
    "    h,b = ac.get_histogram(wat['south'].Table[n], bins = bins, inf = 0.1, sup = 0.5)\n",
    "    pl.bar(b[:-1], 100 * h / h.sum(), 0.05, align='edge', edgecolor = 'k', lw = 3, alpha = alpha, \n",
    "          label = 'Order ' + n[-1])\n",
    "ax.tick_params(labelsize = 23)\n",
    "ax.grid()\n",
    "ax.set_ylim(0,65)\n",
    "ax.set_title('a) $v_0$ for South skunk', size = 24)\n",
    "ax.set_ylabel('Fraction [$\\%$]', size = 23)\n",
    "ax.legend(loc = 0, fontsize = 'xx-large')\n",
    "ax.set_xticklabels([])\n",
    "\n",
    "ax = fig.add_subplot(222)\n",
    "for n,alpha in zip(['vo6','vo5','vo4'],[0.85,0.6,0.4]):\n",
    "    h,b = ac.get_histogram(wat['turkey'].Table[n], bins = bins, inf = 0.1, sup = 0.5)\n",
    "    pl.bar(b[:-1], 100 * h / h.sum(), 0.05, align='edge', edgecolor = 'k', lw = 3, alpha = alpha, \n",
    "          label = 'Order ' + n[-1])\n",
    "ax.tick_params(labelsize = 23)\n",
    "ax.grid()\n",
    "ax.set_ylim(0,65)\n",
    "ax.set_title('b) $v_0$ for Turkey', size = 24)\n",
    "ax.set_yticklabels([])\n",
    "ax.set_xticklabels([])\n",
    "\n",
    "\n",
    "###################### LAMBDA\n",
    "\n",
    "ax = fig.add_subplot(223)\n",
    "for n,alpha in zip(['la6','la5','la4'],[0.85,0.6,0.4]):\n",
    "    h,b = ac.get_histogram(wat['south'].Table[n], bins = bins, inf = 0.1, sup = 0.5)\n",
    "    pl.bar(b[:-1], 100 * h / h.sum(), 0.05, align='edge', edgecolor = 'k', lw = 3, alpha = alpha, \n",
    "          label = 'Order ' + n[-1])\n",
    "ax.tick_params(labelsize = 23)\n",
    "ax.grid()\n",
    "ax.set_ylim(0,65)\n",
    "ax.set_title('c) $\\lambda_1$ for South skunk', size = 24)\n",
    "ax.set_ylabel('Fraction [$\\%$]', size = 23)\n",
    "\n",
    "ax = fig.add_subplot(224)\n",
    "for n,alpha in zip(['la6','la5','la4'],[0.85,0.6,0.4]):\n",
    "    h,b = ac.get_histogram(wat['turkey'].Table[n], bins = bins, inf = 0.1, sup = 0.5)\n",
    "    pl.bar(b[:-1], 100 * h / h.sum(), 0.05, align='edge', edgecolor = 'k', lw = 3, alpha = alpha, \n",
    "          label = 'Order ' + n[-1])\n",
    "ax.tick_params(labelsize = 23)\n",
    "ax.grid()\n",
    "ax.set_ylim(0,65)\n",
    "ax.set_title('d) $\\lambda_1$ for Turkey', size = 24)\n",
    "ax.set_yticklabels([])\n",
    "pl.savefig('figures/Vo_and_L1_distribution_levels_654.png', bbox_inches = 'tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Write Waterhsed data for the HLM model to run "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creates the control points for both watersheds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:33:34.232656Z",
     "start_time": "2020-07-22T21:33:33.435461Z"
    }
   },
   "outputs": [],
   "source": [
    "#Creates a sav file of the control points.\n",
    "for name in wat.keys():\n",
    "    for k in [4,5,6]:\n",
    "        f = open(names[name]['path'] + 'control_vr'+str(k)+'.sav','w')\n",
    "        for i in wat[name].Table['prun_'+str(k)].unique():\n",
    "            f.write('%d\\n' % wat[name].Table.loc[wat[name].Table['prun_'+str(k)] == i, 'Acum'].idxmax())\n",
    "        f.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Setup the global parameters for both cases "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:59:53.669093Z",
     "start_time": "2020-07-22T21:59:53.641579Z"
    }
   },
   "outputs": [],
   "source": [
    "#Set the global parameters for both watersheds.\n",
    "\n",
    "Params = {'v_r': {'value': 0.02,'format': '%.4f'},\n",
    "    'a_r': {'value': 1.67,'format': '%.2f'},\n",
    "    'a': {'value': 1.5e-6,'format': '%.2e'},\n",
    "    'b': {'value': 17,'format': '%.2f'},\n",
    "    'c': {'value': 5.5e-7,'format': '%.2e'}, # 0.001\n",
    "    'd': {'value': 35,'format': '%.2f'}, # 0.03\n",
    "    'm1': {'value': 2.045e-6,'format': '%.2e'},\n",
    "    'ki_fac': {'value': 0.02,'format': '%.4f'},\n",
    "    'TopDepth': {'value': 0.1,'format': '%.2f'},\n",
    "    'NoFlow': {'value': 1.48,'format': '%.2f'},\n",
    "    'Td': {'value': 999,'format': '%.2f'}, # 1.54\n",
    "    'Beta': {'value': 1.67,'format': '%.2f'},\n",
    "    'lambda1': {'value': 0.18,'format': '%.3f'},\n",
    "    'lambda2': {'value': -0.1,'format': '%.2f'},\n",
    "    'channel_vo': {'value': 0.25,'format': '%.3f'},} \n",
    "Order = ['v_r','a_r','a','b','c','d','m1','ki_fac',\n",
    "    'TopDepth','NoFlow','Td','Beta','lambda1', 'lambda2','channel_vo']\n",
    "for name in wat.keys():\n",
    "    wat[name].set_parameters(DictP=Params, DicOrder=Order)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Setup the virtually distributed parameters of both cases."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T21:34:59.254895Z",
     "start_time": "2020-07-22T21:34:50.574666Z"
    }
   },
   "outputs": [],
   "source": [
    "#setupo the distributed virtual parameters of both cases for the different levels.\n",
    "for name in wat.keys():\n",
    "    for i in [4,5,6]:\n",
    "        # setup the parameters for each watershed.\n",
    "        wat[name].Table.at[wat[name].Table.index, 'channel_vo'] = wat[name].Table['vo'+str(i)].values\n",
    "        wat[name].Table.at[wat[name].Table.index, 'lambda1'] = wat[name].Table['la'+str(i)].values\n",
    "        \n",
    "        #Write the param file to run HLM\n",
    "        wat[name].write_prm(names[name]['path']+'Tparam_o'+str(i)+'.prm', \n",
    "            extraNames=Order,\n",
    "            extraFormats=[Params[k]['format'] for k in Order])\n",
    "        "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Write the global files."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-08T17:47:57.533495Z",
     "start_time": "2020-07-08T17:47:56.910614Z"
    }
   },
   "outputs": [],
   "source": [
    "for name in wat.keys():\n",
    "    path = names[name]['path']    \n",
    "    for y in range(2002,2019):\n",
    "        #Writes the global files and run files    \n",
    "        for k in [4,5,6]:\n",
    "            #Set the name of the global file\n",
    "            global_name = 'global_'+str(k)+'_'+str(y)+'.gbl'\n",
    "            #Writes the global file\n",
    "            wat[name].write_Global(path + global_name,model_uid=608,\n",
    "                date1=str(y)+'-03-01 01:00', \n",
    "                date2=str(y)+'-12-31 01:00',\n",
    "                rvrFile='/Users/nicolas/Parameter_identifiability/'+path+'topology.rvr', \n",
    "                rvrLink='',\n",
    "                prmFile='/Users/nicolas/Parameter_identifiability/'+path+'param_o'+str(k)+'.prm',\n",
    "                initialFile='/Users/nicolas/Parameter_identifiability/data/for_hlm/initial.uini',\n",
    "                initialType = 1,\n",
    "                rainPath = '/Dedicated/IFC/data_bin/stage4/st4_'+str(y)+'/iowa_',\n",
    "                controlFile='/Users/nicolas/Parameter_identifiability/'+path+'control_vr5.sav',\n",
    "                nComponents = 1,\n",
    "                Components = [0],\n",
    "                datResults='out_'+str(k)+'_'+str(y)+'.dat',\n",
    "                evpFile='/Users/nicolas/Parameter_identifiability/data/for_hlm/evap.mon')\n",
    "            #Updates the dictionary with the parameters to setup the run\n",
    "            Proc={global_name:{'nproc': 14, 'secondplane': False}}\n",
    "            #Writes the runfile with all the comands\n",
    "            wat[name].write_runfile(path + 'run_'+str(k)+'_'+str(y)+'.sh', \n",
    "                Proc, jobName='o'+str(k)+'_'+str(y), nCores= 14) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Run the prun level 4 setup to save snapshots to then run just relevant events "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T23:40:02.365100Z",
     "start_time": "2020-07-22T23:39:58.882884Z"
    }
   },
   "outputs": [],
   "source": [
    "for name in wat.keys():\n",
    "    path = names[name]['path']    \n",
    "    for y in range(2002,2019):\n",
    "        #Writes the global files and run files    \n",
    "        for k in [4,]:\n",
    "            #Set the name of the global file\n",
    "            global_name = 'global_'+str(k)+'_'+str(y)+'.gbl'\n",
    "            #Writes the global file\n",
    "            wat[name].write_Global(path + global_name,model_uid=608,\n",
    "                date1=str(y)+'-03-01 01:00', \n",
    "                date2=str(y)+'-12-31 01:00',\n",
    "                rvrFile='/Users/nicolas/Parameter_identifiability/'+path+'topology.rvr', \n",
    "                rvrLink='',\n",
    "                prmFile='/Users/nicolas/Parameter_identifiability/'+path+'param_o'+str(k)+'.prm',\n",
    "                initialFile='/Users/nicolas/Parameter_identifiability/data/for_hlm/initial.uini',\n",
    "                initialType = 1,\n",
    "                rainPath = '/Dedicated/IFC/data_bin/stage4/st4_'+str(y)+'/iowa_',\n",
    "                controlFile='/Users/nicolas/Parameter_identifiability/'+path+'control_vr5.sav',\n",
    "                nComponents = 1,\n",
    "                Components = [0],\n",
    "                datResults='out_'+str(k)+'_'+str(y)+'.dat',\n",
    "                snapType = 4,\n",
    "                snapTime = 1440,\n",
    "                snapPath = '/Users/nicolas/Parameter_identifiability/data/for_hlm/south_skunk/snapshots/p4.h5',\n",
    "                evpFile='/Users/nicolas/Parameter_identifiability/data/for_hlm/evap.mon')\n",
    "            #Updates the dictionary with the parameters to setup the run\n",
    "            Proc={global_name:{'nproc': 14, 'secondplane': False}}\n",
    "            #Writes the runfile with all the comands\n",
    "            wat[name].write_runfile(path + 'run_'+str(k)+'_'+str(y)+'.sh', \n",
    "                Proc, jobName='o'+str(k)+'_'+str(y), nCores= 14) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random watersheds of order 5.\n",
    "\n",
    "Here we generate 10 random watersheds that distribute the parameters at the prun level 5. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:06:57.040335Z",
     "start_time": "2020-07-22T22:06:57.037851Z"
    }
   },
   "outputs": [],
   "source": [
    "y = 815"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:00:05.320269Z",
     "start_time": "2020-07-22T22:00:05.252089Z"
    }
   },
   "outputs": [],
   "source": [
    "path = names['south']['path']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "years = [2008,2010,2013,2016,2018]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:18:36.652118Z",
     "start_time": "2020-07-22T22:18:34.039728Z"
    }
   },
   "outputs": [],
   "source": [
    "#setupo the distributed virtual parameters of both cases for the different levels.\n",
    "for name in wat.keys():\n",
    "    for i in range(1,2):\n",
    "        # Create random parameters for the prun 5 level \n",
    "        f_vo = cascade_values_sync(wat['south'].Table, 'vo', 5, [0.15,0.3,0.22], 0.2)\n",
    "        f_l1 = cascade_values_sync(wat['south'].Table, 'la', 5, [0.18,0.23,0.25], 0.2)\n",
    "        \n",
    "        # setup the parameters for each watershed.\n",
    "        wat['south'].Table.at[wat['south'].Table.index, 'channel_vo'] = wat['south'].Table['vo5'].values\n",
    "        wat['south'].Table.at[wat['south'].Table.index, 'lambda1'] = wat['south'].Table['la5'].values\n",
    "        \n",
    "        #Write the param file to run HLM\n",
    "        wat['south'].write_prm(names['south']['path']+'param_s'+str(i)+'.prm', \n",
    "            extraNames=Order,\n",
    "            extraFormats=[Params[k]['format'] for k in Order])\n",
    "        \n",
    "        #Writes the global and the runfile for each experiment\n",
    "        #for y in range(2002,2019):            \n",
    "        #Set the name of the global file\n",
    "        global_name = 'global_s'+str(i)+'_'+str(y)+'.gbl'\n",
    "        #Writes the global file\n",
    "        wat[name].write_Global(path + global_name,model_uid=608,\n",
    "            date1='2016'+'-03-01 01:00', \n",
    "            date2='2016'+'-12-31 01:00',\n",
    "            rvrFile='/Users/nicolas/Parameter_identifiability/'+path+'topology.rvr', \n",
    "            rvrLink='',\n",
    "            prmFile='/Users/nicolas/Parameter_identifiability/'+path+'param_s'+str(i)+'.prm',\n",
    "            initialFile='/Users/nicolas/Parameter_identifiability/data/for_hlm/initial.uini',\n",
    "            initialType = 1,\n",
    "            rainPath = '/Dedicated/IFC/data_bin/stage4/st4_'+str(y)+'/iowa_',\n",
    "            controlFile='/Users/nicolas/Parameter_identifiability/'+path+'control_vr5.sav',\n",
    "            nComponents = 1,\n",
    "            Components = [0],\n",
    "            datResults='out_'+str(i)+'_'+str(y)+'.h5',\n",
    "            datType = 5,\n",
    "            evpFile='/Users/nicolas/Parameter_identifiability/data/for_hlm/evap.mon')\n",
    "        #Updates the dictionary with the parameters to setup the run\n",
    "        Proc={global_name:{'nproc': 14, 'secondplane': False}}\n",
    "        #Writes the runfile with all the comands\n",
    "        wat[name].write_runfile(path + 'run_s'+str(i)+'_'+str(y)+'.sh', \n",
    "            Proc, jobName='o'+str(i)+'_'+str(y), nCores= 14) \n",
    "\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:19:50.304078Z",
     "start_time": "2020-07-22T22:19:50.274788Z"
    }
   },
   "outputs": [],
   "source": [
    "q = pd.read_hdf('data/for_hlm/south_skunk/out_1_815.h5',)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:20:34.836630Z",
     "start_time": "2020-07-22T22:20:34.719595Z"
    }
   },
   "outputs": [],
   "source": [
    "import h5py\n",
    "import h5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:20:59.870390Z",
     "start_time": "2020-07-22T22:20:59.866053Z"
    }
   },
   "outputs": [],
   "source": [
    "h = h5py.File('data/for_hlm/south_skunk/out_1_815.h5','r')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:21:51.984481Z",
     "start_time": "2020-07-22T22:21:51.980385Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KeysView(<HDF5 file \"out_1_815.h5\" (mode r)>)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "h.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:23:41.448035Z",
     "start_time": "2020-07-22T22:23:41.444334Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KeysView(<HDF5 file \"out_1_815.h5\" (mode r)>)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "h.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:24:51.104691Z",
     "start_time": "2020-07-22T22:24:51.102258Z"
    }
   },
   "outputs": [],
   "source": [
    "g = h['/']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:29:34.028152Z",
     "start_time": "2020-07-22T22:29:34.019207Z"
    }
   },
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "'method' object is not subscriptable",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-59-a5af60bdce5d>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mo\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mh\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'outputs'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m: 'method' object is not subscriptable"
     ]
    }
   ],
   "source": [
    "o = h.get['outputs']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-22T22:26:02.281413Z",
     "start_time": "2020-07-22T22:26:02.255198Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([(1.0000000e-06,), (5.6097266e-04,), (1.2954471e-03,), ...,\n",
       "       (5.4735615e-08,), (5.4687831e-08,), (5.4640100e-08,)],\n",
       "      dtype=[('State0', '<f4')])"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "o.value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "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": "3.7.0"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "384px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}