{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ".. _models:\n",
    " \n",
    ".. title:: Models\n",
    " \n",
    "#### Models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " Created on Sat May 1 15:12:38 2019  \n",
    " @author: 'Semeon Risom'  \n",
    " @email: 'semeon.risom@gmail.com'  \n",
    " @url: 'https://semeon.io/d/mdl'  \n",
    " @purpose: 'Build models for analysis.'  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ##### Resources:\n",
    " - ANOVA:  \n",
    "    - https://m-clark.github.io/docs/mixedModels/anovamixed.html  \n",
    "    - http://dwoll.de/rexrepos/posts/anovaMixed.html  \n",
    "    - https://rpsychologist.com/r-guide-longitudinal-lme-lmer  \n",
    "    - https://stats.stackexchange.com/questions/247582/repeated-measures-anova-in-r-errorsubject-vs-errorsubject-day  \n",
    "    - https://cran.r-project.org/web/packages/afex/vignettes/afex_anova_example.html#post-hoc-contrasts-and-plotting  \n",
    "    - http://www.let.rug.nl/nerbonne/teach/rema-stats-meth-seminar/presentations/Wieling-MixedModels-2011.pdf  \n",
    " - negative-binomial poisson:  \n",
    "   - https://stackoverflow.com/questions/47686227/poisson-regression-in-statsmodels-and-r\n",
    "   - https://tsmatz.wordpress.com/2017/08/30/glm-regression-logistic-poisson-gaussian-gamma-tutorial-with-r/\n",
    "   - https://stats.stackexchange.com/questions/311556/help-interpreting-count-data-glmm-using-lme4-glmer-and-glmer-nb-negative-binom\n",
    " - logistic regression:\n",
    "   - https://stats.idre.ucla.edu/other/mult-pkg/whatstat/\n",
    "   - https://stats.idre.ucla.edu/r/dae/mixed-effects-logistic-regression/\n",
    "   - https://tsmatz.wordpress.com/2017/08/30/glm-regression-logistic-poisson-gaussian-gamma-tutorial-with-r/\n",
    "   - https://data.library.virginia.edu/diagnostic-plots/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ##### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#----local\n",
    "from imhr import plot, processing, raw, redcap\n",
    "import imhr.model as model\n",
    "import imhr.settings as settings\n",
    "\n",
    "#----check if required libraries are available\n",
    "is_library = False\n",
    "if is_library:\n",
    "    settings.library()\n",
    "pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ##### Start"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "processing = processing(config, filters)\n",
    "console = settings.console"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ##### Analysis of Varience: Gaze and Dwell Time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "effects = {}\n",
    "\n",
    "#----load config\n",
    "# p = path_['output'] + \"/analysis/config.json\"\n",
    "# with open(p) as f:\n",
    "#     config_ = json.loads(f.read())\n",
    "config_ = config\n",
    "console('fix config', 'red')   \n",
    "\n",
    "#----exclude\n",
    "exclude = config_['metadata']['subjects']['exclude']\n",
    "\n",
    "#----load data\n",
    "p = path_['output'] + \"/analysis/final_data.csv\"\n",
    "df_ = pd.read_csv(p, float_precision='high')\n",
    "\n",
    "#-----parameters\n",
    "# dependent variable\n",
    "y = 'dwell_time'\n",
    "# main effects\n",
    "effects['main'] = {\n",
    "    'cesd_group': 'categorical',\n",
    "    'aoi': 'categorical',\n",
    "    'trialType': 'categorical'\n",
    "}\n",
    "# random effects\n",
    "effects['random'] = {\n",
    "    'participant': 'categorical'\n",
    "}\n",
    "# formula\n",
    "f = \"%s ~ cesd_group + aoi + trialType + (1|participant)\"%(y)\n",
    "\n",
    "#----save data for access by R and for calculating dwell time\n",
    "csv = \"dwell_data.csv\"\n",
    "p = path_['output'] + \"/analysis/html/model/anova/\"\n",
    "\n",
    "#-----calculate dwell time using multiprocessing\n",
    "# use __name__ to protect main module\n",
    "df_dwell, error_dwell = processing.dwell(df=df_, cores=7) if __name__ == '__main__' else None\n",
    "\n",
    "#----normalize dwell_time for comparison between iaps and pofa\n",
    "df_dwell['dwell_time'] = df_dwell.apply(lambda x: (x['dwell_time']/4500) \n",
    "if (x['trialType'] == 'iaps') else (x['dwell_time']/3000), axis=1)\n",
    "\n",
    "#-----exclude participants, group by subject:trialType:aoi\n",
    "# exclude participants \n",
    "df_dwell = df_dwell[~df_dwell['participant'].isin(exclude)]\n",
    "# groupby\n",
    "df_dwell = df_dwell.groupby(['participant','cesd_group','trialType','aoi'])['dwell_time'].mean().reset_index()\n",
    "\n",
    "#-----run\n",
    "anova_, anova_result, anova_r, html = model.anova(config=config_, df=df_dwell, y=y, f=f, csv=csv, path=p, effects=effects)\n",
    "\n",
    "#-----delete\n",
    "del y, f, csv, p"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ##### Linear Mixed Model Regression: Stimulus and Dotloc Onset Error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "effects = {}\n",
    "\n",
    "# #----load config\n",
    "# p = path_['output'] + \"/analysis/config.json\"\n",
    "# with open(p) as f:\n",
    "#     config_ = json.loads(f.read())\n",
    "config_ = config\n",
    "console('fix config', 'red')   \n",
    "\n",
    "#----load data\n",
    "p = path_['output'] + \"/analysis/error.csv\"\n",
    "#df_error = pd.read_csv(p_error, float_precision='high')\n",
    "df_ = pd.read_csv(p, float_precision='high')\n",
    "\n",
    "#----parameters            \n",
    "# dependent variable\n",
    "y = ['diff_stim','diff_dotloc'] #build models for each IV in list\n",
    "# fixed effects\n",
    "effects['fixed'] = {\n",
    "    'os': 'categorical',\n",
    "    'trialType': 'categorical',\n",
    "    'TrialNum': 'factorial'\n",
    "}\n",
    "# random effects\n",
    "effects['random'] = {\n",
    "    'TrialNum': 'factorial',\n",
    "    'participant': 'factorial',\n",
    "}\n",
    "\n",
    "#----save data for access by R and for calculating dwell time\n",
    "csv = \"onset_data.csv\"\n",
    "\n",
    "#----run model for each IV\n",
    "for _y in y:\n",
    "    # path\n",
    "    p = path_['output'] + \"/analysis/html/model/lmer/\"\n",
    "    # formula\n",
    "    f = \"sqrt(%s) ~ os + trialType + TrialNum + (1+TrialNum|participant)\"%(_y)\n",
    "    # run\n",
    "    lmer_, lmer_result, lmer_r, html = model.lmer(config=config_, df=df_, y=_y, f=f, exclude=exclude, csv=csv, path=p, effects=effects)\n",
    "\n",
    "#-----delete\n",
    "del y, _y, f, csv, p"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ##### Analysis of Varience: Gaze and Dotprobe bias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "effects = {}\n",
    "\n",
    "# #----load config\n",
    "# p = path_['output'] + \"/analysis/config.json\"\n",
    "# with open(p) as f:\n",
    "#     config_ = json.loads(f.read())\n",
    "config_ = config\n",
    "console('fix config', 'red')   \n",
    "#----load data\n",
    "p = path_['output'] + \"/analysis/final_data.csv\"\n",
    "df_ = pd.read_csv(p, float_precision='high')\n",
    "\n",
    "#-----exclude participants, group by subject\n",
    "# exclude participants\n",
    "df_ = df_[~df_['participant'].isin(exclude)]\n",
    "\n",
    "# groupby\n",
    "df_ = df_.loc[df_['nested'] == 'trialType']\n",
    "\n",
    "#----parameters\n",
    "# dependent variable\n",
    "y = ['dp_bias','gaze_bias']\n",
    "# main effects\n",
    "effects['main'] = {\n",
    "    'cesd_group': 'categorical',\n",
    "    'trialType': 'categorical'\n",
    "}\n",
    "# random effects\n",
    "effects['random'] = {\n",
    "    'participant': 'categorical'\n",
    "}\n",
    "\n",
    "#----create function for each IV\n",
    "for _y in y:\n",
    "    console('Step: ANOVA (%s)'%(_y), 'red')\n",
    "    #----save data for access by R\n",
    "    csv = \"%s.csv\"%(_y)\n",
    "    #-----path\n",
    "    p = path_['output'] + \"/analysis/html/model/anova/\"\n",
    "    #-----formula\n",
    "    f = \"%s ~ cesd_group + trialType + (1|participant)\"%(_y)\n",
    "    #-----run\n",
    "    anova_, anova_result, anova_r, html = model.anova(config=config, df=df_, y=_y, f=f, csv=csv, path=p, effects=effects)"
   ]
  }
 ],
 "metadata": {
  "file_extension": ".py",
  "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": "3.7.3"
  },
  "mimetype": "text/x-python",
  "name": "python",
  "npconvert_exporter": "python",
  "pygments_lexer": "ipython3",
  "version": 3
 },
 "nbformat": 4,
 "nbformat_minor": 2
}