helper_model_arbitrage_report.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import ipywidgets as widgets\n",
    "from ipywidgets import VBox, HBox\n",
    "import plotly.graph_objs as go\n",
    "import numpy as np\n",
    "\n",
    "\n",
    "\n",
    "class MarketArbitrageReport:\n",
    "    \n",
    "    \n",
    "    def __init__(self, model_arbitrage_df):\n",
    "        self.df=model_arbitrage_df\n",
    "        self.df['Adjusted_call_butterfly']=self.df['Call_butterfly']*self.df['Multiplier']\n",
    "        self.df['Adjusted_put_butterfly']=self.df['Put_butterfly']*self.df['Multiplier']\n",
    "        self.tolerance=1.0e-10\n",
    "        self.tolerance_criteria=VBox([widgets.HTMLMath(value=r'Tolerance level = accepted numerical error')])\n",
    "\n",
    "        \n",
    "    def aggregated_results(self):\n",
    "        self.agg_df=pd.DataFrame({'min_call_difference':self.df.groupby(['DATE','CURRENCY','TENOR', 'EXPIRY'])[['Call_difference']].apply(lambda df: pd.eval('-1*df.Call_difference').min())})\n",
    "        self.agg_df['min_put_difference']=pd.DataFrame({'min_put_difference':self.df.groupby(['DATE','CURRENCY','TENOR', 'EXPIRY'])[['Put_difference']].apply(lambda df: pd.eval('df.Put_difference').min())})\n",
    "        self.agg_df['min_call_butterfly']=pd.DataFrame({'min_call_butterfly':self.df.groupby(['DATE','CURRENCY','TENOR', 'EXPIRY'])[['Adjusted_call_butterfly']].apply(lambda df: pd.eval('df.Adjusted_call_butterfly').min())})\n",
    "        self.agg_df['min_put_butterfly']=pd.DataFrame({'min_put_butterfly':self.df.groupby(['DATE','CURRENCY','TENOR', 'EXPIRY'])[['Adjusted_put_butterfly']].apply(lambda df: pd.eval('df.Adjusted_put_butterfly').min())})\n",
    "        \n",
    "        self.agg_df=self.agg_df.reset_index(drop=False)\n",
    "        self.agg_df['TENOR'] = pd.Categorical(self.agg_df['TENOR'], categories=[f'{i}Y' for i in range(1,50)])\n",
    "        self.agg_df['EXPIRY'] = pd.Categorical(self.agg_df['EXPIRY'], categories=[f'{i}M' for i in range(1,12)]+['1Y']+[f'{i}M' for i in range(13,24)]+[f'{i}Y' for i in range(2,50)])\n",
    "        self.agg_df=self.agg_df.sort_values(by=['DATE','CURRENCY','TENOR', 'EXPIRY'])\n",
    "        \n",
    "        return self.agg_df\n",
    "    \n",
    "    \n",
    "    def aggregated_results_heatmap(self, currency, date):\n",
    "        call_difference_df = self.agg_df.loc[(self.agg_df['DATE']==date) & (self.agg_df['CURRENCY']==currency)].set_index(['TENOR','EXPIRY'])['min_call_difference'].unstack()\n",
    "        call_difference_df = call_difference_df.reindex(call_difference_df.columns.sort_values(), axis=1)\n",
    "        call_difference_df = call_difference_df.reindex(call_difference_df.index.sort_values(), axis=0)\n",
    "        x=call_difference_df.columns.tolist()\n",
    "        y=call_difference_df.index.tolist()\n",
    "        call_difference=call_difference_df.values\n",
    "        \n",
    "        put_difference_df = self.agg_df.loc[(self.agg_df['DATE']==date) & (self.agg_df['CURRENCY']==currency)].set_index(['TENOR','EXPIRY'])['min_put_difference'].unstack()\n",
    "        put_difference_df = put_difference_df.reindex(put_difference_df.columns.sort_values(), axis=1)\n",
    "        put_difference_df = put_difference_df.reindex(put_difference_df.index.sort_values(), axis=0)\n",
    "        put_difference=put_difference_df.values\n",
    "        \n",
    "        call_butterfly_df = self.agg_df.loc[(self.agg_df['DATE']==date) & (self.agg_df['CURRENCY']==currency)].set_index(['TENOR','EXPIRY'])['min_call_butterfly'].unstack()\n",
    "        call_butterfly_df = call_butterfly_df.reindex(call_butterfly_df.columns.sort_values(), axis=1)\n",
    "        call_butterfly_df = call_butterfly_df.reindex(call_butterfly_df.index.sort_values(), axis=0)\n",
    "        call_butterfly=call_butterfly_df.values\n",
    "        \n",
    "        put_butterfly_df = self.agg_df.loc[(self.agg_df['DATE']==date) & (self.agg_df['CURRENCY']==currency)].set_index(['TENOR','EXPIRY'])['min_put_butterfly'].unstack()\n",
    "        put_butterfly_df = put_butterfly_df.reindex(put_butterfly_df.columns.sort_values(), axis=1)\n",
    "        put_butterfly_df = put_butterfly_df.reindex(put_butterfly_df.index.sort_values(), axis=0)\n",
    "        put_butterfly=put_butterfly_df.values\n",
    "        \n",
    "        z=[[np.nan if np.isnan(call_difference[i][j]) else ((call_difference[i][j]<-self.tolerance) or (put_difference[i][j]<-self.tolerance) \n",
    "                                                            or (call_butterfly[i][j]<-self.tolerance/(2**0.5)) or \n",
    "                                                            (put_butterfly[i][j]<-self.tolerance/(2**0.5)))*1 for j in range(len(call_difference[:][i]))] for i in range(len(call_difference))]\n",
    "        \n",
    "        smiles_total=np.count_nonzero(~np.isnan(z))\n",
    "        smiles_failed=int(np.nansum(z))\n",
    "    \n",
    "        self.fig=go.FigureWidget([go.Heatmap(z=z,x=x,y=y, name = '',xgap = 1, ygap = 1,zmin=0,zmax=1,\n",
    "                                             colorbar = dict(nticks=2, \n",
    "                                                             tickmode=\"array\", \n",
    "                                                             tickvals=[0,1], \n",
    "                                                             ticktext=[f'No arbitrage: {(smiles_total-smiles_failed)}/{smiles_total}',\n",
    "                                                                       f'Arbitrage: {smiles_failed}/{smiles_total}'], \n",
    "                                                             tickfont={\"size\":15}), colorscale=[[0.0, \"rgb(0,0,200)\"], \n",
    "                                                                                                [1.0, \"rgb(200,0,0)\"]])])\n",
    "        self.fig['layout'].update(height=600, width=750, title='Success/failure map')\n",
    "        self.fig.update_yaxes(title_text='Tenor')\n",
    "        self.fig.update_xaxes(title_text='Expiry')\n",
    "        \n",
    "        self.fig.data[0].on_click(self.select_smile_on_heatmap)\n",
    "        \n",
    "        return self.fig  \n",
    "    \n",
    "    \n",
    "    def plot_selected_smile(self, currency, date, tenor, expiry, value):\n",
    "        \n",
    "        selected_smile=self.df.loc[(self.df['CURRENCY']==currency) & (self.df['DATE']==date) & (self.df['TENOR']==tenor) & (self.df['EXPIRY']==expiry)].sort_values(by=['MONEYNESS'])\n",
    "        strike=selected_smile['STRIKE'].tolist()\n",
    "        if value=='difference':\n",
    "            name1=r'$C(K_{i})-C(K_{i+1})$'\n",
    "            name2=r'$P(K_{i+1})-P(K_{i})$'\n",
    "            mult=1.\n",
    "            title=r'$C(K_{i})-C(K_{i+1}), \\, P(K_{i+1})-P(K_{i})$'\n",
    "            y1=(-selected_smile['Call_difference']).tolist()\n",
    "            y2=selected_smile['Put_difference'].tolist()\n",
    "            \n",
    "        elif value=='butterfly':\n",
    "            name1='Call butterfly'\n",
    "            name2='Put butterfly'\n",
    "            mult=1/(2**0.5)\n",
    "            title='Adjusted call/put butterflies'\n",
    "            y1=selected_smile['Adjusted_call_butterfly']\n",
    "            y2=selected_smile['Adjusted_put_butterfly']\n",
    "        \n",
    "        data=[go.Scatter(x=strike, y=y1, name=name1,mode='lines', line=dict(color='blue', width=5)), \n",
    "              go.Scatter(x=strike, y=y2, name=name2,mode='lines', line=dict(color='orange', width=5)), \n",
    "              go.Scatter(x=strike, y=[-mult*self.tolerance]*len(strike), name='Threshold',mode='lines', line=dict(color='red', width=5, dash='dash' ))]\n",
    "    \n",
    "        fig = go.FigureWidget(data=data)\n",
    "        fig['layout'].update(height=300, width=750, title=title)\n",
    "        fig.update_xaxes(title_text='Strike, %')\n",
    "\n",
    "        return fig\n",
    "\n",
    "     \n",
    "    def display_plots(self):\n",
    "        self.agg_df=self.aggregated_results()\n",
    "        self.currencies=self.agg_df['CURRENCY'].unique().tolist()\n",
    "        tenor_sorting=[f'{i}Y' for i in range(1,50)]\n",
    "        expiry_sorting=[f'{i}M' for i in range(1,12)]+['1Y']+[f'{i}M' for i in range(13,24)]+[f'{i}Y' for i in range(2,50)]\n",
    "        self.active_tenor='5Y'\n",
    "        self.active_expiry='5Y'\n",
    "\n",
    "        \n",
    "        # Tolerance slider      \n",
    "        self.slider_wdg_tolerance=widgets.FloatLogSlider(value=self.tolerance,base=10,min=-20, max=-4, step=1, description='Tolerance', disabled=False, \n",
    "                                                         continuous_update=False,orientation='horizontal',readout=True)\n",
    "        self.slider_wdg_tolerance.observe(lambda x: self.update_tolerance_level(x['new']), names = 'value')\n",
    "\n",
    "        # Currency tabs\n",
    "        tab_list=[]\n",
    "        for currency in self.currencies:\n",
    "            dates=pd.to_datetime(self.agg_df.loc[self.agg_df['CURRENCY']==currency]['DATE'].unique().tolist())\n",
    "            tenors=sorted(self.agg_df.loc[self.agg_df['CURRENCY']==currency]['TENOR'].unique().tolist(), key=lambda x: tenor_sorting.index(x))\n",
    "            expiries=sorted(self.agg_df.loc[self.agg_df['CURRENCY']==currency]['EXPIRY'].unique().tolist(), key=lambda x: expiry_sorting.index(x))\n",
    "            tab_list.append(VBox([widgets.SelectionSlider(description='Date', options=dates, continuous_update=False, layout={'width': '90%'}),\n",
    "                                  self.aggregated_results_heatmap(currency,dates[0]),\n",
    "                                  HBox([widgets.Dropdown(description='Tenor', options=tenors, value=self.active_tenor), widgets.Dropdown(description='Expiry', options=expiries, value=self.active_expiry)]),\n",
    "                                  self.plot_selected_smile(currency, dates[0],  self.active_tenor,  self.active_expiry, 'difference'), \n",
    "                                  self.plot_selected_smile(currency, dates[0],  self.active_tenor,  self.active_expiry, 'butterfly')]))\n",
    "        \n",
    "        self.tabs = widgets.Tab(children=tab_list)     \n",
    "            \n",
    "        for currency in self.currencies:\n",
    "            self.tabs.set_title(self.currencies.index(currency), currency)\n",
    "            self.tabs.children[self.currencies.index(currency)].children[0].observe(lambda x, currency=currency: self.update_date(currency,x['new']), names = 'value')\n",
    "            self.tabs.children[self.currencies.index(currency)].children[2].children[0].observe(lambda x, currency=currency: self.update_tenor(currency,x['new']), names = 'value')\n",
    "            self.tabs.children[self.currencies.index(currency)].children[2].children[1].observe(lambda x, currency=currency: self.update_expiry(currency,x['new']), names = 'value')\n",
    "        \n",
    "        self.report_page=HBox([self.tabs, VBox([self.slider_wdg_tolerance, self.tolerance_criteria])])\n",
    "        \n",
    "        return self.report_page\n",
    "        \n",
    "        \n",
    "    def update_date(self, currency, value):\n",
    "        tab_index=self.currencies.index(currency)\n",
    "        date=value\n",
    "        tenor=self.tabs.children[tab_index].children[2].children[0].value\n",
    "        expiry=self.tabs.children[tab_index].children[2].children[1].value\n",
    "        self.update_plots(currency, date, tenor, expiry, 1)\n",
    "           \n",
    "        \n",
    "    def update_tolerance_level(self, value):\n",
    "        self.tolerance=value\n",
    "        for currency in self.currencies:\n",
    "            date=self.tabs.children[self.currencies.index(currency)].children[0].value\n",
    "            tenor=self.tabs.children[self.currencies.index(currency)].children[2].children[0].value\n",
    "            expiry=self.tabs.children[self.currencies.index(currency)].children[2].children[1].value\n",
    "            self.update_plots(currency, date, tenor, expiry, 0)\n",
    "    \n",
    "    \n",
    "    def update_tenor(self, currency, value):\n",
    "        tab_index=self.currencies.index(currency)\n",
    "        date=self.tabs.children[tab_index].children[0].value\n",
    "        tenor=value\n",
    "        expiry=self.tabs.children[tab_index].children[2].children[1].value\n",
    "        self.update_plots(currency, date, tenor, expiry,2)\n",
    "        \n",
    "        \n",
    "    def update_expiry(self, currency, value):\n",
    "        tab_index=self.currencies.index(currency)\n",
    "        date=self.tabs.children[tab_index].children[0].value\n",
    "        tenor=self.tabs.children[tab_index].children[2].children[0].value\n",
    "        expiry=value\n",
    "        self.update_plots(currency, date, tenor, expiry,2)\n",
    "    \n",
    "    \n",
    "    def select_smile_on_heatmap(self, trace, points, state):\n",
    "        expiry=points.xs[0]\n",
    "        tenor=points.ys[0]\n",
    "        tab_index=self.tabs.selected_index\n",
    "        self.tabs.children[tab_index].children[2].children[0].value, self.tabs.children[tab_index].children[2].children[1].value=tenor, expiry      \n",
    "    \n",
    "    \n",
    "    def update_plots(self, currency, date, tenor, expiry, flag):\n",
    "        tab_index=self.currencies.index(currency)\n",
    "        \n",
    "        if flag==0:\n",
    "\n",
    "            self.tabs.children[tab_index].children[1].data[0].z=self.aggregated_results_heatmap(currency, date).data[0].z\n",
    "            self.tabs.children[tab_index].children[1].data[0].colorbar=self.aggregated_results_heatmap(currency, date).data[0].colorbar\n",
    "        \n",
    "            self.tabs.children[tab_index].children[3].data[2].y=self.plot_selected_smile(currency, date, tenor, expiry, 'difference').data[2].y\n",
    "            self.tabs.children[tab_index].children[4].data[2].y=self.plot_selected_smile(currency, date, tenor, expiry, 'butterfly').data[2].y\n",
    "\n",
    "                \n",
    "        elif flag==1:\n",
    "\n",
    "            self.tabs.children[tab_index].children[1].data[0].x=self.aggregated_results_heatmap(currency, date).data[0].x\n",
    "            self.tabs.children[tab_index].children[1].data[0].y=self.aggregated_results_heatmap(currency, date).data[0].y\n",
    "            self.tabs.children[tab_index].children[1].data[0].z=self.aggregated_results_heatmap(currency, date).data[0].z\n",
    "            self.tabs.children[tab_index].children[1].data[0].colorbar=self.aggregated_results_heatmap(currency, date).data[0].colorbar\n",
    "\n",
    "            self.tabs.children[tab_index].children[3].data=[]\n",
    "            for j in range(3):\n",
    "                self.tabs.children[tab_index].children[3].add_trace(self.plot_selected_smile(currency, date, tenor, expiry, 'difference').data[j])\n",
    "            self.tabs.children[tab_index].children[4].data=[]\n",
    "            for j in range(3):\n",
    "                self.tabs.children[tab_index].children[4].add_trace(self.plot_selected_smile(currency, date, tenor, expiry, 'butterfly').data[j])\n",
    "\n",
    "                \n",
    "        elif flag==2:\n",
    "\n",
    "            self.tabs.children[tab_index].children[3].data=[]\n",
    "            for j in range(3):\n",
    "                self.tabs.children[tab_index].children[3].add_trace(self.plot_selected_smile(currency, date, tenor, expiry, 'difference').data[j])\n",
    "            self.tabs.children[tab_index].children[4].data=[]\n",
    "            for j in range(3):\n",
    "                self.tabs.children[tab_index].children[4].add_trace(self.plot_selected_smile(currency, date, tenor, expiry, 'butterfly').data[j])\n",
    "\n",
    "                \n",
    "\n",
    "    "
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}