{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook prepare the catalogues that will be analysed by CIGALE for SED fitting and physical parameter estimation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import os\n",
    "os.environ['LOG_LEVEL'] = 'INFO'\n",
    "\n",
    "from astropy.table import Table\n",
    "\n",
    "from herschelhelp.filters import correct_galactic_extinction\n",
    "from herschelhelp.external import convert_table_for_cigale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "SUFFIX = '20180219'\n",
    "master_catalogue = Table.read(\"../../dmu32/dmu32_Lockman-SWIRE/data/Lockman-SWIRE_{}_cigale.fits\".format(SUFFIX))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4366298"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(master_catalogue)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Best sources\n",
    "\n",
    "Define a good far-IR measurement as:\n",
    "- an existing flux in the band;\n",
    "- the flag from XID+ must not be set;\n",
    "- the signal to noise ratio must be over 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "good = {}\n",
    "for band in ['pacs_green', 'pacs_red', 'spire_250', 'spire_350', 'spire_500']:\n",
    "    good[band] = (~np.isnan(master_catalogue['f_{}'.format(band)]) & \n",
    "                                   ~master_catalogue['flag_{}'.format(band)])\n",
    "    good[band][good[band]] &= (master_catalogue[good[band]]['f_{}'.format(band)] /\n",
    "                               master_catalogue[good[band]]['ferr_{}'.format(band)] >= 2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will keep only sources with at leat 2 good far-IR measurements (we may actually use less sources are not all may have a redshift)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "combined_good = np.sum(list(good.values()), axis=0) >= 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of good sources: 50661\n"
     ]
    }
   ],
   "source": [
    "print(\"Number of good sources: {}\".format(np.sum(combined_good)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Only sources with at least two optical and at least two near infrared detections\n",
    "optnir = ((master_catalogue['flag_optnir_det'] == 3) \n",
    "          | (master_catalogue['flag_optnir_det'] == 7))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Main catalogue for CIGALE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "best_catalogue = master_catalogue[combined_good].copy()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Correction for galactic extinction\n",
    "best_catalogue = correct_galactic_extinction(best_catalogue, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:herschelhelp.external:For 1817 sources, the band wfc_u should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 1054 sources, the band megacam_u should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 353 sources, the band megacam_g should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 366 sources, the band gpc1_g should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 333 sources, the band wfc_g should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 108 sources, the band gpc1_r should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 90 sources, the band wfc_r should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 91 sources, the band megacam_r should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 13 sources, the band gpc1_i should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 13 sources, the band megacam_i should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 13 sources, the band megacam_y should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 13 sources, the band wfc_i should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n"
     ]
    }
   ],
   "source": [
    "# Convertion to CIGALE format\n",
    "best_catalogue = convert_table_for_cigale(best_catalogue, inplace=True, remove_zerofluxes=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Band selection\n",
    "\n",
    "We want to use only one filter for similar bands. We define an order of preference and set to NaN the flux in the lower prefered bands when a prefered band is available. Some band may have a 0 flux, we set there values to NaN."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "u_bands = [\"megacam_u\", \"wfc_u\" ]\n",
    "g_bands = [\"megacam_g\", \"wfc_g\", \"gpc1_g\"]\n",
    "r_bands = [\"megacam_r\", \"wfc_r\", \"gpc1_r\"]\n",
    "i_bands = [\"megacam_i\", \"wfc_i\", \"gpc1_i\"]\n",
    "z_bands = [\"megacam_z\", \"wfc_z\", \"gpc1_z\"]\n",
    "y_bands = [\"megacam_y\",          \"gpc1_y\"]\n",
    "\n",
    "def remove_unneeded_fluxes(list_of_bands):\n",
    "    for band_idx, band in enumerate(list_of_bands[:-1]):\n",
    "        mask = ~np.isnan(best_catalogue[band])\n",
    "        for lower_band in list_of_bands[band_idx+1:]:\n",
    "            best_catalogue[lower_band][mask] = np.nan\n",
    "            best_catalogue[\"{}_err\".format(lower_band)][mask] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "remove_unneeded_fluxes(g_bands)\n",
    "remove_unneeded_fluxes(u_bands)\n",
    "remove_unneeded_fluxes(r_bands)\n",
    "remove_unneeded_fluxes(i_bands)\n",
    "remove_unneeded_fluxes(z_bands)\n",
    "remove_unneeded_fluxes(y_bands)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "best_catalogue.write(\"data_tmp/Lockman-SWIRE_cigale_best_extcor_{}.fits\".format(SUFFIX), overwrite=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# All optnir objects with photo-z"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of optnir sources with z-phot: 351369\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:herschelhelp.external:For 19737 sources, the band wfc_u should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 14668 sources, the band megacam_u should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 7379 sources, the band megacam_g should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 7488 sources, the band gpc1_g should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 7164 sources, the band wfc_g should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 2983 sources, the band gpc1_r should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 2505 sources, the band wfc_r should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 2549 sources, the band megacam_r should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 150 sources, the band gpc1_i should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 162 sources, the band megacam_i should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 155 sources, the band megacam_y should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 157 sources, the band wfc_i should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n"
     ]
    }
   ],
   "source": [
    "best_catalogue = master_catalogue[optnir].copy()\n",
    "best_catalogue = best_catalogue[~np.isnan(best_catalogue[\"redshift\"])]\n",
    "print(\"Number of optnir sources with z-phot: {}\".format(len(best_catalogue)))\n",
    "# Correction for galactic extinction\n",
    "best_catalogue = correct_galactic_extinction(best_catalogue, inplace=True)\n",
    "# Convertion to CIGALE format\n",
    "os.environ['LOG_LEVEL'] = 'INFO'\n",
    "best_catalogue = convert_table_for_cigale(best_catalogue, inplace=True, remove_zerofluxes=True)\n",
    "remove_unneeded_fluxes(g_bands)\n",
    "remove_unneeded_fluxes(u_bands)\n",
    "remove_unneeded_fluxes(r_bands)\n",
    "remove_unneeded_fluxes(i_bands)\n",
    "remove_unneeded_fluxes(z_bands)\n",
    "remove_unneeded_fluxes(y_bands)\n",
    "best_catalogue.write(\"data_tmp/Lockman-SWIRE_cigale_optnir_extcor_{}.fits\".format(SUFFIX), overwrite=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Catalogue using spectroscopic redshift"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "best_catalogue = master_catalogue[optnir].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "best_catalogue.remove_column(\"redshift\")\n",
    "best_catalogue[\"zspec\"].name = \"redshift\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "best_catalogue = best_catalogue[~np.isnan(best_catalogue[\"redshift\"])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of sources with z-spec: 3011\n"
     ]
    }
   ],
   "source": [
    "print(\"Number of sources with z-spec: {}\".format(len(best_catalogue)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Correction for galactic extinction\n",
    "best_catalogue = correct_galactic_extinction(best_catalogue, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:herschelhelp.external:For 141 sources, the band wfc_u should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 92 sources, the band megacam_u should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 27 sources, the band megacam_g should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 29 sources, the band gpc1_g should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 24 sources, the band wfc_g should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 8 sources, the band gpc1_r should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 8 sources, the band wfc_r should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 8 sources, the band megacam_r should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 2 sources, the band gpc1_i should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 2 sources, the band megacam_i should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 2 sources, the band megacam_y should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n",
      "INFO:herschelhelp.external:For 2 sources, the band wfc_i should not be used because it overlaps or is below the Lyman limit at the redshift of these sources. These fluxes were set to NaN.\n"
     ]
    }
   ],
   "source": [
    "# Convertion to CIGALE format\n",
    "os.environ['LOG_LEVEL'] = 'INFO'\n",
    "best_catalogue = convert_table_for_cigale(best_catalogue, inplace=True, remove_zerofluxes=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "remove_unneeded_fluxes(g_bands)\n",
    "remove_unneeded_fluxes(u_bands)\n",
    "remove_unneeded_fluxes(r_bands)\n",
    "remove_unneeded_fluxes(i_bands)\n",
    "remove_unneeded_fluxes(z_bands)\n",
    "remove_unneeded_fluxes(y_bands)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "best_catalogue.write(\"data_tmp/Lockman-SWIRE_cigale_optnir_extcor_zspec_{}.fits\".format(SUFFIX), overwrite=True)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (herschelhelp_internal)",
   "language": "python",
   "name": "helpint"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}