{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "445b60dd-9f3c-4655-85f2-bcd0285c10c5",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import huracanpy"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "58add8dc-7537-4512-b192-6594ba527eaf",
   "metadata": {},
   "source": [
    "# Embedded IBTrACS Subsets\n",
    "IBTrACS can be loaded in two ways using HuracanPy's `load` function:\n",
    "* Online: You can load the latest version of any IBTrACS subset using the load function, provided you are connected to internet\n",
    "* Offline: Your installation of huracanpy embeds parts of the IBTrACS database which can be loaded even if you are not connected to internet.\n",
    "\n",
    "\n",
    "Two offline subsets are currently available:\n",
    "* \"wmo\" contains the data provided in the \"wmo\" columns, which correspond to the data provided by the center\n",
    "      responsible for the area of a given point. (see https://community.wmo.int/en/tropical-cyclone-regional-bodies)\n",
    "      Note that within this dataset, wind units are not homogeneous: they are provided as collected from the\n",
    "      meteorological agencies, which means that they have different time-averaging for wind extrema. (default)\n",
    "* \"usa\" contains the data provided in the \"wmo\" columns, which is provided by the NHC or the JTWC.\n",
    "\n",
    "Loading these will raise a warning to remind you that these datasets are offline versions with caveats and some post-treatment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "219b4b1f-7224-4e22-bee0-0d0a04ac91f9",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# WMO subset\n",
    "ib_wmo = huracanpy.load(source=\"ibtracs\", ibtracs_online=False, ibtracs_subset=\"wmo\")\n",
    "huracanpy.plot.tracks(ib_wmo.lon, ib_wmo.lat, intensity_var=ib_wmo.wind)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b449b6e9-8c66-4cf1-b3dd-0d7cfd55d169",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": [
     "nbsphinx-thumbnail"
    ]
   },
   "outputs": [],
   "source": [
    "# USA subset\n",
    "ib_usa = huracanpy.load(source=\"ibtracs\", ibtracs_online=False, ibtracs_subset=\"usa\")\n",
    "huracanpy.plot.tracks(ib_usa.lon, ib_usa.lat, intensity_var=ib_usa.wind)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a96453f3-9ea4-493f-b421-215d8f0bb201",
   "metadata": {},
   "source": [
    "Both subsets currently cover the 1980-2022. The WMO subset contains more tracks, because\n",
    "the JTWC columns does not always contain data for all tracks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b9bb2fb8-cfce-4dd9-b1b0-2edfdcabf3ce",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"WMO\")\n",
    "print(ib_wmo.time.values.min(), ib_wmo.time.values.max())\n",
    "print(ib_wmo.track_id.hrcn.nunique(), \"tracks\", len(ib_wmo.record), \"points\\n\")\n",
    "\n",
    "print(\"JTWC\")\n",
    "print(ib_usa.time.values.min(), ib_usa.time.values.max())\n",
    "print(ib_usa.track_id.hrcn.nunique(), \"tracks\", len(ib_usa.record), \"points\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5fd7eabc-df10-4e57-89fd-6e33f8f8a997",
   "metadata": {},
   "source": [
    "One of the main differences among these two subsets is the way winds are reported: In WMO, the maximum winds as reported by the WMO agencies are provided. This is inhomogeneous: the USA report 1-minute sustained winds, CMA reports 3-minutes sustained winds, and most other centers report 10-minutes sustained winds. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "751d56de-1f2c-46ec-bdf9-451d8c0e7154",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Add basin data to ib_wmo\n",
    "ib_wmo = ib_wmo.hrcn.add_basin()\n",
    "# Match tracks between ib_wmo and ib_usa, then retrieve LMI\n",
    "m = huracanpy.assess.match([ib_wmo, ib_usa], names=[\"wmo\", \"usa\"])\n",
    "max_winds = m.join(\n",
    "    ib_wmo[[\"wind\"]].groupby(ib_wmo.track_id).max().to_dataframe(), on=\"id_wmo\"\n",
    ").join(\n",
    "    ib_usa[[\"wind\"]].groupby(ib_usa.track_id).max().to_dataframe(),\n",
    "    on=\"id_usa\",\n",
    "    lsuffix=\"_wmo\",\n",
    "    rsuffix=\"_usa\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bb31f486-51a4-4c49-b1ad-d67569ec9a6a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add basin with separate groupby\n",
    "max_winds = max_winds.join(\n",
    "    ib_wmo[[\"basin\"]].groupby(ib_wmo.track_id).first().to_dataframe(), on=\"id_wmo\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8cebc0c0-e0b8-4aab-8059-74c0bb57e3a9",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Plot difference between WMO and USA winds in each basin\n",
    "import seaborn as sns\n",
    "\n",
    "p = sns.displot(\n",
    "    data=max_winds,\n",
    "    x=\"wind_wmo\",\n",
    "    y=\"wind_usa\",\n",
    "    col=\"basin\",\n",
    "    col_wrap=3,\n",
    ")\n",
    "for ax in p.axes.flatten():\n",
    "    ax.plot([0, 175], [0, 175], color=\"k\", linestyle=\"--\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.12.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}