added weekdays plot

roocs · Aug 2, 2024 · 6795dfd · 6795dfd
1 parent c34eeb9
commit 6795dfd
Show file tree

Hide file tree

Showing 4 changed files with 80 additions and 112 deletions.
diff --git a/notebooks/dashboard.ipynb b/notebooks/dashboard.ipynb
@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "id": "59109053-8451-421a-930c-534bfe75e2f8",
    "metadata": {},
    "outputs": [],
@@ -12,7 +12,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 2,
    "id": "7384ef24-299e-42c9-8a7a-fc990a02ce14",
    "metadata": {},
    "outputs": [],
@@ -22,142 +22,62 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 3,
    "id": "1cd42832-c7fe-405a-b927-663d996fddff",
    "metadata": {},
    "outputs": [],
    "source": [
     "dash.load(\n",
-    "        \"http://rook3.cloud.dkrz.de:80/outputs/rook/34369610-d351-11eb-8f86-fa163e466023/wps_requests.csv\",\n",
+    "        \"http://rook8.cloud.dkrz.de:80/outputs/rook/ba7a8b60-50b6-11ef-a299-fa163e934c9b/wps_requests.csv\",\n",
     "        filter=\"orchestrate\",\n",
     "    )\n",
     "\n",
     "dash.load_downloads(\n",
-    "        \"http://rook3.cloud.dkrz.de:80/outputs/rook/34369610-d351-11eb-8f86-fa163e466023/downloads.csv\",\n",
+    "        \"http://rook8.cloud.dkrz.de:80/outputs/rook/ba7a8b60-50b6-11ef-a299-fa163e934c9b/downloads.csv\",\n",
     "    )"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 4,
    "id": "9fc29c58-067b-409b-b047-1c7ae1e5d3c8",
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/Users/pingu/Documents/GitHub/roocs/rook/rook/dashboard/tables/overview.py:32: FutureWarning: The default value of numeric_only in DataFrameGroupBy.sum is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n",
+      "  tdf = self.df_downloads.groupby(pd.Grouper(key=\"datetime\", freq=\"1D\")).sum()\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "'/tmp/dashboard.html'"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
    "source": [
     "dash.write()"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "3175d999-2052-4741-b19d-3ec371e0713c",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import IPython\n",
-    "IPython.display.HTML(\"/tmp/dashboard.html\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "80388a16-d341-416b-8b9c-940fb0cd77e9",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import pandas as pd"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "78d6ebd9-6717-4f9f-a608-f7fec855bac1",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import requests\n",
-    "from io import StringIO\n",
-    "ghc_url = \"https://geohealthcheck.cloud.dkrz.de/resource/45/history/csv\"\n",
-    "req = requests.get(ghc_url, verify=False)\n",
-    "df_ghc = pd.read_csv(StringIO(req.text), parse_dates=['checked_datetime'])\n",
-    "df_ghc"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "dbb74983-ba8e-41c6-8224-3a28f01a0b8f",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "df_ghc[\"down\"] = df_ghc.status.apply(lambda x: 1 if x=='False' else 0)\n",
-    "gdf = df_ghc.groupby(pd.Grouper(key=\"checked_datetime\", freq=\"1D\"))\n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "c229d50b-c852-4db5-a9d0-e5455a5680df",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "df = gdf.sum()\n",
-    "df"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "98058ba3-8ead-4618-bceb-1c62d62a2fea",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from bokeh.plotting import figure, show\n",
-    "from bokeh.models import ColumnDataSource"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "90d07d4c-9534-461a-89a7-91e6bfa2584f",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "plot = figure(\n",
-    "        title=\"Up\",\n",
-    "        tools=\"\",\n",
-    "        toolbar_location=None,\n",
-    "        # x_axis_label=\"Date\",\n",
-    "        x_axis_type=\"datetime\",\n",
-    "        # y_axis_label=\"Requests per day\",\n",
-    "        sizing_mode=\"scale_width\",\n",
-    "        plot_height=100,\n",
-    "    )\n",
-    "plot.line(x=\"checked_datetime\", y=\"down\", source=ColumnDataSource(df), color=\"green\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "aaf0dab6-97a2-4065-83ba-8cf3040607ad",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "show(plot)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "15846309-44f1-407d-ab45-9df83b803486",
+   "id": "51ed711f-cd44-4597-8320-3cb99fb287f2",
    "metadata": {},
    "outputs": [],
    "source": []
   }
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3",
+   "display_name": "Python 3 (ipykernel)",
    "language": "python",
    "name": "python3"
   },
@@ -171,7 +91,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.7.10"
+   "version": "3.11.9"
   }
  },
  "nbformat": 4,

diff --git a/rook/dashboard/dashboard.py b/rook/dashboard/dashboard.py
@@ -3,7 +3,7 @@
 import humanize
 import bokeh
 
-from .plots import ActivityPlot, DurationPlot, ConcurrencyPlot, PulsePlot
+from .plots import ActivityPlot, DurationPlot, ConcurrencyPlot, DayPlot
 from .plots import DownloadsPlot, TransferPlot
 from .tables import OverviewTable, MessageTable
 
@@ -56,7 +56,7 @@ def render(self):
         script_p1, plot_1 = ActivityPlot(self.df).components()
         script_p2, plot_2 = ConcurrencyPlot(self.df).components()
         script_p3, plot_3 = DurationPlot(self.df).components()
-        script_p4, plot_4 = PulsePlot(self.df).components()
+        script_p4, plot_4 = DayPlot(self.df).components()
         script_p5, plot_5 = DownloadsPlot(self.df_downloads).components()
         script_p6, plot_6 = TransferPlot(self.df_downloads).components()
         script_t1, table_1 = OverviewTable(self.df, self.df_downloads).components()

diff --git a/rook/dashboard/plots/__init__.py b/rook/dashboard/plots/__init__.py
@@ -1,6 +1,6 @@
 from .activity import ActivityPlot
 from .duration import DurationPlot
 from .concurrency import ConcurrencyPlot
-from .pulse import PulsePlot
+from .day import DayPlot
 from .downloads import DownloadsPlot
 from .transfer import TransferPlot
diff --git a/rook/dashboard/plots/day.py b/rook/dashboard/plots/day.py
@@ -0,0 +1,48 @@
+import pandas as pd
+from bokeh.plotting import figure
+from bokeh.models import ColumnDataSource
+
+from .base import PlotView
+
+DAYS = ["Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"]
+
+
+class DayPlot(PlotView):
+    def data(self):
+        pdf = pd.DataFrame()
+        pdf["day"] = self.df["time_start"].dt.dayofweek
+        pdf["day"] = pdf["day"].apply(lambda x: DAYS[x])
+        # pdf["time"] = self.df["time_start"].dt.time
+
+        day_counts = pdf["day"].value_counts().sort_index()
+        # total_count = day_counts.sum()
+        # day_percentages = (day_counts / total_count) * 100
+        data_ = dict(days=day_counts.index, counts=day_counts.values)
+        return data_
+
+    def plot(self):
+        plot = figure(
+            title="Requests per weekday",
+            tools="",
+            toolbar_location=None,
+            x_range=DAYS,
+            # x_axis_type="datetime",
+            # sizing_mode="scale_width",
+            plot_height=300,
+        )
+        plot.vbar(
+            x="days",
+            top="counts",
+            source=ColumnDataSource(self.data()),
+            width=0.9,
+            color="blue",
+            alpha=0.5,
+        )
+        plot.y_range.start = 0
+        # plot.x_range.range_padding = 0.1
+        plot.xgrid.grid_line_color = None
+        plot.axis.minor_tick_line_color = None
+        plot.outline_line_color = None
+        # plot.legend.location = "top_left"
+        # plot.legend.orientation = "horizontal"
+        return plot