Adding comment on papermill_launcher.ipynb

destination-earth · Aug 30, 2024 · 28c978b · 28c978b
1 parent 43b7a0a
commit 28c978b
Showing 1 changed file with 21 additions and 7 deletions.
diff --git a/docs/papermill_launcher.ipynb b/docs/papermill_launcher.ipynb
@@ -186,10 +186,24 @@
    "source": [
     "___\n",
     "### Explantion of the code below \n",
-    "- nbs is a list of the notebooks that has been processed, wether they failed or not.\n",
-    "- The code loops over the tag id present in tag list and calculates the time difference in the tagging events.\n",
     "- If the fish has observation over 2 days and has not been processed yet, it starts running a parametrized notebook.\n",
-    "- If it succeds, the generated notebook is placed papermill_output/done, else, it goes at papermill_output/failed"
+    "- nbs is a list of the notebooks that has been processed, wether they failed or not. This list is used to keep a track of the tags that already has been generated.\n",
+    "These two conditions are used in the following way.\n",
+    "```\n",
+    "observation_length = (recapture_date - release_date) / np.timedelta64(1, \"D\")\n",
+    "\n",
+    "if (\n",
+    "    (tag_name not in nbs) and observation_length > 2\n",
+    "):  # Use this statement if you already start a computation that has been interrupted but the generation is still valid\n",
+    "    # if time_difference > 2: # Use this if you want every tag in tag list to be processed\n",
+    "```\n",
+    "First, it means that if the fish, based on the tagging events, has a observation period of less than two days, his trajectory will not be computed.\n",
+    "Second, It means that you can either choose to regenerate for all the tags that you generated once if you noticed that there was an issue in the results.\n",
+    "If the generation was interrupted during the process but the results are valid, you can start back you computation where it has stopped.\n",
+    "\n",
+    "- The code loops over the tag id present in tag list and calculates the time difference in the tagging events.\n",
+    "- If it succeds, the generated notebook is placed papermill_output/done, else, it goes at papermill_output/failed\n",
+    " "
    ]
   },
   {
@@ -213,16 +227,16 @@
     "for tag_name in tqdm(tag_list, desc=\"Processing tags\"):\n",
     "    try:\n",
     "        te = pd.read_csv(s3.open(f\"{cloud_root}/cleaned/{tag_name}/tagging_events.csv\"))\n",
-    "        np_datetime1 = np.datetime64(\n",
+    "        release_date = np.datetime64(\n",
     "            datetime.strptime(te[\"time\"][0], \"%Y-%m-%dT%H:%M:%SZ\")\n",
     "        )\n",
-    "        np_datetime2 = np.datetime64(\n",
+    "        recapture_date = np.datetime64(\n",
     "            datetime.strptime(te[\"time\"][1], \"%Y-%m-%dT%H:%M:%SZ\")\n",
     "        )\n",
-    "        time_difference = (np_datetime2 - np_datetime1) / np.timedelta64(1, \"D\")\n",
+    "        observation_length = (recapture_date - release_date) / np.timedelta64(1, \"D\")\n",
     "\n",
     "        if (\n",
-    "            (tag_name not in nbs) and time_difference > 2\n",
+    "            (tag_name not in nbs) and observation_length > 2\n",
     "        ):  # Use this statement if you already start a computation that has been interrupted but the generation is still valid\n",
     "            # if time_difference > 2: # Use this if you want every tag in tag list to be processed\n",
     "\n",