Improve time series filtering based on cutoff, horizon and min_context_length (#18)

Author: shchur

README.md

@@ -55,13 +55,12 @@ Get an evaluation summary
 task.evaluation_summary(predictions, model_name="naive")
 # {'model_name': 'naive',
 #  'dataset_name': 'chronos_datasets_monash_kdd_cup_2018',
-#  'dataset_fingerprint': '8a50d3417859652b',
 #  'dataset_path': 'autogluon/chronos_datasets',
 #  'dataset_config': 'monash_kdd_cup_2018',
 #  'horizon': 12,
 #  'cutoff': -12,
 #  'lead_time': 1,
-#  'min_ts_length': 13,
+#  'min_context_length': 1,
 #  'max_context_length': None,
 #  'seasonality': 1,
 #  'eval_metric': 'MASE',
@@ -70,15 +69,16 @@ task.evaluation_summary(predictions, model_name="naive")
 #  'id_column': 'id',
 #  'timestamp_column': 'timestamp',
 #  'target_column': 'target',
-#  'multiple_target_columns': None,
+#  'generate_univariate_targets_from': None,
 #  'past_dynamic_columns': [],
 #  'excluded_columns': [],
-#  'test_error': 3.3784518,
+#  'test_error': 3.3784518866750513,
 #  'training_time_s': None,
 #  'inference_time_s': None,
+#  'dataset_fingerprint': 'a22d13d4c1e8641c',
 #  'trained_on_this_dataset': False,
-#  'fev_version': '0.3.0',
-#  'MASE': 3.3784518}
+#  'fev_version': '0.5.0',
+#  'MASE': 3.3784518866750513}
 ```
 The evaluation summary contains all information necessary to uniquely identify the forecasting task.
 

docs/01-quickstart.ipynb

@@ -38,12 +38,54 @@
     {
      "data": {
       "application/vnd.jupyter.widget-view+json": {
-       "model_id": "351a9be21f444cfc90440176e41763d5",
+       "model_id": "da69c728429643ca867244061497eeca",
        "version_major": 2,
        "version_minor": 0
       },
       "text/plain": [
-       "Validating dataset format (num_proc=10):   0%|          | 0/270 [00:00<?, ? examples/s]"
+       "Validating dataset format (num_proc=48):   0%|          | 0/270 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "b285297ae35c45ceac7c55115bded869",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Filtering short time series (num_proc=48):   0%|          | 0/270 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "5d148bdd05444470a6e8387d74f0e5ef",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Selecting past data (num_proc=48):   0%|          | 0/270 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "cad0f53a10854ebba967c941804e2616",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Selecting future data (num_proc=48):   0%|          | 0/270 [00:00<?, ? examples/s]"
       ]
      },
      "metadata": {},
@@ -87,15 +129,15 @@
     {
      "data": {
       "text/plain": [
-       "{'id': np.str_('T000000'),\n",
+       "{'id': 'T000000',\n",
        " 'timestamp': array(['2017-01-01T14:00:00.000', '2017-01-01T15:00:00.000',\n",
        "        '2017-01-01T16:00:00.000', ..., '2018-03-31T01:00:00.000',\n",
        "        '2018-03-31T02:00:00.000', '2018-03-31T03:00:00.000'],\n",
        "       dtype='datetime64[ms]'),\n",
        " 'target': array([453., 417., 395., ...,  99., 102.,  97.], dtype=float32),\n",
-       " 'city': np.str_('Beijing'),\n",
-       " 'station': np.str_('aotizhongxin_aq'),\n",
-       " 'measurement': np.str_('PM2.5')}"
+       " 'city': 'Beijing',\n",
+       " 'station': 'aotizhongxin_aq',\n",
+       " 'measurement': 'PM2.5'}"
       ]
      },
      "execution_count": 5,
@@ -139,17 +181,17 @@
     {
      "data": {
       "text/plain": [
-       "{'id': np.str_('T000000'),\n",
+       "{'id': 'T000000',\n",
        " 'timestamp': array(['2018-03-31T04:00:00.000', '2018-03-31T05:00:00.000',\n",
        "        '2018-03-31T06:00:00.000', '2018-03-31T07:00:00.000',\n",
        "        '2018-03-31T08:00:00.000', '2018-03-31T09:00:00.000',\n",
        "        '2018-03-31T10:00:00.000', '2018-03-31T11:00:00.000',\n",
        "        '2018-03-31T12:00:00.000', '2018-03-31T13:00:00.000',\n",
        "        '2018-03-31T14:00:00.000', '2018-03-31T15:00:00.000'],\n",
        "       dtype='datetime64[ms]'),\n",
-       " 'city': np.str_('Beijing'),\n",
-       " 'station': np.str_('aotizhongxin_aq'),\n",
-       " 'measurement': np.str_('PM2.5')}"
+       " 'city': 'Beijing',\n",
+       " 'station': 'aotizhongxin_aq',\n",
+       " 'measurement': 'PM2.5'}"
       ]
      },
      "execution_count": 7,
@@ -189,7 +231,7 @@
     {
      "data": {
       "application/vnd.jupyter.widget-view+json": {
-       "model_id": "3034f4b2cdc3471eac2bc3caeab28d75",
+       "model_id": "5a5df4eb7fd94875a6b9fb0cf4f757c8",
        "version_major": 2,
        "version_minor": 0
       },
@@ -200,6 +242,20 @@
      "metadata": {},
      "output_type": "display_data"
     },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "fc07c39251fb416a8de6b007b43490fa",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Map (num_proc=48):   0%|          | 0/270 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
     {
      "data": {
       "text/plain": [
@@ -210,7 +266,7 @@
        " 'horizon': 12,\n",
        " 'cutoff': -12,\n",
        " 'lead_time': 1,\n",
-       " 'min_ts_length': 13,\n",
+       " 'min_context_length': 1,\n",
        " 'max_context_length': None,\n",
        " 'seasonality': 1,\n",
        " 'eval_metric': 'MASE',\n",
@@ -219,15 +275,15 @@
        " 'id_column': 'id',\n",
        " 'timestamp_column': 'timestamp',\n",
        " 'target_column': 'target',\n",
-       " 'multiple_target_columns': None,\n",
+       " 'generate_univariate_targets_from': None,\n",
        " 'past_dynamic_columns': [],\n",
        " 'excluded_columns': [],\n",
        " 'test_error': 3.3784518866750513,\n",
        " 'training_time_s': None,\n",
        " 'inference_time_s': None,\n",
-       " 'dataset_fingerprint': 'e91f446bd0b05d16',\n",
+       " 'dataset_fingerprint': 'a22d13d4c1e8641c',\n",
        " 'trained_on_this_dataset': False,\n",
-       " 'fev_version': '0.2.1',\n",
+       " 'fev_version': '0.5.0',\n",
        " 'MASE': 3.3784518866750513}"
       ]
      },
@@ -250,7 +306,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 10,
    "metadata": {},
    "outputs": [
     {
@@ -634,7 +690,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.11.10"
+   "version": "3.11.11"
   }
  },
  "nbformat": 4,