breaking: raise error for gaps in series (#504)

Nixtla · Oct 31, 2024 · 8b0660a · 8b0660a
1 parent 08a9f13
commit 8b0660a
Show file tree

Hide file tree

Showing 17 changed files with 425 additions and 1,951 deletions.
diff --git a/.github/workflows/lint.yaml b/.github/workflows/lint.yaml
@@ -19,7 +19,7 @@ jobs:
           python-version: '3.10'
 
       - name: Install dependencies
-        run: pip install black nbdev pre-commit
+        run: pip install black nbdev==2.3.25 pre-commit
 
       - name: Run pre-commit
-        run: pre-commit run --files nixtla/*
+        run: pre-commit run --show-diff-on-failure --files nixtla/*
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -6,10 +6,12 @@ repos:
         entry: sh -c 'nbdev_clean && nbdev_clean --fname nbs/src --clear_all'
         language: system
 
-  - repo: https://github.com/fastai/nbdev
-    rev: 2.2.10
+  - repo: local
     hooks:
       - id: nbdev_export
+        name: nbdev_export
+        entry: sh -c 'nbdev_export'
+        language: system
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
     rev: v0.2.1

diff --git a/README.md b/README.md
@@ -70,7 +70,7 @@ nixtla_client.plot(df, fcst_df, level=[80, 90])
 nixtla_client = NixtlaClient(api_key = 'YOUR API KEY HERE')
 
 # 2. Read Data # Wikipedia visits of NFL Star (
-df = pd.read_csv('https://raw.githubusercontent.com/Nixtla/transfer-learning-time-series/main/datasets/peyton_manning.csv')
+df = pd.read_csv('https://datasets-nixtla.s3.amazonaws.com/peyton-manning.csv')
 
 
 # 3. Detect Anomalies

diff --git a/nbs/docs/capabilities/anomaly-detection/01_quickstart.ipynb b/nbs/docs/capabilities/anomaly-detection/01_quickstart.ipynb
diff --git a/nbs/docs/capabilities/anomaly-detection/03_anomaly_detection_date_features.ipynb b/nbs/docs/capabilities/anomaly-detection/03_anomaly_detection_date_features.ipynb
diff --git a/nbs/docs/capabilities/anomaly-detection/04_confidence_levels.ipynb b/nbs/docs/capabilities/anomaly-detection/04_confidence_levels.ipynb
diff --git a/nbs/docs/capabilities/forecast/11_irregular_timestamps.ipynb b/nbs/docs/capabilities/forecast/11_irregular_timestamps.ipynb
@@ -48,7 +48,10 @@
    "source": [
     "# Irregular timestamps\n",
     "\n",
-    "TimeGPT can handle data with irregular timestamps. Simply specify the `freq` parameter with the right frequency of your data."
+    "* For pandas dataframes TimeGPT will try to infer the frequency of your timestamps. If you don't want TimeGPT to infer the frequency for you, you have to set the  `freq` argument to a valid [pandas frequency string](https://pandas.pydata.org/docs/user_guide/timeseries.html#dateoffset-objects), e.g. 'MS', 'YE'.\n",
+    "* For polars dataframes you have to set the `freq` argument to a valid [polars offset](https://docs.pola.rs/api/python/stable/reference/expressions/api/polars.Expr.dt.offset_by.html), e.g. '1d', '2w'.\n",
+    "\n",
+    "If your data has missing timestamps please refer to our [tutorial on missing values](https://docs.nixtla.io/docs/tutorials-missing_values)."
    ]
   },
   {
@@ -83,7 +86,8 @@
    "outputs": [],
    "source": [
     "import pandas as pd\n",
-    "from nixtla import NixtlaClient"
+    "from nixtla import NixtlaClient\n",
+    "from utilsforecast.data import generate_series"
    ]
   },
   {
@@ -120,6 +124,24 @@
     "    nixtla_client = NixtlaClient()"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "INFO:nixtla.nixtla_client:Querying model metadata...\n"
+     ]
+    }
+   ],
+   "source": [
+    "#| hide\n",
+    "_ = nixtla_client._get_model_params('timegpt-1', 'B')"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -130,9 +152,8 @@
      "output_type": "stream",
      "text": [
       "INFO:nixtla.nixtla_client:Validating inputs...\n",
+      "INFO:nixtla.nixtla_client:Inferred freq: B\n",
       "INFO:nixtla.nixtla_client:Preprocessing dataframes...\n",
-      "WARNING:nixtla.nixtla_client:You did not provide X_df. Exogenous variables in df are ignored. To surpress this warning, please add X_df with exogenous variables: Open, High, Low, Adj Close, Volume, Dividends, Stock Splits\n",
-      "WARNING:nixtla.nixtla_client:The specified horizon \"h\" exceeds the model horizon. This may lead to less accurate forecasts. Please consider using a smaller horizon.\n",
       "INFO:nixtla.nixtla_client:Restricting input...\n",
       "INFO:nixtla.nixtla_client:Calling Forecast Endpoint...\n"
      ]
@@ -141,19 +162,57 @@
    "source": [
     "# Read data\n",
     "# Dates for the weekends are missing\n",
-    "df = pd.read_csv('https://raw.githubusercontent.com/Nixtla/transfer-learning-time-series/main/datasets/openbb/pltr.csv')\n",
+    "df = pd.read_csv(\n",
+    "    'https://datasets-nixtla.s3.amazonaws.com/pltr.csv',\n",
+    "    usecols=['date', 'Close'],\n",
+    ")\n",
     "\n",
     "# Forecast\n",
-    "# We use B for the freq, as only business days are represented in the dataset\n",
+    "# the frequency is inferred as B, as only business days are represented in the dataset\n",
     "forecast_df = nixtla_client.forecast(\n",
-    "    df=df, \n",
-    "    h=14, \n",
+    "    df=df,\n",
+    "    h=5,\n",
+    "    time_col='date', \n",
+    "    target_col='Close',\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "INFO:nixtla.nixtla_client:Validating inputs...\n",
+      "INFO:nixtla.nixtla_client:Preprocessing dataframes...\n",
+      "INFO:nixtla.nixtla_client:Restricting input...\n",
+      "INFO:nixtla.nixtla_client:Calling Forecast Endpoint...\n"
+     ]
+    }
+   ],
+   "source": [
+    "# manually set the frequency\n",
+    "forecast_df2 = nixtla_client.forecast(\n",
+    "    df=df,\n",
     "    freq='B',\n",
+    "    h=5,\n",
     "    time_col='date', \n",
     "    target_col='Close',\n",
     ")"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pd.testing.assert_frame_equal(forecast_df, forecast_df2)"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -168,13 +227,6 @@
     "> \n",
     "> By default, `timegpt-1` is used. Please see [this tutorial](https://docs.nixtla.io/docs/tutorials-long_horizon_forecasting) on how and when to use `timegpt-1-long-horizon`."
    ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "For more details on handling datasets with irregular timesteps, check out our [tutorial](https://docs.nixtla.io/docs/tutorials-irregular_timestamps). "
-   ]
   }
  ],
  "metadata": {
@@ -185,5 +237,5 @@
   }
  },
  "nbformat": 4,
- "nbformat_minor": 2
+ "nbformat_minor": 4
 }
diff --git a/nbs/docs/getting-started/1_introduction.ipynb b/nbs/docs/getting-started/1_introduction.ipynb
@@ -32,7 +32,7 @@
     "\n",
     "* **[Prediction Intervals](https://docs.nixtla.io/docs/tutorials-prediction_intervals)**: Provide intervals in your predictions to quantify uncertainty effectively.\n",
     "\n",
-    "* **[Irregular Timestamps](https://docs.nixtla.io/docs/tutorials-irregular_timestamps)**: Handle data with irregular timestamps, accommodating non-uniform interval series without preprocessing.\n",
+    "* **[Irregular Timestamps](https://docs.nixtla.io/docs/capabilities-forecast-irregular_timestamps)**: Handle data with irregular timestamps, accommodating non-uniform interval series without preprocessing.\n",
     "\n",
     "* **[Anomaly Detection](https://docs.nixtla.io/docs/tutorials-anomaly_detection)**: Automatically detect anomalies in time series, and use exogenous features for enhanced performance."
    ]
@@ -90,7 +90,13 @@
    "source": []
   }
  ],
- "metadata": {},
+ "metadata": {
+  "kernelspec": {
+   "display_name": "python3",
+   "language": "python",
+   "name": "python3"
+  }
+ },
  "nbformat": 4,
- "nbformat_minor": 2
+ "nbformat_minor": 4
 }
diff --git a/nbs/docs/getting-started/5_faq.ipynb b/nbs/docs/getting-started/5_faq.ipynb
@@ -324,7 +324,7 @@
     "<details>\n",
     "    <summary>Can TimeGPT handle missing values?</summary>\n",
     "\n",
-    "`TimeGPT` cannot handle missing values or series with irregular timestamps. For more information, see the [Forecasting Time Series with Irregular Timestamps](https://docs.nixtla.io/docs/tutorials-irregular_timestamps) and the [Dealing with Missing Values](https://docs.nixtla.io/docs/tutorials-dealing_with_missing_values_in_timegpt) tutorial. \n",
+    "`TimeGPT` cannot handle missing values or series with irregular timestamps. For more information, see the [Forecasting Time Series with Irregular Timestamps](https://docs.nixtla.io/docs/capabilities-forecast-irregular_timestamps) and the [Dealing with Missing Values](https://docs.nixtla.io/docs/tutorials-dealing_with_missing_values_in_timegpt) tutorial. \n",
     "\n",
     "</details>"
    ]

diff --git a/nbs/docs/tutorials/08_cross_validation.ipynb b/nbs/docs/tutorials/08_cross_validation.ipynb
diff --git a/nbs/docs/tutorials/120_special_topics.ipynb b/nbs/docs/tutorials/120_special_topics.ipynb
@@ -19,7 +19,7 @@
     "\n",
     "### What You Will Learn\n",
     "\n",
-    "1. **[Irregular Timestamps](https://docs.nixtla.io/docs/tutorials-irregular_timestamps)**\n",
+    "1. **[Irregular Timestamps](https://docs.nixtla.io/docs/capabilities-forecast-irregular_timestamps)**\n",
     "\n",
     "    - Learn how to deal with irregular timestamps for correct usage of `TimeGPT`.\n",
     "\n",