rename _ProcessedDF (#95)

Author: jmoralez

nbs/processing.ipynb

@@ -1630,33 +1630,24 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "4de12264-0bd1-4eed-935b-7b7fb1cbebc0",
+   "id": "62293bd2-b921-40b2-b1af-25f0b8e55006",
    "metadata": {},
    "outputs": [],
    "source": [
-    "#| exporti\n",
-    "class _ProcessedDF(NamedTuple):\n",
+    "#| export\n",
+    "class ProcessedDF(NamedTuple):\n",
     "    uids: Series\n",
-    "    times: np.ndarray\n",
+    "    last_times: np.ndarray\n",
     "    data: np.ndarray\n",
     "    indptr: np.ndarray\n",
-    "    sort_idxs: Optional[np.ndarray]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "62293bd2-b921-40b2-b1af-25f0b8e55006",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "#| export\n",
+    "    sort_idxs: Optional[np.ndarray]\n",
+    "\n",
     "def process_df(\n",
     "    df: DataFrame,\n",
     "    id_col: str,\n",
     "    time_col: str,\n",
     "    target_col: Optional[str],\n",
-    ") -> _ProcessedDF:\n",
+    ") -> ProcessedDF:\n",
     "    \"\"\"Extract components from dataframe\n",
     "    \n",
     "    Parameters\n",
@@ -1699,7 +1690,7 @@
     "        data = data[sort_idxs]\n",
     "        last_idxs = sort_idxs[last_idxs]\n",
     "    times = df[time_col].to_numpy()[last_idxs]\n",
-    "    return _ProcessedDF(uids, times, data, indptr, sort_idxs)"
+    "    return ProcessedDF(uids, times, data, indptr, sort_idxs)"
    ]
   },
   {

utilsforecast/_modidx.py

@@ -84,8 +84,8 @@
                                                                                                     'utilsforecast/processing.py'),
                                           'utilsforecast.processing.DataFrameProcessor.process': ( 'processing.html#dataframeprocessor.process',
                                                                                                    'utilsforecast/processing.py'),
-                                          'utilsforecast.processing._ProcessedDF': ( 'processing.html#_processeddf',
-                                                                                     'utilsforecast/processing.py'),
+                                          'utilsforecast.processing.ProcessedDF': ( 'processing.html#processeddf',
+                                                                                    'utilsforecast/processing.py'),
                                           'utilsforecast.processing._ensure_month_ends': ( 'processing.html#_ensure_month_ends',
                                                                                            'utilsforecast/processing.py'),
                                           'utilsforecast.processing._multiply_pl_freq': ( 'processing.html#_multiply_pl_freq',

utilsforecast/processing.py

@@ -6,7 +6,7 @@
            'horizontal_concat', 'copy_if_pandas', 'join', 'drop_index_if_pandas', 'rename', 'sort', 'offset_times',
            'offset_dates', 'time_ranges', 'repeat', 'cv_times', 'group_by', 'group_by_agg', 'is_in', 'between',
            'fill_null', 'cast', 'value_cols_to_numpy', 'make_future_dataframe', 'anti_join', 'ensure_sorted',
-           'process_df', 'DataFrameProcessor', 'backtest_splits', 'add_insample_levels']
+           'ProcessedDF', 'process_df', 'DataFrameProcessor', 'backtest_splits', 'add_insample_levels']
 
 # %% ../nbs/processing.ipynb 2
 import re
@@ -641,20 +641,20 @@ def ensure_sorted(df: DataFrame, id_col: str, time_col: str) -> DataFrame:
     return df
 
 # %% ../nbs/processing.ipynb 75
-class _ProcessedDF(NamedTuple):
+class ProcessedDF(NamedTuple):
     uids: Series
-    times: np.ndarray
+    last_times: np.ndarray
     data: np.ndarray
     indptr: np.ndarray
     sort_idxs: Optional[np.ndarray]
 
-# %% ../nbs/processing.ipynb 76
+
 def process_df(
     df: DataFrame,
     id_col: str,
     time_col: str,
     target_col: Optional[str],
-) -> _ProcessedDF:
+) -> ProcessedDF:
     """Extract components from dataframe
 
     Parameters
@@ -697,9 +697,9 @@ def process_df(
         data = data[sort_idxs]
         last_idxs = sort_idxs[last_idxs]
     times = df[time_col].to_numpy()[last_idxs]
-    return _ProcessedDF(uids, times, data, indptr, sort_idxs)
+    return ProcessedDF(uids, times, data, indptr, sort_idxs)
 
-# %% ../nbs/processing.ipynb 78
+# %% ../nbs/processing.ipynb 77
 class DataFrameProcessor:
     def __init__(
         self,
@@ -716,7 +716,7 @@ def process(
     ) -> Tuple[Series, np.ndarray, np.ndarray, np.ndarray, Optional[np.ndarray]]:
         return process_df(df, self.id_col, self.time_col, self.target_col)
 
-# %% ../nbs/processing.ipynb 83
+# %% ../nbs/processing.ipynb 82
 def _single_split(
     df: DataFrame,
     i_window: int,
@@ -781,7 +781,7 @@ def _single_split(
         )
     return cutoffs, train_mask, valid_mask
 
-# %% ../nbs/processing.ipynb 84
+# %% ../nbs/processing.ipynb 83
 def backtest_splits(
     df: DataFrame,
     n_windows: int,
@@ -813,7 +813,7 @@ def backtest_splits(
         valid = filter_with_mask(df, valid_mask)
         yield cutoffs, train, valid
 
-# %% ../nbs/processing.ipynb 88
+# %% ../nbs/processing.ipynb 87
 def add_insample_levels(
     df: DataFrame,
     models: List[str],