Fix_typo (#812)

Co-authored-by: 0lgaF <olgafilippova@MacBook-Air-Olga.local>

examples/how_to_questions/how_to_calculate_recsys_metrics.ipynb

@@ -1,6 +1,6 @@
 {
   "nbformat": 4,
-  "nbformat_minor": 0,
+  "nbformat_minor": 2,
   "metadata": {
     "colab": {
       "provenance": []
@@ -17,19 +17,20 @@
     {
       "cell_type": "code",
       "execution_count": null,
-      "metadata": {
-        "id": "8aJYBZFNMyXc"
-      },
-      "outputs": [],
       "source": [
         "try:\n",
         "    import evidently\n",
         "except:\n",
         "    !pip install git+https://github.com/evidentlyai/evidently.git"
-      ]
+      ],
+      "outputs": [],
+      "metadata": {
+        "id": "8aJYBZFNMyXc"
+      }
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "import pandas as pd\n",
         "import numpy as np\n",
@@ -40,22 +41,21 @@
         "\n",
         "import requests"
       ],
+      "outputs": [],
       "metadata": {
         "id": "UfuNPLwjO99K"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "!pip install implicit"
       ],
+      "outputs": [],
       "metadata": {
         "id": "8A_dH0K0082d"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "markdown",
@@ -68,6 +68,7 @@
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "content = requests.get(\"http://files.grouplens.org/datasets/movielens/ml-100k.zip\").content\n",
         "\n",
@@ -77,14 +78,14 @@
         "  movies = arc.read(\"ml-100k/u.item\").decode(encoding='latin-1').split(\"\\n\")\n",
         "  users = arc.read(\"ml-100k/u.user\").decode(encoding='latin-1').split(\"\\n\")"
       ],
+      "outputs": [],
       "metadata": {
         "id": "f1wLolXpM02U"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "columns = ['user_id', 'movie_id', 'rating', 'timestamp']\n",
         "\n",
@@ -107,11 +108,10 @@
         "movies.drop(columns=['-', 'url'], inplace=True)\n",
         "movies[genres] = movies[genres].astype(int)"
       ],
+      "outputs": [],
       "metadata": {
         "id": "-V1w4P5LeV4X"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "markdown",
@@ -124,6 +124,7 @@
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "def transform_predictions(k, user_ids, item_ids):\n",
         "  return pd.DataFrame(\n",
@@ -138,11 +139,10 @@
         "  preds['rank'] = preds.groupby('user_id')['rank'].transform(lambda x: x.fillna(x.max() + 1))\n",
         "  return preds"
       ],
+      "outputs": [],
       "metadata": {
         "id": "MqP6bLDv92hY"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "markdown",
@@ -155,6 +155,7 @@
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "from implicit.cpu.als import AlternatingLeastSquares\n",
         "from scipy.sparse import csr_matrix\n",
@@ -163,66 +164,66 @@
         "als_model = AlternatingLeastSquares(factors=20, iterations=5, random_state=0)\n",
         "als_model.fit(csr_matrix(pivot_table))"
       ],
+      "outputs": [],
       "metadata": {
         "id": "-FQDsHEA3OKw"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "test_users = test.user_id.unique()\n",
         "len(test_users)"
       ],
+      "outputs": [],
       "metadata": {
         "id": "LLVvpQOvFIzl"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "ref_true = test[test.user_id.isin(test_users[:471])]\n",
         "curr_true = test[test.user_id.isin(test_users[471:])]"
       ],
+      "outputs": [],
       "metadata": {
         "id": "Sa8bO0CHFAJV"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "ids, scores = als_model.recommend(ref_true.user_id.unique() - 1, csr_matrix(pivot_table.loc[ref_true.user_id.unique()]), N=30, filter_already_liked_items=True)\n",
         "ref = get_full_df(30, ref_true.user_id.unique(), ids, ref_true)\n",
         "\n",
         "ids, scores = als_model.recommend(curr_true.user_id.unique() - 1, csr_matrix(pivot_table.loc[curr_true.user_id.unique()]), N=30, filter_already_liked_items=True)\n",
         "curr = get_full_df(30, curr_true.user_id.unique(), ids, curr_true)"
       ],
+      "outputs": [],
       "metadata": {
         "id": "LipvjnXIH3Y5"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "from evidently.metrics import PrecisionTopKMetric, RecallTopKMetric, FBetaTopKMetric, MAPKMetric, NDCGKMetric\n",
         "from evidently.pipeline.column_mapping import ColumnMapping\n",
         "from evidently.report import Report"
       ],
+      "outputs": [],
       "metadata": {
         "id": "vxU6s88ism0_"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [
         "report = Report(metrics=[\n",
         "    PrecisionTopKMetric(k=5),\n",
@@ -231,24 +232,23 @@
         "    MAPKMetric(k=5),\n",
         "    NDCGKMetric(k=5),\n",
         "])\n",
-        "column_mapping=ColumnMapping(recomendations_type='rank', target='rating', prediction='rank')\n",
+        "column_mapping=ColumnMapping(recommendations_type='rank', target='rating', prediction='rank')\n",
         "report.run(reference_data=ref.fillna(0), current_data=curr.fillna(0), column_mapping=column_mapping)\n",
         "report"
       ],
+      "outputs": [],
       "metadata": {
         "id": "7KIQreI6tKEA"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     },
     {
       "cell_type": "code",
+      "execution_count": null,
       "source": [],
+      "outputs": [],
       "metadata": {
         "id": "jSiTOBzhOl1s"
-      },
-      "execution_count": null,
-      "outputs": []
+      }
     }
   ]
 }

src/evidently/calculations/recommender_systems.py

@@ -9,7 +9,7 @@ def collect_dataset(
     users: pd.Series,
     target: pd.Series,
     preds: pd.Series,
-    recomendations_type: str,
+    recommendations_type: str,
     min_rel_score: Optional[int],
     no_feedback_users: bool,
     bin_data: bool,
@@ -18,7 +18,7 @@ def collect_dataset(
     df.columns = ["users", "target", "preds"]
     if min_rel_score:
         df["target"] = (df["target"] >= min_rel_score).astype(int)
-    if recomendations_type == "score":
+    if recommendations_type == "score":
         df["preds"] = df.groupby("users")["preds"].transform("rank", ascending=False)
     if bin_data:
         df["target"] = (df["target"] > 0).astype(int)
@@ -37,10 +37,10 @@ def get_curr_and_ref_df(
     if target_column is None or prediction is None:
         raise ValueError("Target and prediction were not found in data.")
     _, target_current, target_reference = data.get_data(target_column.column_name)
-    recomendations_type = data.column_mapping.recomendations_type
-    if recomendations_type is None:
-        recomendations_type = "scores"
-    if recomendations_type == "rank" and prediction.predicted_values is not None:
+    recommendations_type = data.column_mapping.recommendations_type
+    if recommendations_type is None:
+        recommendations_type = "scores"
+    if recommendations_type == "rank" and prediction.predicted_values is not None:
         pred_name = prediction.predicted_values.column_name
     elif prediction.prediction_probas is not None:
         pred_name = prediction.prediction_probas[0].column_name
@@ -53,7 +53,7 @@ def get_curr_and_ref_df(
         user_current,
         target_current,
         prediction_current,
-        recomendations_type,
+        recommendations_type,
         min_rel_score,
         no_feedback_users,
         bin_data,
@@ -65,7 +65,7 @@ def get_curr_and_ref_df(
             user_reference,
             target_reference,
             prediction_reference,
-            recomendations_type,
+            recommendations_type,
             min_rel_score,
             no_feedback_users,
             bin_data,

src/evidently/pipeline/column_mapping.py

@@ -32,7 +32,7 @@ class ColumnMapping:
     embeddings: Optional[Embeddings] = None
     user_id: Optional[str] = "user_id"
     item_id: Optional[str] = "item_id"
-    recomendations_type: Optional[str] = "score"
+    recommendations_type: Optional[str] = "score"
 
     def is_classification_task(self):
         return self.task == TaskType.CLASSIFICATION_TASK

src/evidently/utils/data_preprocessing.py

@@ -195,9 +195,9 @@ def _prediction_column(
                 raise ValueError("Prediction type is categorical but task is regression")
             if prediction_type == ColumnType.Numerical:
                 return PredictionColumns(predicted_values=ColumnDefinition(prediction, prediction_type))
-        if mapping is not None and mapping.recomendations_type == "rank":
+        if mapping is not None and mapping.recommendations_type == "rank":
             return PredictionColumns(predicted_values=ColumnDefinition(prediction, prediction_type))
-        if task == TaskType.RECOMMENDER_SYSTEMS and mapping is not None and mapping.recomendations_type == "score":
+        if task == TaskType.RECOMMENDER_SYSTEMS and mapping is not None and mapping.recommendations_type == "score":
             return PredictionColumns(prediction_probas=[ColumnDefinition(prediction, prediction_type)])
         if task is None:
             if prediction_type == ColumnType.Numerical and target_type == ColumnType.Categorical:

tests/metrics/recsys/test_f_beta_top_k.py

@@ -17,7 +17,7 @@ def test_fbeta_values():
 
     metric = FBetaTopKMetric(k=2)
     report = Report(metrics=[metric])
-    column_mapping = ColumnMapping(recomendations_type="rank")
+    column_mapping = ColumnMapping(recommendations_type="rank")
     report.run(reference_data=None, current_data=current, column_mapping=column_mapping)
 
     results = metric.get_result()

tests/metrics/recsys/test_map_k.py

@@ -17,7 +17,7 @@ def test_map_value():
 
     metric = MAPKMetric(k=2)
     report = Report(metrics=[metric])
-    column_mapping = ColumnMapping(recomendations_type="rank")
+    column_mapping = ColumnMapping(recommendations_type="rank")
     report.run(reference_data=None, current_data=current, column_mapping=column_mapping)
 
     results = metric.get_result()
@@ -38,7 +38,7 @@ def test_map_value_judged_only():
 
     metric = MAPKMetric(k=3, no_feedback_users=True)
     report = Report(metrics=[metric])
-    column_mapping = ColumnMapping(recomendations_type="rank")
+    column_mapping = ColumnMapping(recommendations_type="rank")
     report.run(reference_data=None, current_data=current, column_mapping=column_mapping)
 
     results = metric.get_result()

tests/metrics/recsys/test_mar_k.py

@@ -16,7 +16,7 @@ def test_mar_values():
 
     metric = MARKMetric(k=2)
     report = Report(metrics=[metric])
-    column_mapping = ColumnMapping(recomendations_type="rank")
+    column_mapping = ColumnMapping(recommendations_type="rank")
     report.run(reference_data=None, current_data=current, column_mapping=column_mapping)
 
     results = metric.get_result()

tests/metrics/recsys/test_ndcg_k.py

@@ -16,7 +16,7 @@ def test_ndcg_values():
 
     metric = NDCGKMetric(k=2)
     report = Report(metrics=[metric])
-    column_mapping = ColumnMapping(recomendations_type="rank")
+    column_mapping = ColumnMapping(recommendations_type="rank")
     report.run(reference_data=None, current_data=current, column_mapping=column_mapping)
 
     results = metric.get_result()

tests/metrics/recsys/test_precision_top_k.py

@@ -17,7 +17,7 @@ def test_precision_value():
 
     metric = PrecisionTopKMetric(k=2)
     report = Report(metrics=[metric])
-    column_mapping = ColumnMapping(recomendations_type="rank")
+    column_mapping = ColumnMapping(recommendations_type="rank")
     report.run(reference_data=None, current_data=current, column_mapping=column_mapping)
 
     results = metric.get_result()
@@ -38,7 +38,7 @@ def test_precision_value_judged_only():
 
     metric = PrecisionTopKMetric(k=3, no_feedback_users=True)
     report = Report(metrics=[metric])
-    column_mapping = ColumnMapping(recomendations_type="rank")
+    column_mapping = ColumnMapping(recommendations_type="rank")
     report.run(reference_data=None, current_data=current, column_mapping=column_mapping)
 
     results = metric.get_result()

tests/metrics/recsys/test_recall_top_k.py

@@ -17,7 +17,7 @@ def test_recall_values():
 
     metric = RecallTopKMetric(k=2)
     report = Report(metrics=[metric])
-    column_mapping = ColumnMapping(recomendations_type="rank")
+    column_mapping = ColumnMapping(recommendations_type="rank")
     report.run(reference_data=None, current_data=current, column_mapping=column_mapping)
 
     results = metric.get_result()
@@ -59,7 +59,7 @@ def test_recsll_include_no_feedback():
 
     metric = RecallTopKMetric(k=2, no_feedback_users=True)
     report = Report(metrics=[metric])
-    column_mapping = ColumnMapping(recomendations_type="rank")
+    column_mapping = ColumnMapping(recommendations_type="rank")
     report.run(reference_data=None, current_data=current, column_mapping=column_mapping)
 
     results = metric.get_result()