fix the represantion of nan values in bivariate plots

src/syngen/VERSION

@@ -1 +1 @@
-0.9.47
+0.9.48rc2

src/syngen/ml/metrics/accuracy_test/accuracy_test.py

@@ -112,12 +112,11 @@ def __init__(
         synthetic: pd.DataFrame,
         paths: dict,
         table_name: str,
-        infer_config: Dict,
-        columns_nan_labels: Dict,
+        infer_config: Dict
     ):
         super().__init__(original, synthetic, paths, table_name, infer_config)
         self.reports_path = f"{self.paths['reports_path']}/accuracy"
-        self.columns_nan_labels = columns_nan_labels
+        self.dataset_pickle_path = self.paths["dataset_pickle_path"]
         self.univariate = UnivariateMetric(
             self.original,
             self.synthetic,
@@ -129,7 +128,7 @@ def __init__(
             self.synthetic,
             self.plot_exists,
             self.reports_path,
-            self.columns_nan_labels
+            self.dataset_pickle_path
         )
         self.correlations = Correlations(
             self.original,

src/syngen/ml/metrics/metrics_classes/metrics.py

@@ -26,7 +26,11 @@
 from slugify import slugify
 from loguru import logger
 
-from syngen.ml.utils import timestamp_to_datetime, timing
+from syngen.ml.utils import (
+    timestamp_to_datetime,
+    timing,
+    fetch_config
+)
 matplotlib.use("Agg")
 
 
@@ -338,11 +342,20 @@ def __init__(
         synthetic: pd.DataFrame,
         plot: bool,
         reports_path: str,
-        columns_nan_labels: Dict,
+        dataset_pickle_path: str
     ):
         super().__init__(original, synthetic, plot, reports_path)
-        self.cmap = LinearSegmentedColormap.from_list("rg", ["#0D5598", "#3E92E0", "#E8F4FF"])
-        self.columns_nan_labels = columns_nan_labels
+        self.cmap = LinearSegmentedColormap.from_list(
+            "rg", ["#0D5598", "#3E92E0", "#E8F4FF"]
+        )
+        self.dataset = fetch_config(dataset_pickle_path)
+        nan_labels_dict = self.dataset.nan_labels_dict
+        na_values = self.dataset.format.get("na_values", [])
+        self.missing_values: Dict[str, str] = (
+            {col: na_values[0] for col in self.dataset.order_of_columns}
+            if na_values
+            else nan_labels_dict
+        )
 
     @staticmethod
     def _format_date_labels(heatmap_orig_data, heatmap_synthetic_data, axis):
@@ -471,7 +484,7 @@ def get_common_min_max(original, synthetic):
         return vmin, vmax
 
     @staticmethod
-    def __format_float_tick_labels(labels: List, nan_label: str = 'NaN') -> List:
+    def __format_float_tick_labels(labels: List, nan_label: str = "nan") -> List:
         labels = [nan_label if pd.isna(l) else l for l in labels]
         if all([isinstance(i, float) for i in labels]) and (
             max(labels) > 1e5 or min(labels) < 1e-03
@@ -494,8 +507,16 @@ def _plot_heatmap(
         ax = self._axes.flat[plt_index]
         ax.tick_params(labelsize=14)
         heatmap, x_tick_labels, y_tick_labels = heatmap_data
-        x_tick_labels = self.__format_float_tick_labels(x_tick_labels, self.columns_nan_labels.get(xfeature, 'NaN'))
-        y_tick_labels = self.__format_float_tick_labels(y_tick_labels, self.columns_nan_labels.get(yfeature, 'NaN'))
+        print(f"!!!!!!!!!!!!!!!!!!!!!")
+        print(f"{self.missing_values}")
+        x_tick_labels = self.__format_float_tick_labels(
+            x_tick_labels,
+            self.missing_values.get(xfeature, "nan")
+        )
+        y_tick_labels = self.__format_float_tick_labels(
+            y_tick_labels,
+            self.missing_values.get(yfeature, "nan")
+        )
         ax = sns.heatmap(
             heatmap,
             xticklabels=x_tick_labels,

src/syngen/ml/reporters/reporters.py

@@ -37,6 +37,7 @@ def __init__(
         self.loader = loader
         self.dataset = None
         self.columns_nan_labels = dict()
+        self.na_values = dict()
 
     def _fetch_dataframe(self) -> pd.DataFrame:
         """
@@ -63,7 +64,6 @@ def _extract_report_data(self) -> Tuple[pd.DataFrame, pd.DataFrame]:
 
     def fetch_data_types(self):
         self.dataset = fetch_config(self.paths["dataset_pickle_path"])
-        self.columns_nan_labels = self.dataset.nan_labels_dict
         types = (
             self.dataset.str_columns,
             self.dataset.date_columns,
@@ -294,8 +294,7 @@ def report(self):
             synthetic,
             self.paths,
             self.table_name,
-            self.config,
-            self.columns_nan_labels,
+            self.config
         )
         accuracy_test.report(
             cont_columns=list(float_columns | int_columns),