convert frame improved

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "perpetual"
-version = "0.7.5"
+version = "0.7.6"
 edition = "2021"
 authors = ["Mutlu Simsek <[email protected]>"]
 homepage = "https://perpetual-ml.com"

README.md

@@ -66,7 +66,7 @@ conda install conda-forge::perpetual
 
 To use in a Rust project and to get the package from [crates.io](https://crates.io/crates/perpetual):
 
-```toml
+```shell
 cargo add perpetual
 ```
 

python-package/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "py-perpetual"
-version = "0.7.5"
+version = "0.7.6"
 edition = "2021"
 authors = ["Mutlu Simsek <[email protected]>"]
 homepage = "https://perpetual-ml.com"
@@ -19,7 +19,7 @@ crate-type = ["cdylib", "rlib"]
 
 [dependencies]
 pyo3 = { version = "0.22.6", features = ["extension-module"] }
-perpetual_rs = {package="perpetual", version = "0.7.5", path = "../" }
+perpetual_rs = {package="perpetual", version = "0.7.6", path = "../" }
 numpy = "0.22.1"
 ndarray = "0.16.1"
 serde_plain = { version = "1.0" }

python-package/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "maturin"
 
 [project]
 name = "perpetual"
-version = "0.7.5"
+version = "0.7.6"
 description = "A self-generalizing gradient boosting machine which doesn't need hyperparameter optimization"
 license = { file = "LICENSE" }
 keywords = [

python-package/python/perpetual/utils.py

@@ -117,38 +117,33 @@ def convert_input_frame(
     cat_to_num = []
     if categorical_features_:
         for i in categorical_features_:
-            categories = np.unique(X_[:, i].astype(dtype="str", copy=False))
+            categories, inversed = np.unique(X_[:, i].astype(str), return_inverse=True)
+            if categories[-1] == "nan":
+                categories = list(categories)
+            else:
+                categories = list(categories)
+                categories.append("nan")
+            inversed = inversed.astype(np.float32)
+
             if len(categories) > max_cat:
                 cat_to_num.append(i)
                 logger.warning(
                     f"Feature {features_[i]} will be treated as numerical since the number of categories ({len(categories)}) exceeds max_cat ({max_cat}) threshold."
                 )
-            categories = [c for c in list(categories) if c != "nan"]
-            categories.insert(0, "nan")
-            cat_mapping[features_[i]] = categories
+
+            feature_name = features_[i]
+            cat_mapping[feature_name] = categories
+            ind_nan = len(categories) - 1
+            inversed[inversed == ind_nan] = np.nan
+            X_[:, i] = inversed
+
         categorical_features_ = [
             x for x in categorical_features_ if x not in cat_to_num
         ]
 
-    if cat_mapping:
         logger.info(f"Categorical features: {categorical_features_}")
         logger.info(f"Mapping of categories: {cat_mapping}")
 
-        for feature_name, categories in cat_mapping.items():
-
-            def f(x):
-                try:
-                    return (
-                        np.nan
-                        if str(x[feature_index]) == "nan"
-                        else float(categories.index(str(x[feature_index])))
-                    )
-                except (ValueError, IndexError):
-                    return np.nan
-
-            feature_index = features_.index(feature_name)
-            X_[:, feature_index] = np.apply_along_axis(f, 1, X_)
-
     if not np.issubdtype(X_.dtype, "float64"):
         X_ = X_.astype(dtype="float64", copy=False)
     flat_data = X_.ravel(order="F")
@@ -183,18 +178,12 @@ def transform_input_frame(X, cat_mapping) -> Tuple[List[str], np.ndarray, int, i
     if cat_mapping:
         for feature_name, categories in cat_mapping.items():
             feature_index = features_.index(feature_name)
-
-            def f(x):
-                try:
-                    return (
-                        np.nan
-                        if str(x[feature_index]) == "nan"
-                        else float(categories.index(str(x[feature_index])))
-                    )
-                except (ValueError, IndexError):
-                    return np.nan
-
-            X_[:, feature_index] = np.apply_along_axis(f, 1, X_)
+            x_enc = np.searchsorted(
+                categories, X_[:, feature_index].astype(str)
+            ).astype(np.float32)
+            ind_nan = len(categories) - 1
+            x_enc[x_enc == ind_nan] = np.nan
+            X_[:, feature_index] = x_enc
 
     if not np.issubdtype(X_.dtype, "float64"):
         X_ = X_.astype(dtype="float64", copy=False)