From 98c5ecee5702f10d2daf630ec171f50a7abc632f Mon Sep 17 00:00:00 2001
From: Simon Adorf <sadorf@nvidia.com>
Date: Wed, 12 Apr 2023 11:55:39 -0700
Subject: [PATCH] Use more concise names for the array and store conversion
 functions.

---
 legate/raft/__init__.py    |  6 +++---
 legate/raft/core.py        |  4 ++--
 naive_bayes/multinomial.py | 24 ++++++++++++------------
 3 files changed, 17 insertions(+), 17 deletions(-)

diff --git a/legate/raft/__init__.py b/legate/raft/__init__.py
index 2ab49b3272..28fb076aa3 100644
--- a/legate/raft/__init__.py
+++ b/legate/raft/__init__.py
@@ -1,10 +1,11 @@
 from .array_api import add, exp, fill, log, negative, subtract, sum_over_axis
-from .core import array_to_store, convert, store_to_array
+from .core import as_array, as_store, convert
 from .multiarray import bincount, categorize, matmul, multiply
 
 __all__ = [
     "add",
-    "array_to_store",
+    "as_array",
+    "as_store",
     "bincount",
     "categorize",
     "convert",
@@ -14,7 +15,6 @@
     "matmul",
     "multiply",
     "negative",
-    "store_to_array",
     "subtract",
     "sum_over_axis",
 ]
diff --git a/legate/raft/core.py b/legate/raft/core.py
index 3832529ca8..a086611b2e 100644
--- a/legate/raft/core.py
+++ b/legate/raft/core.py
@@ -45,7 +45,7 @@ def __array_interface__(self):
         }
 
 
-def array_to_store(array: np.ndarray) -> Store:
+def as_store(array: np.ndarray) -> Store:
     store = context.create_store(
         pa.from_numpy_dtype(array.dtype),
         shape=array.shape,
@@ -59,7 +59,7 @@ def array_to_store(array: np.ndarray) -> Store:
     return store
 
 
-def store_to_array(store: Store) -> np.ndarray:
+def as_array(store: Store) -> np.ndarray:
     if store.kind is Future:
         dtype = store.get_dtype()
         buf = store.storage.get_buffer(dtype.size)
diff --git a/naive_bayes/multinomial.py b/naive_bayes/multinomial.py
index 57397a0fea..5a9fcde3d3 100644
--- a/naive_bayes/multinomial.py
+++ b/naive_bayes/multinomial.py
@@ -11,13 +11,13 @@
 
 from legate.raft import (
     add,
-    array_to_store,
+    as_array,
+    as_store,
     convert,
     exp,
     fill,
     log,
     matmul,
-    store_to_array,
     subtract,
     sum_over_axis,
 )
@@ -34,12 +34,12 @@ def __init__(self, alpha: float = 1.0, fit_prior=True, class_prior=None):
 
     def fit(self, X, y):
         # Convert to a legate stores
-        X = array_to_store(X)
+        X = as_store(X)
         _, n_features = X.shape
         self.n_features_in_ = n_features
 
         labelbin = LabelBinarizer()
-        Y = array_to_store(labelbin.fit_transform(y))
+        Y = as_store(labelbin.fit_transform(y))
 
         self.classes_ = labelbin.classes_
         assert Y.shape[1] != 1
@@ -69,12 +69,12 @@ def _update_feature_log_proba(self, alpha, dtype=pa.float64()):
     @property
     def feature_log_prob_(self):
         if self._feature_log_prob_ is not None:
-            return store_to_array(self._feature_log_prob_)
+            return as_array(self._feature_log_prob_)
 
     @property
     def class_log_prior_(self):
         if self._class_log_prior_ is not None:
-            return store_to_array(self._class_log_prior_)
+            return as_array(self._class_log_prior_)
 
     def _update_class_log_prior(self, class_prior=None):
         n_classes = len(self.classes_)
@@ -104,17 +104,17 @@ def _predict_log_proba(self, X):
         return subtract(x1.transpose((1, 0)), x2).transpose((1, 0))
 
     def predict_log_proba(self, X):
-        X = array_to_store(X)
+        X = as_store(X)
         ret = self._predict_log_proba(X)
-        return store_to_array(ret)
+        return as_array(ret)
 
     def predict_proba(self, X):
-        X = array_to_store(X)
+        X = as_store(X)
         ret = exp(self._predict_log_proba(X))
-        return store_to_array(ret)
+        return as_array(ret)
 
     def predict(self, X):
-        X = array_to_store(X)
-        jll = store_to_array(self._joint_log_likelihood(X))
+        X = as_store(X)
+        jll = as_array(self._joint_log_likelihood(X))
         ret = self.classes_[np.argmax(jll, axis=1)]
         return ret