feat!: prepare for 2.5.0 (#2798)

* fix: drop lengths recursively (#2775) * fix: drop lengths recursively * test: ensure slicing from-typetracer works * chore: bump pyproject.toml * chore: remove to_categorical (#2779)
scikit-hep · Nov 7, 2023 · bd4cb8d · bd4cb8d
1 parent 3012363
commit bd4cb8d
Show file tree

Hide file tree

Showing 26 changed files with 64 additions and 508 deletions.
diff --git a/docs/redirects.json b/docs/redirects.json
@@ -120,7 +120,7 @@
   "_auto/ak.to_arrow.any-ext": "../reference/generated/ak.to_arrow.html",
   "_auto/ak.to_arrow_table.any-ext": "../reference/generated/ak.to_arrow_table.html",
   "_auto/ak.to_buffers.any-ext": "../reference/generated/ak.to_buffers.html",
-  "_auto/ak.to_categorical.any-ext": "../reference/generated/ak.to_categorical.html",
+  "_auto/ak.to_categorical.any-ext": "../reference/generated/ak.str.to_categorical.html",
   "_auto/ak.to_cupy.any-ext": "../reference/generated/ak.to_cupy.html",
   "_auto/ak.to_jax.any-ext": "../reference/generated/ak.to_jax.html",
   "_auto/ak.to_json.any-ext": "../reference/generated/ak.to_json.html",

diff --git a/docs/reference/toctree.txt b/docs/reference/toctree.txt
@@ -232,10 +232,10 @@
 .. toctree::
     :caption: Arrays of categorical data
 
-    generated/ak.to_categorical
     generated/ak.from_categorical
     generated/ak.is_categorical
     generated/ak.categories
+    generated/ak.str.to_categorical
 
 .. toctree::
     :caption: Indexing and grouping

diff --git a/docs/user-guide/how-to-create-strings.md b/docs/user-guide/how-to-create-strings.md
@@ -95,15 +95,15 @@ Categorical strings
 
 A large set of strings with few unique values are more efficiently manipulated as integers than as strings. In Pandas, this is [categorical data](https://pandas.pydata.org/pandas-docs/stable/user_guide/categorical.html), in R, it's called a [factor](https://www.rdocumentation.org/packages/base/versions/3.6.2/topics/factor), and in Arrow and Parquet, it's [dictionary encoding](https://arrow.apache.org/blog/2019/09/05/faster-strings-cpp-parquet/).
 
-The {func}`ak.to_categorical` function makes Awkward Arrays categorical in this sense. {func}`ak.to_arrow` and {func}`ak.to_parquet` recognize categorical data and convert it to the corresponding Arrow and Parquet types.
+The {func}`ak.str.to_categorical` (requires PyArrow) function makes Awkward Arrays categorical in this sense. {func}`ak.to_arrow` and {func}`ak.to_parquet` recognize categorical data and convert it to the corresponding Arrow and Parquet types.
 
 ```{code-cell} ipython3
 uncategorized = ak.Array(["three", "one", "two", "two", "three", "one", "one", "one"])
 uncategorized
 ```
 
 ```{code-cell} ipython3
-categorized = ak.to_categorical(uncategorized)
+categorized = ak.str.to_categorical(uncategorized)
 categorized
 ```
 

diff --git a/pyproject.toml b/pyproject.toml
@@ -7,7 +7,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "awkward"
-version = "2.4.9"
+version = "2.5.0rc1"
 description = "Manipulate JSON-like data with NumPy-like idioms."
 license = { text = "BSD-3-Clause" }
 requires-python = ">=3.8"

diff --git a/src/awkward/contents/bitmaskedarray.py b/src/awkward/contents/bitmaskedarray.py
@@ -312,7 +312,7 @@ def _to_typetracer(self, forget_length: bool) -> Self:
         tt = TypeTracer.instance()
         return BitMaskedArray(
             self._mask.to_nplike(tt),
-            self._content._to_typetracer(False),
+            self._content._to_typetracer(forget_length),
             self._valid_when,
             unknown_length if forget_length else self.length,
             self._lsb_order,

diff --git a/src/awkward/contents/bytemaskedarray.py b/src/awkward/contents/bytemaskedarray.py
@@ -235,7 +235,7 @@ def _to_typetracer(self, forget_length: bool) -> Self:
         mask = self._mask.to_nplike(tt)
         return ByteMaskedArray(
             mask.forget_length() if forget_length else mask,
-            self._content._to_typetracer(False),
+            self._content._to_typetracer(forget_length),
             self._valid_when,
             parameters=self._parameters,
         )

diff --git a/src/awkward/contents/indexedarray.py b/src/awkward/contents/indexedarray.py
@@ -225,7 +225,7 @@ def _to_typetracer(self, forget_length: bool) -> Self:
         index = self._index.to_nplike(TypeTracer.instance())
         return IndexedArray(
             index.forget_length() if forget_length else index,
-            self._content._to_typetracer(False),
+            self._content._to_typetracer(forget_length),
             parameters=self._parameters,
         )
 

diff --git a/src/awkward/contents/indexedoptionarray.py b/src/awkward/contents/indexedoptionarray.py
@@ -214,7 +214,7 @@ def _to_typetracer(self, forget_length: bool) -> Self:
         index = self._index.to_nplike(TypeTracer.instance())
         return IndexedOptionArray(
             index.forget_length() if forget_length else index,
-            self._content._to_typetracer(False),
+            self._content._to_typetracer(forget_length),
             parameters=self._parameters,
         )
 

diff --git a/src/awkward/contents/listarray.py b/src/awkward/contents/listarray.py
@@ -229,7 +229,7 @@ def _to_typetracer(self, forget_length: bool) -> Self:
         return ListArray(
             starts.forget_length() if forget_length else starts,
             self._stops.to_nplike(tt),
-            self._content._to_typetracer(False),
+            self._content._to_typetracer(forget_length),
             parameters=self._parameters,
         )
 

diff --git a/src/awkward/contents/listoffsetarray.py b/src/awkward/contents/listoffsetarray.py
@@ -212,7 +212,7 @@ def _to_typetracer(self, forget_length: bool) -> Self:
         offsets = self._offsets.to_nplike(TypeTracer.instance())
         return ListOffsetArray(
             offsets.forget_length() if forget_length else offsets,
-            self._content._to_typetracer(False),
+            self._content._to_typetracer(forget_length),
             parameters=self._parameters,
         )
 

diff --git a/src/awkward/contents/unionarray.py b/src/awkward/contents/unionarray.py
@@ -487,7 +487,7 @@ def _to_typetracer(self, forget_length: bool) -> Self:
         return UnionArray(
             tags.forget_length() if forget_length else tags,
             self._index.to_nplike(tt),
-            [x._to_typetracer(False) for x in self._contents],
+            [x._to_typetracer(forget_length) for x in self._contents],
             parameters=self._parameters,
         )
 

diff --git a/src/awkward/operations/__init__.py b/src/awkward/operations/__init__.py
@@ -78,7 +78,6 @@
 from awkward.operations.ak_to_arrow_table import *
 from awkward.operations.ak_to_backend import *
 from awkward.operations.ak_to_buffers import *
-from awkward.operations.ak_to_categorical import *
 from awkward.operations.ak_to_cupy import *
 from awkward.operations.ak_to_dataframe import *
 from awkward.operations.ak_to_feather import *

diff --git a/src/awkward/operations/ak_categories.py b/src/awkward/operations/ak_categories.py
@@ -18,7 +18,7 @@ def categories(array, highlevel=True):
     #ak.contents.IndexedOptionArray labeled with parameter
     `"__array__" = "categorical"`), then this function returns its categories.
 
-    See also #ak.is_categorical, #ak.to_categorical, #ak.from_categorical.
+    See also #ak.is_categorical, #ak.str.to_categorical, #ak.from_categorical.
     """
     # Dispatch
     yield (array,)

diff --git a/src/awkward/operations/ak_from_categorical.py b/src/awkward/operations/ak_from_categorical.py
@@ -23,8 +23,7 @@ def from_categorical(array, *, highlevel=True, behavior=None):
     size of the dataset. (Conversion to categorical is expensive; conversion
     from categorical is cheap.)
 
-    See also #ak.is_categorical, #ak.categories, #ak.to_categorical,
-    #ak.str.to_categorical, #ak.from_categorical.
+    See also #ak.is_categorical, #ak.categories, #ak.str.to_categorical.
     """
     # Dispatch
     yield (array,)

diff --git a/src/awkward/operations/ak_is_categorical.py b/src/awkward/operations/ak_is_categorical.py
@@ -15,7 +15,7 @@ def is_categorical(array):
     `"__array__" = "categorical"`), then this function returns True;
     otherwise, it returns False.
 
-    See also #ak.categories, #ak.to_categorical, #ak.from_categorical.
+    See also #ak.categories, #ak.str.to_categorical, #ak.from_categorical.
     """
     # Dispatch
     yield (array,)

diff --git a/src/awkward/operations/ak_to_categorical.py b/src/awkward/operations/ak_to_categorical.py