Collect dependencies in one place

Author: phofl

pandas_expr/_accessor.py

@@ -14,7 +14,7 @@ class Accessor:
     """
 
     def __init__(self, series):
-        from dask_expr import Series
+        from pandas_expr import Series
 
         if not isinstance(series, Series):
             raise ValueError("Accessor cannot be initialized")
@@ -53,7 +53,7 @@ def _delegate_method(obj, accessor, attr, args, kwargs):
         return maybe_wrap_pandas(obj, out)
 
     def _function_map(self, attr, *args, **kwargs):
-        from dask_expr._collection import new_collection
+        from pandas_expr._collection import new_collection
 
         return new_collection(
             FunctionMap(self._series.expr, self._accessor_name, attr, args, kwargs)

pandas_expr/_categorical.py

@@ -66,7 +66,7 @@ def as_known(self, **kwargs):
         """
         if self.known:
             return self._series
-        from dask_expr._collection import new_collection
+        from pandas_expr._collection import new_collection
 
         categories = (
             new_collection(PropertyMap(self._series.expr, "cat", "categories"))
@@ -80,7 +80,7 @@ def as_unknown(self):
         if not self.known:
             return self._series
 
-        from dask_expr import new_collection
+        from pandas_expr import new_collection
 
         return new_collection(AsUnknown(self._series.expr))
 
@@ -115,7 +115,7 @@ def codes(self):
                 "`df.categorize()` beforehand to ensure known categories"
             )
             raise AttributeNotImplementedError(msg)
-        from dask_expr._collection import new_collection
+        from pandas_expr._collection import new_collection
 
         return new_collection(PropertyMap(self._series.expr, "cat", "codes"))
 

pandas_expr/_collection.py

@@ -3,14 +3,14 @@
 import functools
 
 import pandas as pd
-from dask.base import collections_to_dsk
 from fsspec.utils import stringify_path
 from pandas._typing import Axes, Dtype
 from pandas.core.accessor import CachedAccessor
 
 from pandas_expr import _expr as expr
 from pandas_expr._categorical import CategoricalAccessor
 from pandas_expr._concat import Concat
+from pandas_expr._deps import collections_to_dsk
 from pandas_expr._expr import Eval, no_default
 from pandas_expr._merge import JoinRecursive, Merge
 from pandas_expr._reductions import (

pandas_expr/_concat.py

@@ -1,9 +1,8 @@
 import functools
 
 import pandas as pd
-from dask.dataframe.dispatch import make_meta, meta_nonempty
-from dask.dataframe.utils import strip_unknown_categories
 
+from pandas_expr._deps import make_meta, meta_nonempty, strip_unknown_categories
 from pandas_expr._expr import Blockwise, Projection
 
 

pandas_expr/_deps.py

@@ -0,0 +1,7 @@
+from dask.base import collections_to_dsk, normalize_token, tokenize  # noqa: F401
+from dask.dataframe.core import make_meta  # noqa: F401
+from dask.dataframe.dispatch import meta_nonempty  # noqa: F401
+from dask.dataframe.utils import (  # noqa: F401
+    clear_known_categories,
+    strip_unknown_categories,
+)

pandas_expr/_expr.py

@@ -7,16 +7,17 @@
 from collections import defaultdict
 from collections.abc import Generator, Mapping
 
-import dask
 import pandas as pd
 import toolz
-from dask.base import normalize_token
-from dask.dataframe.core import make_meta
-from dask.dataframe.dispatch import meta_nonempty
-from dask.dataframe.utils import clear_known_categories
 from pandas.compat._optional import import_optional_dependency
 from pandas.core.dtypes.common import is_array_like
 
+from pandas_expr._deps import (
+    clear_known_categories,
+    make_meta,
+    meta_nonempty,
+    normalize_token,
+)
 from pandas_expr._util import M, _tokenize_deterministic, apply, funcname, ishashable
 
 replacement_rules = []
@@ -1893,6 +1894,8 @@ def _task(self):
     def _execute_task(graph, name, *deps):
         for i, dep in enumerate(deps):
             graph["_" + str(i)] = dep
+        import dask
+
         return dask.core.get(graph, name)
 
 

pandas_expr/_groupby.py

@@ -2,9 +2,9 @@
 
 import numpy as np
 import pandas as pd
-from dask.dataframe.dispatch import meta_nonempty
 
 from pandas_expr._collection import DataFrame, Series, new_collection
+from pandas_expr._deps import meta_nonempty
 from pandas_expr._expr import Projection
 from pandas_expr._reductions import Reduction
 from pandas_expr._util import M, apply

pandas_expr/_reductions.py

@@ -1,6 +1,6 @@
 import pandas as pd
-from dask.dataframe.core import make_meta, meta_nonempty
 
+from pandas_expr._deps import make_meta, meta_nonempty
 from pandas_expr._expr import Elemwise, Expr, Index, Projection
 from pandas_expr._util import M, apply
 

pandas_expr/_util.py

@@ -4,7 +4,7 @@
 from collections.abc import Sequence
 from types import LambdaType
 
-from dask.base import normalize_token, tokenize
+from pandas_expr._deps import normalize_token, tokenize
 
 
 def _convert_to_list(column) -> list | None:

pandas_expr/_version.py

@@ -51,8 +51,8 @@ def get_config() -> VersioneerConfig:
     cfg.VCS = "git"
     cfg.style = "pep440"
     cfg.tag_prefix = ""
-    cfg.parentdir_prefix = "dask-"
-    cfg.versionfile_source = "dask_expr/_version.py"
+    cfg.parentdir_prefix = "pandas-"
+    cfg.versionfile_source = "pandas_expr/_version.py"
     cfg.verbose = False
     return cfg