Fix typos

docs/api.md

@@ -347,7 +347,7 @@ For each standard type values are converted to the type using the type as constr
 
 When using the collection definition from `typing` the inner types will also be checked recursively. For example, one can define a `#!python typing.Tuple[int,float,typing.List[int]]` type and the conversion will be done accordingly.
 
-Reading sequences and dictionaries is done using the `yaml` syntax. Notice that tuples and sets use the same syntax as lists, where the conversion happens after loading the arguments. For dictionaries, the keys must be space seperated so the entire dictionary should be wrapped in string quotation marks.
+Reading sequences and dictionaries is done using the `yaml` syntax. Notice that tuples and sets use the same syntax as lists, where the conversion happens after loading the arguments. For dictionaries, the keys must be space separated so the entire dictionary should be wrapped in string quotation marks.
 
 ```console
 $ python train_model.py --worker_inds=[2,18,42] --worker_names="{2: 'George', 18: 'Ben'}"

pyrallis/utils.py

@@ -90,7 +90,7 @@ def get_item_type(container_type: Type[Container[T]]) -> T:
 
 def _mro(t: Type) -> List[Type]:
     # TODO: This is mostly used in 'is_tuple' and such, and should be replaced with
-    # either the build-in 'get_origin' from typing, or from typing-inspect.
+    # either the built-in 'get_origin' from typing, or from typing-inspect.
     if t is None:
         return []
     if hasattr(t, "__mro__"):
@@ -121,13 +121,13 @@ def is_set(t: Type) -> bool:
 
 
 def is_dataclass_type(t: Type) -> bool:
-    """Returns wether t is a dataclass type or a TypeVar of a dataclass type.
+    """Returns whether t is a dataclass type or a TypeVar of a dataclass type.
 
     Args:
         t (Type): Some type.
 
     Returns:
-        bool: Wether its a dataclass type.
+        bool: Whether its a dataclass type.
     """
     return dataclasses.is_dataclass(t) or (
             tpi.is_typevar(t) and dataclasses.is_dataclass(tpi.get_bound(t))

pyrallis/wrappers/docstring.py

@@ -83,7 +83,7 @@ def get_attribute_docstring(
 
 
 def _contains_attribute_definition(line_str: str) -> bool:
-    """Returns wether or not a line contains a an dataclass field definition.
+    """Returns whether or not a line contains a an dataclass field definition.
 
     Arguments:
         line_str {str} -- the line content

pyrallis/wrappers/field_wrapper.py

@@ -57,7 +57,7 @@ def arg_options(self) -> Dict[str, Any]:
         well as Python's type annotations.
 
         By passing additional keyword arguments to the `field()`
-        function, the autogenerated arguments can be overwriten,
+        function, the autogenerated arguments can be overwritten,
         giving access to all of the usual argparse features know and love.
 
         NOTE: When passing an `action` keyword argument, we remove all the

pyrallis/wrappers/wrapper.py

@@ -1,4 +1,4 @@
-"""Abstact Wrapper base-class for the FieldWrapper and DataclassWrapper."""
+"""Abstract Wrapper base-class for the FieldWrapper and DataclassWrapper."""
 
 from abc import ABC, abstractmethod
 from typing import Generic, Optional, List

tests/conftest.py

@@ -129,18 +129,18 @@ class TaskHyperParameters(TestSetup):
         num_units: int = 8  # units per layer
         activation: Activations = field(default=Activations.TANH)  # activation function
         use_batchnorm: bool = (
-            False  # wether or not to use batch normalization after each dense layer
+            False  # whether or not to use batch normalization after each dense layer
         )
-        use_dropout: bool = True  # wether or not to use dropout after each dense layer
+        use_dropout: bool = True  # whether or not to use dropout after each dense layer
         dropout_rate: float = 0.1  # the dropout rate
         use_image_features: bool = (
-            True  # wether or not image features should be used as input
+            True  # whether or not image features should be used as input
         )
-        use_likes: bool = True  # wether or not 'likes' features should be used as input
+        use_likes: bool = True  # whether or not 'likes' features should be used as input
         l1_reg: float = 0.005  # L1 regularization coefficient
         l2_reg: float = 0.005  # L2 regularization coefficient
 
-        # Wether or not a task-specific Embedding layer should be used on the 'likes' features.
+        # Whether or not a task-specific Embedding layer should be used on the 'likes' features.
         # When set to 'True', it is expected that there no shared embedding is used.
         embed_likes: bool = False
 
@@ -177,7 +177,7 @@ class HyperParameters(TestSetup):
 
         shared_likes_embedding: bool = True
 
-        # Wether or not to better filtering of liked pages
+        # Whether or not to better filtering of liked pages
         use_custom_likes: bool = True
 
         # Gender model settings