Make __repr__() more efficient

View casting back to ndarray prevents the overhead in array creation with Galois field arrays. Each element displayed is re-created due to the scalar to 0-D array conversion done with Galois field arrays.

galois/_fields/_main.py

@@ -2736,31 +2736,28 @@ def dot(self, b, out=None):
 
  def __str__(self):
  return self.__repr__()
- # formatter = type(self)._formatter(self)
-
- # with np.printoptions(formatter=formatter):
- # string = super().__str__()
-
- # return string
 
  def __repr__(self):
+ # View the array as an ndarray so that the scalar -> 0-D array conversion in __array_finalize__() for Galois field
+ # arrays isn't continually invoked. This improves performance slightly.
+ x = self.view(np.ndarray)
+
+ separator = ", "
+ prefix = "GF("
+ order = type(self)._order_str
+ suffix = ")"
  formatter = type(self)._formatter(self)
 
- cls = type(self)
- class_name = cls.__name__
- with np.printoptions(formatter=formatter):
- cls.__name__ = "GF" # Rename the class so very large fields don't create large indenting
- string = super().__repr__()
- cls.__name__ = class_name
-
- # Remove the dtype from the repr and add the Galois field order
- dtype_idx = string.find("dtype")
- if dtype_idx == -1:
- string = string[:-1] + f", {cls._order_str})"
- else:
- string = string[:dtype_idx] + f"{cls._order_str})"
+ string = np.array2string(x, separator=separator, prefix=prefix, suffix=suffix, formatter=formatter)
 
- return string
+ # Determine the width of the last line in the string
+ idx = string.rfind("\n") + 1
+ last_line_width = len(string[idx:] + ", " + order + suffix)
+
+ if last_line_width <= np.get_printoptions()["linewidth"]:
+ return prefix + string + ", " + order + suffix
+ else:
+ return prefix + string + ",\n" + " "*len(prefix) + order + suffix
 
 
 ###############################################################################