Arraykit

Common errors

• ModuleNotFoundError: No module named 'arraykit'

  cause The 'arraykit' package has not been installed in the current Python environment.
  fix Run `pip install arraykit` to install the package.

• AttributeError: module 'arraykit' has no attribute 'astype_array'

  cause The specific function `astype_array` was added in version 1.1.0. This error occurs if you are using an older version of arraykit or have a typo in the function name.
  fix Upgrade your arraykit installation to the latest version (`pip install --upgrade arraykit`) or verify the function name against the documentation for your installed version.

• ValueError: invalid literal for int() with base 10: 'a'

  cause This error typically occurs when using `astype_array` (or similar functions) to convert a sequence containing non-numeric strings to a numeric NumPy dtype (e.g., `np.int64`, `np.float64`).
  fix Ensure that all elements in the input sequence are compatible with the target `dtype`. If the sequence contains mixed types (e.g., strings and numbers), consider using `dtype=object` or pre-process the data to handle non-numeric values.

Warnings

• breaking NumPy `datetime64` scalar lookups in `AutoMap` and `FrozenAutoMap` (internal `arraykit` classes) explicitly require matching units for comparisons and indexing.
  Fix: Ensure that `datetime64` scalars used for lookups or comparisons have identical time units (e.g., 'ns', 's') to the `datetime64` values stored within `AutoMap` or `FrozenAutoMap`.
• gotcha Arraykit relies heavily on NumPy. Users unfamiliar with NumPy dtypes and array-like structures may encounter unexpected behavior or errors when passing incompatible data types to arraykit functions.
  Fix: Familiarize yourself with NumPy's type system and how it handles type conversions. Ensure data passed to arraykit functions conforms to expected array-like structures.
• gotcha As a utility library, arraykit functions are typically imported directly from the top-level package (e.g., `from arraykit import some_function`). Attempting to access functions via nested modules might fail or be incorrect.
  Fix: Always refer to the official documentation or source code for the correct import path of specific functions. Avoid `from arraykit import *` to prevent namespace pollution and improve clarity.

Install

• pip install arraykit Install stable version

Imports

• astype_array

  from arraykit import astype_array

  Commonly used for flexible conversion of various inputs to NumPy arrays.
• is_objectable

  from arraykit import is_objectable

  Useful for checking if a value can be placed in a NumPy object array.
• delimited_to_arrays

  from arraykit import delimited_to_arrays

  For parsing delimited string data into arrays.

Quickstart

This quickstart demonstrates how to use `astype_array` to convert various Python sequences into NumPy arrays with optional dtype specification. It also shows `is_objectable` for checking type compatibility with object arrays.

import numpy as np
from arraykit import astype_array

# Convert a list to a NumPy array with a specified dtype
data_int = [1, 2, 3, 4]
array_float = astype_array(data_int, dtype=np.float64)
print(f"Original list: {data_int}")
print(f"Converted array (float64): {array_float}, dtype: {array_float.dtype}\n")

# Convert mixed data; typically results in an object dtype if types are incompatible
mixed_data = [10, 'hello', 20.5, True]
array_object = astype_array(mixed_data)
print(f"Original mixed data: {mixed_data}")
print(f"Converted array (object): {array_object}, dtype: {array_object.dtype}\n")

# Using is_objectable
from arraykit import is_objectable
print(f"Is 'hello' objectable? {is_objectable('hello')}")
print(f"Is np.datetime64('2023-01-01') objectable? {is_objectable(np.datetime64('2023-01-01'))}")