NumPy User Guide

Overview: This lesson introduces NumPy as the fundamental library for scientific computing in Python, explaining its performance advantages through vectorization. Students will learn how to install the library across various platforms (Windows, Raspberry Pi, Conda, PyCharm) and resolve common installation hurdles like ImportError by managing environment variables and system dependencies.

Learning Outcomes:

• Define NumPy and identify its role in the scientific Python ecosystem.
• Explain why NumPy is significantly faster than standard Python loops using the concept of vectorization.
• Execute installation commands for various environments including Pip, Conda, and Raspberry Pi.

Overview: This lesson provides a comprehensive introduction to NumPy's ndarray object, covering its core attributes, creation methods, and basic mathematical operations. It extends into advanced topics including fancy indexing (boolean and integer-based), shape manipulation, and essential linear algebra routines. By the end of this module, learners will be able to efficiently store, transform, and analyze multi-dimensional data structures.

Learning Outcomes:

• Identify and interpret core ndarray attributes such as ndim, shape, and dtype.
• Execute array creation and manipulation using functions like linspace, reshape, vstack, and hstack.
• Apply elementwise operations, universal functions (ufuncs), and linear algebra solvers to numerical datasets.

Overview: This lesson covers the advanced mechanics of NumPy, focusing on precise data type management, handling overflow, and sophisticated I/O operations using genfromtxt. Students will master the internal logic of Broadcasting for arithmetic operations, the nuances of memory-level byte ordering, and the creation of structured/record arrays for heterogeneous datasets. The final sections detail the extensibility of NumPy through custom array containers and the formal protocols for subclassing ndarray.

Learning Outcomes:

• Manage data precision and mitigate overflow errors using NumPy's scalar types and info tools (iinfo, finfo).
• Implement flexible data ingestion from disk using genfromtxt with custom delimiters, headers, and column selections.
• Apply General Broadcasting Rules to predict and control interactions between arrays of differing shapes.

Overview: This lesson explores the advanced nuances of NumPy subclassing, specifically regarding memory management and downstream compatibility. It further examines how NumPy implements the IEEE 754 floating-point standard to handle special values and numerical exceptions, concluding with the mechanisms for interfacing NumPy arrays with low-level languages like C, C++, and Fortran.

Learning Outcomes:

• Manage memory references and avoid "gotchas" in custom ndarray subclasses using the .base attribute.
• Maintain downstream compatibility in subclasses by correctly implementing __array_wrap__ and method signatures.
• Identify and manipulate IEEE 754 special values (NaN, Inf) and configure global numerical exception behaviors.

Overview: This lesson explores the various methods for extending NumPy's functionality by interfacing with compiled languages like C, C++, and Fortran. It covers automated tools like f2py and Cython, manual wrapping with ctypes, the creation of high-performance universal functions (ufuncs), and advanced C-API techniques for array iteration, custom data types, and subtyping the ndarray.

Learning Outcomes:

• Compare and implement various methods for "gluing" compiled code to Python (f2py, Cython, ctypes).
• Construct and register custom NumPy universal functions (ufuncs) for single and multiple data types, including structured arrays.
• Utilize the NumPy C-API to perform efficient array iteration, handle broadcasting, and define user-defined data types or ndarray subtypes.