Inconsistent behavior of np.linalg.solve with singular matrices in NumPy 1.24.0
I'm getting frustrated with I'm updating my dependencies and I've been struggling with this for a few days now and could really use some help... I'm relatively new to this, so bear with me. I’m working with an scenario when using `np.linalg.solve` to solve a system of linear equations where the coefficient matrix is singular. For some inputs, it throws a `LinAlgError: Singular matrix`, but for others, it returns a solution that doesn't seem valid. Here’s the code snippet I’m working with: ```python import numpy as np # Coefficient matrix (singular) A = np.array([[1, 2], [2, 4]]) # Right-hand side vector b = np.array([3, 6]) try: x = np.linalg.solve(A, b) print('Solution:', x) except np.linalg.LinAlgError as e: print('behavior:', e) ``` When I run this code, I get the expected `LinAlgError`, which is fine. However, if I change `b` to `np.array([1, 2])`, it returns a solution `Solution: [0.5, 1.0]`, which I suspect might not be correct since `A` is still singular. Shouldn’t it consistently raise an behavior regardless of the values in `b`? I’ve checked the documentation, and it mentions that a singular matrix want to have a unique solution, so I’m confused why I’m getting a valid output in some cases. Is there a way to handle or check for singular matrices more reliably before attempting to solve the system? I'm using NumPy version 1.24.0. Any insights or recommendations would be greatly appreciated! I'm working on a application that needs to handle this. What am I doing wrong? The project is a microservice built with Python. I've been using Python for about a year now. Has anyone dealt with something similar?