Restrictive version pins causing issues with upstream libraries.

[cmvcordova]

Hi heatgeo team, thanks for the great work!

heatgeo is used in multiple libraries that build on it. Many of these interact with other tools in the scientific computing ecosystem, such as numpy, scikitlearn and the like. When trying to create projects that build on these libraries AND heatgeo, there's often "unresolvable" package mismatches due to the overly restrictive nature of the exact version pins. While I'm sure there was a rationale behind the decision to pin exact versions, would it be a possibility to relax version constraints? e.g.

# FROM (current - too restrictive):
'numpy==1.23.5'

# TO (flexible):
'numpy>=1.23.5'

and maybe other similarly restrictive packages?

graphtools==1.5.3 → graphtools>=1.5.3
networkx==2.8.6 → networkx>=2.8.6
PyGSP==0.5.1 → PyGSP>=0.5.1
s-gd2==1.8.1 → s-gd2>=1.8.1
scikit-learn==1.2.2 → scikit-learn>=1.2.2

Thanks in advance!

[cmvcordova]

Created PR #2 to address this issue: #2

[cmvcordova]

Since merged. Closing! Thank you to the team.