v1.0 will be the next major release of delicatessen. There are some important updates to the internal structure, how the bread is computed, and the dependencies. These changes warrant the jump to v1.0 since there are major changes for compatibility.
v1.0 will be limited to
SciPy v1.9.0+ which requires NumPy 1.18.5. This is a major change in the supported versions
SciPy v1.9.0 also necessitates the reduction of support to 3.8+
All previous versions will have their continual support, but future versions will be restricted to these versions.
The major change is the replacement of my partial_derivative function with SciPy's approx_fprime. The SciPy offers a simple solution to compute the bread (the Jacobian of the estimating equations). This is substantial faster when delicatessen interacts with >50 parameters. The speed difference is stark, so I made the decision to switch over.
Summary of changes
[x] Update setup.py to new versions and dropping support for 3.6 and 3.7
[x] Update GitHub tests for new versions
[x] Add support for Python 3.11
[x] Remove old API for ee_regression (as warned in previous versions)
v1.0 will be the next major release of
delicatessen
. There are some important updates to the internal structure, how the bread is computed, and the dependencies. These changes warrant the jump to v1.0 since there are major changes for compatibility.v1.0 will be limited to
All previous versions will have their continual support, but future versions will be restricted to these versions.
The major change is the replacement of my
partial_derivative
function with SciPy'sapprox_fprime
. The SciPy offers a simple solution to compute the bread (the Jacobian of the estimating equations). This is substantial faster whendelicatessen
interacts with >50 parameters. The speed difference is stark, so I made the decision to switch over.Summary of changes
setup.py
to new versions and dropping support for 3.6 and 3.7ee_regression
(as warned in previous versions)Other changes still TBD