gretl4py 0.62 released
by Marcin Błażejowski
Hi gretl users,
I have just released *gretl4py 0.62*, available via PyPI and Conda (a
little later, maybe tomorrow).
This release brings a substantial extension and cleanup of the Python
API. Among the most notable changes are:
* an extended model-specific API, including post-estimation tests and
diagnostics;
* further development of the |gretl.Matrix|, |gretl.Dataset|,
|gretl.Bundle|, and |gretl.Package|interfaces;
* improved interoperability with NumPy and pandas;
* improved support for calling gretl function packages directly from
Python;
* further development of VAR/VECM functionality;
* improved plotting support;
* numerous smaller API improvements, fixes, and documentation updates.
The documentation has also been substantially revised and extended.
Installation via PyPI is as usual:
python3 -m pip install --upgrade gretl4py
For more information, documentation, examples, and source code, see:
https://gretl.sourceforge.net/gretl4py.html
<https://gretl.sourceforge.net/gretl4py.html>
Have fun!
Marcin
1 month, 1 week
- 1
- 0
gretl4py 0.62 pre-release
by Marcin Błażejowski
Hi,
here: https://sourceforge.net/projects/gretl/files/gretl4py/snapshots/
you'll find pre-release builds of gretl4py 0.62.
The most important change is the new dataset- and model-based testing API:
<python>
import gretl
d = gretl.get_data("denmark")
r1 = d.adf(2, [1])
m = gretl.ols("1 0 2 3 4").fit()
r2 = m.reset(squares_only=True)
</python>
The old API, i.e. obj.test(type: str, **kwargs), is still available, but
will be removed in version 0.7.
Marcin
1 month, 1 week
- 1
- 0