In recent years, X-ray speckle tracking techniques have emerged as viable tools for wavefront metrology and sample imaging applications, and have been actively developed for use at synchrotron light sources. Speckle techniques can recover an image free of aberrations and can be used to measure wavefronts with a high angular sensitivity. Since they are compatible with low-coherence sources they can be also used with laboratory X-ray sources. A new implementation of the ptychographic X-ray speckle tracking method, suitable for the metrology of highly divergent wavefields, such as those created by multilayer Laue lenses, is presented here. This new program incorporates machine learning techniques such as Huber and non-parametric regression and enables robust and quick wavefield measurements and data evaluation even for low brilliance X-ray beams, and the imaging of low-contrast samples. To realize this, a software suite was written in Python 3, with a C back-end capable of concurrent calculations for high performance. It is accessible as a Python module and is available as source code under Version 3 or later of the GNU General Public License.