Releases: dynamicslab/pysensors
JOSS paper submission release
pre-JOSS release
This release introduces significant changes and features to the API which include
Reconstruction
Introduces regularized reconstruction, based on Bayesian MAP state estimate and pointwise standard deviation of the error std. Renames previous reconstruction method as unregularized reconstruction.
Optimizers
Adds the following optimizers
GQR(General QR) - Provides access to the pivoting procedure for spatially constrained sensor placement.TPGR(Two Point GReedy) - Handles sensor undersampling and oversampling, and provides pointwise sensor interaction landscapes.
Additional Features
- Adds three spatial constraint types - region, predetermined, and distance.
 - Adds support for different shapes of spatial constraints - 
circle,cylinder,line,parabola,ellipse,polygonandUserDefinedConstraints - Adds custom basis support (e.g., DMD basis).
 - Adds uncertainty heatmaps showing noise propagation in sensor measurements to state reconstruction.
 
Example Notebooks
Introduces these example notebooks
- Constrained sensing for simulation of a fuel rod - demonstrates the use of spatial constraints.
 - Two-point greedy (
TPGR) sensor placement on sea surface temperature data - demonstrates uncertainty quantification and landscapes. - Reconstruction comparison on Olivetti Dataset - compares unregularized vs regularized reconstruction using 
TPGRoptimizer, and compares theQRandTPGRoptimizers with unregularized and regularized reconstruction. 
Updates all existing notebooks to comply with the new functionalities.
Add warning
This release merely adds a warning when the number of sensors exceeds the number of samples and the CCQR optimizer is used. In this case sensors may be selected from constrained regions. See #14.
Minor bug fix: cross-validation with SSPOC
Minor (one line) bug fix related to #13.
Zenodo DOI
This release is the same as the previous one. It was created to be archived by Zenodo to produce a DOI.
Additional JOSS revisions
This release contains minor updates to documentation and the JOSS paper made in response to feedback from JOSS reviewers.
JOSS Revisions
This release contains a collection of changes, mostly to documentation and notebooks, made based on journal of open source software (JOSS) feedback: openjournals/joss-reviews#2828.
Refactoring of PySensors
In addition to fixing minor bugs and adding content to the examples, the primary change in this release is the renaming of the SensorSelector class to SSPOR (Sparse Sensor Placement Optimization for Reconstruction). See #8 for more details.
Test automatic versioning
This release is essentially the same as v0.2.0, but it is designed to test a GitHub action that should automatically deploy it to PyPI.
Automatic versioning
Incremental improvement on v0.2.0 allowing for automatic versioning via setuptools_scm_git_archive.