Integration of DL1DH Reader, Processor, and ImageMapper completed. Numerous bugfixes and improvements for compatibility with CTLearn.
Added option for serial/parallel writing modes to DL1DH Writer. Fixes to setuptools and conda build process.
Major Features And Improvements
* Finished changes to reader and image mapper APIs to be compatible with CTLearn
* Finished incorporating image mapper and DL1DH processor.
* Added example notebook.
* Added six new Transform functions for the data processor and moved to transform.py
* Added single threaded/parallel writes modes as a config option for DL1DataWriter.
* Updated conda package build files and setup.py dependencies.
Bug Fixes and Other Changes
* Fixed output_dir option of DataWriter to correctly write output files to directory.
* Added output_dir option to generate_runlists.py script.
* Moved example config files and runlists to /config folder
* Added separate expected_mc_events config setting for optimizing chunking on the MC_Events table.
* Fixed various issues with setuptools dependencies.
* Fixed FACT camera shape mismatch by upgrading to ctapipe-extra 0.2.17.
* Changed mapping table indices to start at 0 (removed 0 padding).
* Added option for camera rotation for LST, NectarCam, MAGICCam in image mapper.
* Fixed output_dir option in write_data.py script.
* Moved config files to config folder
* Changed indexing convention in image mapper.
* Changed get_image() to return a vector when map method is axial addressing
* Fixed status badge
Known Issues
* The current version of ctapipe (v0.6.2) does not correctly support a version tag that can be read by pkg_resources (instead __VERSION__ = "unknown") so it is not being written correctly to the output file headers.
* Calling file.close() in the destructor of CTAMLDataDumper causes a KeyError on some data files due to an empty NodeManager registry. The cause of this exception is unclear. The exception resulting from the call is currently being caught and ignored.