This version fixes a few issues found in the previous version. Firstly, the rebin function has been updated to increase the robustness of the conditions used to determine whether the spectrum is either binned or interpolated. Secondly, when generating the planet spectrum convolved with the instrument response, the stellar spectrum and the planet spectrum must be rebinned to the same wavelength grid to permit matrix multiplication. In the previous version the model stellar spectrum was downsampled to the model planet spectrum resolution to achieve this if the model planet spectrum was provided at the lower resolution (which is usually the case). This could result inconsistencies in the the representation of the stellar spectrum in terms of its final resolution depending on the planet spectrum used. Although the impact of this effect in most cases will not be significant, we have chosen to change the code so that the model planet spectrum is always rebinned to the model stellar spectrum resolution. This eliminates the above issue, and the impact on the final planet spectrum is insignificant so that it maintains its fidelity. Other changes made were as follows. 1) The NIRISS wavelength solution is processed differently by first fitting a polynomial solution which is then used to assign wavelengths per pixel. This results in smoother pixel-to-pixel wavelength gradient than using the previous method which directly interpolates the wavelength solution obtained from file. 2) Small fix made to code that deals with selecting the complex planet spectrum. 3) Fix made in the results.py file that allows it to be run from within an IDE.
When running the example cases given in the tutorial using this version, there is no significant change in the results obtained for both signal and noise compared to the previous version.