Conversation
Allow overriding the spectroscopic slit via a slit_name parameter at both the global (params) and per-recipe (YAML) level. The parameter flows through to cmd["!OBS.slit"] in ScopeSim before OpticalTrain creation. The updateHeaders() slit handling now reads the actual slit from the FITS header (HIERARCH ESO INS OPTI3 NAME) instead of being hardcoded to C-38_1. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Set up a docs/ folder with Sphinx configuration and four pages: - Repository structure overview (modules, data flow, CI) - Workflow summary (all 9 simulation blocks with outputs) - How to write a workflow (YAML format, parameters, sources) - FITS Wrangler reference (module API, complete keyword catalogue) Also add .readthedocs.yaml build config and .claude/ to .gitignore. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
|
A problem with documentation is that it will deteriorate. Experience shows that we will not maintain it. We have a hard time maintaining outdated code. Case in point (both about not properly maintained software, and an example of documentation that will be outdated):
So my proposal: why not just run whatever you ran on demand? That is, we don't merge this in, but just ask the LLM to produce the info whenever we need it. Related, there must be a "summarize this github repo for me" service somewhere. If we do want to have this in the repo itself, then I suggest we automatically update it; e.g. run whatever you ran, but then in a github action that we run 'regularly'. |
2 participants
I made these for myself (with help from Claude) to get a better understanding of the layout of METIS_Simulations.
Feel free to reject the pull request, but I though if it's 90% accurate, it might be useful for Shannon and Yannis, and any of his minions to transition towards wrtiting YAML files for their AIT activities