Closed gcanasherrera closed 2 weeks ago
What's in your global config file that it just tried to read?
In principle, there shouldn't be any global config file. I haven't used cobaya-install to install anything, I literally only did pip install cobaya
and tried the gaussian example in the docs.
Debug: True doesn't return any more info. What's the variable I should print within cobaya source code to figure out the config file it's trying to load?
See load_config_file() source for the file name it constructs.
See load_config_file() source for the file name it constructs.
So, I've put some prints in the function. Here is the output:
ketelmeer [143] > cobaya-run gaussian.yaml
2024-09-20 23:03:43,493 [output] Output to be read-from/written-into folder 'chains', with prefix 'gaussian'
2024-09-20 23:03:43,494 [output] Found existing info files with the requested output prefix: 'chains/gaussian'
2024-09-20 23:03:43,609 [model] Parameters were assigned as follows:
2024-09-20 23:03:43,609 [model] - gaussian_mixture:
2024-09-20 23:03:43,609 [model] Input: ['a', 'b']
2024-09-20 23:03:43,609 [model] Output: ['derived_a', 'derived_b']
2024-09-20 23:03:43,609 [gaussian_mixture] Initializing
In load_config_file
get_config_path(): /home/canasherrera/.config/cobaya
packages_path_config_file: config.yaml
This file: /home/canasherrera/.config/cobaya/config.yaml
does exist:
ketelmeer [162] > pwd
/home/canasherrera/.config/cobaya
ketelmeer [163] > ls
config.yaml
However, it is empty.
Not sure how an empty file could result, looks like it must have come from the install() function being called sometime
Still not sure how it happens, but changed to return empty dict in this case, which hopefully leads to more helpful error message https://github.com/CobayaSampler/cobaya/commit/8eb34eb052cf402ea9c0b86b2b61ef4cd39a036e
Dear @cmbant @JesusTorrado,
I have encountered for the very first time a very weird error when running cobaya in a new machine that never saw before. I pip install cobaya as usual, and it does not complain when I import it. However, when I try to simply run the gaussian example, I get the following error.
I'm running out of possible ideas, any suggestion for trouble-shooting?