Hi, by chance I compared the info about registered parsers from our instance with https://nomad-lab.eu/prod/v1/gui/about/information and noticed that most of the available parsers are not registered in our Oasis. Is this intentional and if so is there a simple way to add all parsers of the official instance?
If you are using a current image, e.g. v1.2.1
, latest
, stable
and your nomad.yaml
is not disabling plugins, e.g. via plugins/include
or plugins/exclude
, the list of parsers shown in your UI should be very similar if not the same.
strange, I set up the instance quite a while ago but am now using the 1.2.1 image (see screenshot) I suppose but get the following list:
parsers: random, template, chaos, apbs, electrospinning, lightforge, battery, cg, graphene, tabular, archive, broken, octopus, crystal, wien2k, fhi-aims, broken
Is it possible, that if we specify custom parsers It doesn’t use all the default ones?
If you specify them via plugins/include
, only the ones listed there (plus a handful of build-in parsers like tabular, archive, broken, etc.) will be used.
The plugins
config has tree keys include
, exclude
and options
. With options
you would extend the potential list of possible parsers. With include
, exclude
you really set the list of parser to be used or not used. With include=*
you would enable all plugins in options
(the default options form the image + the ones from your nomad.yaml
).
Thanks, removing the plugin section and just adding the parsers via the option syntax apparently resolved the issue. Is this documented somewhere and I missed it?
Out of curiosity how would include=*
translate to yaml? As follows:
plugins:
include: "*"
options:
parsers/my_parser:
python_package: my_parser
...
Yes, but include: "*"
is semantically the same as just leaving it out.
Decide for yourself on the sufficiency of the documentation: Configuration - Documentation
Thanks, I was only looking at the “How to deploy an Oasis” and “how to write a custom parser” sections and obviously missed the link the Configuration documentation part even though I’m not fully sure if this would have been as helpful as the specific reply here