Open holtgrewe opened 10 months ago
New-style imports deposit files in external storage. We thus need to make projects known to varfish. This should be done in the ~/.varfishrc.toml
file.
Here, is how to create a list of projects in general in toml
# ...
[[projects]]
uuid = "..."
[[projects]]
uuid = "..."
This will be loaded as {'projects': [{'uuid': '...'}, {'uuid': '...'}]}
in JSON/Python.
Users can configure projects with the following schema:
[[projects]]
title = "..." # optional; user-readable project title
uuid = "..." # SODAR project UUID
# protocol to use for import
import_data_protocol = "s3" # one of "s3" | "http" | "https" | "file"
import_data_path = "..." # path prefix to use
import_data_port = 80 # optional; port to user for connecting on import
import_data_user = "user" # user/S3 access key
import_data_password = "key" # password/S3 secret key to use
We should support users with the possibility to download these settings via the following command. This should fetch the settings from above from the server and append to the projects
configuration in the TOML.
varfish-cli projects project-load-config PROJECT_UUID
This follows the phenopackets YAML format supported by VarFish Server.
General note on files:
varfish-cli
Notes on individuals' files:
s3://varfish-server/seqmeta/enrichment-kits
and refers to the internal filesNotes on family files:
Precondition:
~/.varfishrc.toml
Then:
Is your feature request related to a problem? Please describe. The new-style imports (based on depositing files in an external storage and registering the case as phenopackets) is currently unsupported in VarFish.
Describe the solution you'd like Implement the import.
varfish-cli projects project-load-config PROJECT_UUID
(see below)Describe alternatives you've considered N/A
Additional context N/A