Closed metasoarous closed 6 years ago
adding chalkshot from defunct issue
@csmall lmk what you think /fh/fast/matsen_e/processed-data/partis/laura-mb/latest/info.yaml
@psathyrella Looks good! Here are my requests/suggestions:
dataset-id
; should be unique between any combination of study, build version, and any other setting which might lead to a separate run of partispartition-cluster-annotations
as appropriatesequence-metadata
pointer (probably in general via a sequence-metadata-filename` attr); I could handle whatever is easier for you but imagine that specifying either per sample or per dataset would work just finePerhaps something like this would accommodate these issues?
dataset-id: blablah
sequence-metadata-filename: /src/bin/whatevs.csv
samples:
<sample-id>:
meta: ...
partition-file: ...
cluster-annotation-file: ...
seeds:
<seed-id>:
meta: ?
partition-file: ...
cluster-annotation-file: ...
ok v2 /fh/fast/matsen_e/processed-data/partis/laura-mb/latest/info.yaml
I don't remember where you put the translation files, so I just put a placeholder in there for 'per-sequence-meta-file'.
Also, not sure if it was obvious, but general design philosophy is: write initial file with all meta data from datascripts/meta (hence all the empty dicts), then gradually fill things in with actual paths as we check that they've run successfully.
@psathyrella I've got this more or less "working" now. There are still a few more kinks to work out, like #228, but I think we're getting close to a point where it would be nice to have a complete yaml fleshed out with all of the data you've build so far, so I can get all of the datasets built out (there were some schema updates that came out of the yaml refactor, so I'll need to rebuild all the datasets).
ok, I should be able to make yamls for everything in the next few days. LMK if that isn't soon enough.
That's perfect. Thanks.
We need a yaml file per dataset, according to spec from #216.