| Age | Commit message (Collapse) | Author |
|---|
|
Set up the scaffolding for the flows and UI that will be used when
running QC against the uploaded R/qtl2 bundle.
This will be fleshed out later, and the UI is likely to change
somewhat, down the line.
|
|
Provide a better error message that can help the user figure out what
they did wrong and fix it for themselves instead of the older, cryptic
and extremely unhelpful error message.
|
|
Extract the common structure into a separate function and pass in
checkers that return the errors they find.
|
|
|
|
The ShortName is optional, and can be set to the same value as
Name. This commit removes the check that made it mandatory.
|
|
|
|
|
|
|
|
|
|
Fix the routing issues that were redirecting the flow to the wrong
endpoints.
|
|
|
|
Add a function to ensure the values in the geno files are all listed
in the control data under the "genotypes" key.
|
|
Add a QC step to the R/qtl2 bundle processing - check that all files
that are listed in the control file also exist in the bundle.
|
|
|
|
* Display any and all errors on the UI
* Move `validate_bundle` to QC module and refactor to use
`missing_files`
|
|
Add QC a function to list all files listed in the control file, and
another to list only the files missing from the bundle.
|
|
Improve the UI documentation by making it more explicit what each
section deals with to help aleviate confusion of what the different
features do.
|
|
|
|
|
|
Without specifying the `cursorclass`, the result is a tuple of the
form:
((query_col01_val, query_col02_val, ...), ...)
where the ellipsis specify the possibility of more than one of the
previous form.
We specify the DictCursor class instead so that the form changes to:
({query_col01_name: query_col01_val, ...}, ...)
which allows us to pick the value we want with a string index.
|
|
|
|
|
|
|
|
Update the file validation script and routes to use the redis prefix
for jobs.
|
|
|
|
Since the GNQC application might find itself sharing the same redis
server and/or database with other applications, add a prefix to its
keys.
|
|
It's easier, and more robust to put the data under a prefix within a
redis database rather than try to run a redis server, especially
within the test environment. This commit updates the tests to use a
redis prefix instead.
|
|
|
|
|
|
|
|
Fetch the sample/case names from the database rather than from a
static file in the repository.
Issue:
https://issues.genenetwork.org/issues/quality-control/read-samples-from-database-by-species
|
|
Build script to start the processing of the R/qtl2 bundle.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
We do not have the option to modify the request and then redirect with
the modification in place. To go around that, this commit creates an
intermediate step that informs the user of their progress, while
allowing us to store the filename for future steps.
|
|
Build a sha512 message digest from the file's name and the timestamp
of when the file was uploaded.
This ensures that the filename is URL-safe, and reduces chances of
name conflicts in the upload directory which could lead to data
corruption when processing the files and saving the files into the
database.
|
|
The ProbeSetData table does not allow `NULL` for the `value` column
for good reason - if there is no value, just don't have a record for
that particular sample.
|
|
`na.strings` has a default value of "NA" as stated in
https://kbroman.org/qtl2/assets/vignettes/input_files.html#CSV_files
quote:
> Missing value codes will be specified in the control file (as
> na.strings, with default value "NA") and will apply across all
> files, so a missing value code for one file cannot be an allowed
> value in another file.
for `comment.char`
> The CSV files can include a header with a set of comment lines
> initiated by a value specified in the control file as comment.char
> (with default value "#").
for `sep`:
The default separator is expected to be the comma, as stated in
https://kbroman.org/qtl2/assets/vignettes/input_files.html#field-separator
quote:
> If the data files use a separator other than a comma ...
|
|
|
|
There was a bug where the `na.strings` were not processed correctly if
the user called the `r_qtl.r_qtl2.file_data(...)` function without
explicitly providing the `process_*` arguments.
This commit fixes that.
|
