| Age | Commit message (Collapse) | Author |
|---|
|
As part of debugging the "Too many connections" error to the database,
this commit adds logging to track the creation of connections to the
database, and their eventual closing.
|
|
* wqflask/wqflask/metadata_edits.py: Remove
"get_allowable_sampledata_headers" import.
(display_phenotype_metadata): Use
"get_case_attributes" to fetch all the allowable
case attributes.
|
|
|
|
There is no need to create a database engine for requests for static
files like CSS files and javascript files. This commit checks for and
skips the creation of the engine for such requests.
|
|
Remove code for connecting to the database that is not actually used
in the application in any sensible form.
|
|
|
|
|
|
|
|
|
|
|
|
* wqflask/wqflask/metadata_edits.py (show_case_attribute_columns):
Create a connection and pass it to "get_case_attributes".
|
|
* wqflask/wqflask/metadata_edits.py: Run "black -l 79
wqflask/wqflask/metadata_edits.py"
|
|
* wqflask/wqflask/metadata_edits.py: Import
"gn3.db.sample_data.get_case_attributes".
(show_case_attribute_columns): New function/ end-point to show
case-attributes.
* wqflask/tests/integration/wqflask/test_metadata_edits.py: New
integration tests for the above.
* wqflask/tests/integration/wqflask/__init__.py: New file.
* wqflask/wqflask/templates/case_attributes.html: New template file to
display the above.
|
|
|
|
The logic for building the SQL query wasn't working correctly when
aliases were involved. The query also checks for a gene symbol's aliases
in the DB, but this wasn't properly handling the AND/OR logic, leading
to no results being shown when doing a "Get Combined" search involving a
gene symbol.
This also removed some unused code related to aliases.
|
|
For some reason that isn't clear to me yet, a logged-in user session ID
was set even when a user wasn't logged in, causing the user session code
to attempt to fetch user information for a non-existent user ID
|
|
It's not entirely clear to me exactly why the previous logic wasn't
working correctly, but this change (that removes the user cookie when
there's no user_session after the request and returns None if there
aren't any user_details) appears to fix the issue.
The main confusing part is why the user_cookie still exists even when
not logged in
|
|
metadata_edits.py
For some reason there were two places where it wasn't being called as a
function. Fixing this appears to have fixed the error
|
|
There were two places this needed to be checked and I previously missed
this one
|
|
property to user_session
Also changed the order of the dropdowns to have Sign Out after the user
drop-down, to more closely match NCBI's style
|
|
The function returns a call to mdb.connect, which only returns the
connection object. It was throwing an error due to trying to unpack
multiple items.
One thing I noticed while looking at this is that we should probably
decide on a single way to create these connections. In some other files,
we instead use database_connection from wqflask.database (instead of the
gn3 database_connector from gn3.db_utils)
|
|
Chore/ctl wgcna enhancements
|
|
Passing the data into `requests.post` as a `json=...` argument will
automatically encode the data to JSON and set up the correct
Content-Type header.
The call `json.dumps(post_data)` was pre-encoding the data to a JSON
string, that was the re-encoded to JSON yet again, which is not what
we want. This commit fixes that.X
|
|
|
|
Some of this was caused by heatmaps supporting code; that code should probably pass the traits differently than the way it does in the "trait_info_str" function
|
|
Bug/fix qtlreaper snp order
|
|
This is meant to deal with edge cases where there's only a single coordinate for a chromosome when doing interval mapping (which causes an error, since interval mapping requires that markers be linked within each chromosome)
|
|
end)
Not sure how best to do this; I just had the sort algorithm change "M" to "z" (which will always be sorted last, behind X and Y)
|
|
|
|
|
|
|
|
Feature/add rqtl pairscan
|
|
The parameter should never be interval if pair_scan is being run; not checking for this causes the --interval tag to pointlessly be included when running pair-scan, which also interferes with caching
|
|
Previously it would always register as True, due to being in the start_vars
|
|
|
|
|
|
|
|
Add missing argument to the command in the documentation
|
|
The CI has been migrated from penguin2.genenetwork.org to ci.genenetwork.org.
* README.md: Update CI badge subdomain.
|
|
Chore/gn2 enhancements
|
|
|
|
|
|
|
|
|
|
|
|
Replace pca rpy2 code
|
|
* test/requests/test-website.py: Run "black test/requests/test
test-website.py".
|
|
* test/requests/test-website.py: Delete parser option "-i" for running
integration tests.
(integration_tests): Delete it.
(integration_suite): Ditto.
(run_integration_tests): Ditto.
* test/requests/test_registration.py: Delete it.
|
|
* test/requests/test-website: These imports were deleted in an earlier
PR here: <https://github.com/genenetwork/genenetwork2/pull/574>
|
|
* test/requests/test-website.py: Delete "from wqflask import
app". This has the undesired effect of starting an instance of
Genenetwork2.
|
