* As part of https://github.com/chanzuckerberg/cellxgene/pull/1548 we
accidentally removed the part of the "get_or_else_dev_env_default"
function that allowed users to override the environment variables.
This commit adds that back.
* When environment.default was changed from a shell script file to json,
the shell commands used to parameterize DATASET were not evaluated.
This commit fixes this issue as well.
Update documentation
* on installing the server
* on how to get nodejs hot-reloading to work
* on running standard and annotations smoke tests with hot-reloading
There seem to be breaking changes in Chrome that are causing this to
fail. We've weighed the risk of disabling the feature with the issues
we've encountered using it and decided to disable it for now.
6.7 does not have the `hidden` flag used in the code. Users building the
app with an older version of click within the current range specified by
requirements.txt may fail.
* Refactor CSS layout and react logic for layout
Fixes https://github.com/chanzuckerberg/cellxgene/issues/1022
* Menubar should wrap inside middle pane instead of overlapping left
sidebar when window is scrunched
* cellxgene should have a minimum width of 1240px
1. Replace absolute positioning and dimension calculation with css grid
2. Use flexbox for wrapping menubar buttons
* Middle pane (graph) can calculate its own size
* Removing components calculating their size/position relative to
eachother increases modularity, decreases use of global variables
* Improved some scrollbar behavior
* Removed responsive reducer, propagating window size to components
triggers unnecessary events and encourages breaking modularity; doing
this made some components state agnostic
Reference: https://css-tricks.com/snippets/css/complete-guide-grid/
* Reposition the continuous legend
* Small fixes
* Respond to feedback from @colinmegill
* Respond to feedback from @colinmegill
Add more documentation on the renderGraph method.
* Add server plugin system
Plugins are optional modules loaded at runtime. Specification:
* Plugins are loaded from the server.plugins module (directory
server/plugins)
* The import_plugins method is run as part of the initialization of the
server module in __init__.py
* Add plugins to the EB build process
* Remove bit of dead code
* Respond to feedback from @bmccandless
* Return empty colors for .cxg v0.0 files
Fixes https://github.com/chanzuckerberg/cellxgene/issues/1440
The CxgAdaptor.get_colors method currently assumes that the .cxg file has
cxg_group_metadata. As a result, the /api/v0.2/colors endpoint always fails for
.cxg v0.0 files.
* Add test fixture
Fixes https://github.com/chanzuckerberg/cellxgene/issues/1433
When selection is deselected the world is reset to the universe the
crossfilter is reset to that in the resetCache, including the embeddings
in the layout_XY dim. However, the embedding selection stays the same.
If the embedding selected is not the default, the embedding shown to the
user will be different than the embedding layout_XY in the crossfilter,
causing lasso selections to be made against the wrong embedding
coordinates
* Add user-defined category-label colors
Fixes https://github.com/chanzuckerberg/cellxgene/issues/1152
As described in https://github.com/chanzuckerberg/cellxgene/issues/1307
* Respond to feedback from @bkmartinjr in nodejs
* Respond to feedback from @bkmartinjr in python
* Add tests to the server module
* Autoformat python, run linter
* Make colors_get error handling specific
* Respond to feedback from @bkmartinjr
* Respond to feedback from @bkmartinjr
* Fix whitespace
* Fix python lint errrors
* Update documentation
* Add --disable-user-colors option to launch and cxgtool.py
* Fix python formatting
* Rename '--disable-user-colors' to '--disable-custom-colors'
* Cleanup the backend-dev convenience method
* Add the frontend_dev convenience method
frontend_dev is a soup-to-nuts convenience method for setting up the FE
development environment with node running a the client code on port 3000
with the a separate cellxgene package serving the API over port 5005 in
the background.
The script can be run from Finder.
* Update the developer scripts documentation
* Remove the 'test' make target in the client Makefile
Rationale:
* Given how long the smoke tests take to run, it is unlikely that
developers will want to run all tests together.
* It is unlikely that developers will have set up the backend server
properly for the tests to pass.
* Available commands should be safe-ish and not lend themselves to
confusing errors.
* You can still group tests by concatenating them in a make command, as
in `make unit-test smoke-test`.
* This target isn't used in any of our CI pipelines -- KISS.
* Some version of python3...
* Minor typos in docs
* Respond to feedback from @bkmartinjr
* Make adjustments so that DATASET path is predictable
* Simplify environment defaults a bit
Fixes https://github.com/chanzuckerberg/cellxgene/issues/1349
For more information see https://github.com/d3/d3-format
Note that does not _fully_ fix the issue described in #1349, but rather
makes the formatting issue far less likely. It is _still_ possible for
this to occur if the difference between two ticks in axes happes in the
a significant digit cropped by the scientific notation format
Currently, if there is no build directory, the `make install` target
will attempt to cd into the build directory, fail, and run pip install
-e . in the root directory anyway. This causes cellxgene to be installed
from the source tree instead of what the user would expect.
This commit changes the behavior such that the `make install` will fail
if there is no build directory.
See sample of current behavior below:
```
venv❯ cellxgene prepare example-dataset/pbmc3k.h5ad
[cellxgene] Starting CLI...
Error: [cellxgene] cellxgene prepare has not been installed. Please run
`pip install cellxgene[prepare]` to install the necessary requirements.
~/workspace/cellxgene mweiden/446-custom-color-palette*
1 venv❯ pip install cellxgene[prepare]
zsh: no matches found: cellxgene[prepare]
```
Fix:
Wrap cellxgene[prepare] in single quotes.
* Add color mapping to the bar chart.
* Change histogram color generation from Viridis to InterpolateCool.
* Coloring of Histrogram based up the x axis instead of y axis.
Coloring of Histrogram based up the x axis instead of y axis.
* Respond to feedback from @colinmegill
* Only color histograms that are selected for colorby
* Add some small refactors to BrushableHistogram's componentDidUpdate
* Fix histogram coloring and binning
* Reuse binning functions from util/dataframe/histogram.js; this fixes
an issue with there being near-zero width bins
* Fix color mapping so that it matches the scale in the legend
* Do not attempt to plot bins if the calculated binWidth is zero; this
can happen if all values are the same
* Refactor the function that draws the histogram a bit
* Respond to feedback from @bkmartinjr
Co-authored-by: Donald Paul Herman <Donaldpherman@hotmail.com>
* Make sure apt is up to date before pulling hdf5
* Only install py dev reqs in cxg release vs anndata master test
* Don't need bu flag when using sed on ubuntu
* Don't re-install package reqs in python x anndata ver tests
* Minor documentation fix
* Unpin anndata ver in tests against anndata master
* Add branch config to work on push
* Add links to tests in GitHub Actions status badges
* Remove temporary branch name
* Remove unused variables and imports
* Simplify conditional
* Fix typo
* Do not overprune var data cache
There is a bug in how the universe and world gene sets are constructed
and passed to `ControlsHelpers.pruneVarDataCache` that causes the var
data cache to be over-pruned. This commit fixes the issue.
Consider the following example from the node console:
```
❯ node
Welcome to Node.js v13.5.0.
Type ".help" for more information.
> new Set([1], [2], [3])
Set(1) { 1 }
```
What we really want is the set `Set(3) { 1, 2, 3 }`, which can be
constructed as:
```
> new Set([].concat([1], [2], [3]))
Set(3) { 1, 2, 3 }
```
* Disable ColorBy button for truncated categories
Fixes https://github.com/chanzuckerberg/cellxgene/issues/1156
For categories that have more than 100 labels we truncate the labels in
the UI, but still allowed users to ColorBy these categories. Coloring by
these categories can cause browsers to get bogged down.
This commit disables ColorBy for truncated categories.
* Minor documentation spelling and typo fixes
* Respond to feedback from @liaprins-czi
* Respond to feedback from @colinmegill and @bkmartinjr
* Undo selection appends diffExp genes to user gene list
Fixes https://github.com/chanzuckerberg/cellxgene/issues/1171
Need: When a user performs a differential expression from within
a sub-selection (world) of the data and then resets the selection to all
cells (universe), the differential expression results are no longer
valid.
Approach:
* When the selection is reset, move the top (maxUserDefinedGenes
- len(userDefinedGenes) from the differential expression results to the
list of user defined genes
* Raise maxUserDefinedGenes to 25 to give users more room and
accommodate the extra genes transferred in from differential expression
Other commits:
* Choose different button icons
* Add diff exp genes to user defined genes on subset too
* Respond to feedback from @liaprins-czi and @bkmartinjr
* Add undo/redo tests for annotations
Fixes https://github.com/chanzuckerberg/cellxgene/issues/969
... also refactor the tests for DRY.
* Add done()
* Make e2e annotations tests safer to concurrency
* Add data-testclass for save state.
* Simplify tests and make them dependent on save state
* Add codecov to Push Test workflow
* Empty commit
* Clear reports and tag each with flags
* Tag code reports by test
* Fix codecov tags
* One more fix
* Add user-generated annotations tests to the server
Partially completes https://github.com/chanzuckerberg/cellxgene/issues/969
* Auto-format python code
* @skip_if: passing lambdas > than property strings
* Respond to feedback from @bkmartinjr
* Add smoke test for annotations features
* Do not save during annotations tests
* Fix botched rebase in dev guidelines
* Revert "Do not save during annotations tests"
This reverts commit f0bd970bb2.
* Respond to feedback from @bkmartinjr
* Collect all env vars in one, easy-to-find place
Past state:
* Default environement variables were stored in both client/package.json
and client/__tests__/e2e/config.js
* Constants that should have been linked--like the cellxgene server port
during testing--were repeated.
With this commit:
* All environment variables are parameterized
* All environment variables are packaged in default env files
* Move npm scripts to client Makefile
* Respond to feedback from @seve and @bkmartinjr
* Undo feature: fix case where previous state has no state filter
Fixes https://github.com/chanzuckerberg/cellxgene/issues/1099
When the previous state that the undo feature is trying to roll back to
has no filter state, merging javascript dictionaries result in keeping
the current state filter, preventing the actionFilter from saving the
new state.
* Fix whitespace
Currently the client build is not reproducible since, each time you run
`make build-client` the package lockfile is updated. This should be
handled separately by `make gen-package-lock` when developers actually
want to update the dependencies.
`npm ci` installs dependencies directly from the lockfile without
updating them, making builds reproducible.
* Notify users of new versions of cellxgene
Fixes https://github.com/chanzuckerberg/cellxgene/issues/683
* Do not use PyGithub client
* Protect against AttributeError
* Document that all version tags must follow SemVer
* Release tags `should -> MUST` follow semantic versioning
* Fix Makefile whitespace and .PHONY use
* Fix Makefile filename
* Modularize Makefile into client and server Makefiles
Part of the reason that the Makefile in the root directory is a bit
complicated is that it tries to handle tasks that can be handled
separately in the client and server modules.
This commit pushes some of the make logic specific to each module into
their own makefiles and calls out to those makefiles from that in the
project root.
* Add auto-formatting to client and server modules
One thing that can make linting faster is auto-formatting. This commit
adds the yapf auto-formatting tool to the server module and uses
eslint's "fix" functionality to speed up the linting/formatting process.
* Add yapf for automatic code formatting
* Add a root test target that calls sub-tests
* Apply yapf to python files
* Do not duplicate npm commands, simply pass through
* Update documentation
* Do not shadow reserved word len
* Add general test target
* Fix make call in dev-env
* Use black instead of yapf
* Run flake8 from the root directory
* Revert "Apply yapf to python files"
This reverts commit cdca128a01.
* Apply black to python code
* Resolve lint errors resulting from black format
* Add explanation of server unit tests in dev guidelines
Matt Weiden