maayanlab / appyter-catalog Goto Github PK
View Code? Open in Web Editor NEWA catalog of appyter notebooks
Home Page: https://appyters.maayanlab.cloud/
License: Other
A catalog of appyter notebooks
Home Page: https://appyters.maayanlab.cloud/
License: Other
jupyter-template==v0.2.1 introduces a few breaking changes, specifically being more careful about input sanitation, .value
or .safe_value
should no longer be called explicitly. In the case that it is, you end up with double quotes.
Add counter for card clicks
Add counter for times the appyter was executed
Add counter for the number of downloads
Add a likes button and a counter for number of likes (that one is a maybe)
As we accumulate more and more appyters, it would be great to help user to decide which ones they should use...
It is expected that the load green bars will turn solid green after the loading is complete
Appyters are ranked by retrievals, but there are some widely used Appyters
with 0 retrieval counts that should be upgraded, perhaps rank/sort Appyters by runs
Currently, many of our notebooks use tons of RAM. As it is, our server has 15G of ram and multiple notebooks running at once can easily consume all resources. To improve the resiliency of this system, we'll likely need to trigger containers as jobs and let an orchestration system like kubernetes (perhaps even locally) execute them when there are resources available / kill jobs if they use too much memory.
The simplest way to start making this a reality is to setup an execution service where we send jobs to be executed. It is up to the execution service how exactly / when it gets dispatched.
These are the tags that should be listed below the search bar:
RNA-seq
scRNA-seq
Enrichr
Machine Learning
TCGA
Harmonizome
Drugmonizome
L1000
Compare Sets
microRNAs
Kinome
Aging
PCA shows very few samples. Remove tissues with less than 30 samples (check first why you have too few).
It would be great to see the date when the appyter was published or updated.
We were basically out of memory this morning, there were several long-lived high memory usage processes running which I ended up killing because they made updating of anything impossible. We should:
We do not want users contacting authors directly for help. Users should be directed to GitHub if they need help. This way we can monitor user needs and keep proper response for support.
Ideally, addList
and enrich
(cell [10]) should contain sleep()
for a few seconds (identical to this), as enrichment for long lists and retrieving results for big libraries take long time and it instantaneous results retrieval will fail.
Also the title of the form should be centered.
Since Appyters typically produce data visualizations, it would be nice to have a representative image with each appyter on the page that have the launch button. It is ok to specify specific dimensions.
This needs to be done carefully...
The size of the logo appears bigger on the appyter execution pages. It should be made smaller and match the size of it on the catalog landing page.
We need to add data volume mounts in the docker-compose
in order to preserve data files across restarts.
The appyter that Nicole developed is TCGA specific, so it needs to have TCGA in appyter name.
The appyter that was developed by Sherry is more general for analysis of any patient cohort, so the name can be similar to the name of Nicole's appyter.
The tissue selection should be at the top of the form.
The form should have more details about what it is.
Title should be RNA-seq not RNAseq.
Fix this text in the About page
Currently, we construct a static docker-compose with a static ui page. One convenient modification was made when we added traefik
which allows us to detect new containers and mount them without needing to stop the ingest. It would be very convenient if the same could be done with the catalog -- that-is, appyters (which have their catalog metadata in a predictable place in each of the docker images) can be automatically added/removed from the catalog when the docker container gets added or removed.
Technically all we'd have to do is detect updates to the docker containers and re-construct public/appyters.json
which can be done with
docker ps -q \ # list all docker containers
| xargs -I'{}' \ # execute for each docker container
docker exec -it '{}' \ # run in container
cat /app/appyter.json \ # produce appyter metadata
2> /dev/null \ # ignore missing file in non-appyter containers
| jq -rcs \ # collect json into a compact list
> static/appyters.json # update appyters.json
fetch with {cache: 'no-cache'}
will still allow us to cache public/appyters.json
Attempting to start the catalog throws the following error:
charlesdai@Charle-PC:~/Projects/appyter-catalog$ make start
( make pull || make build ) && docker-compose up -d
make[1]: Entering directory '/home/charlesdai/Projects/appyter-catalog'
docker-compose pull
make[1]: docker-compose: Command not found
Makefile:42: recipe for target 'pull' failed
make[1]: *** [pull] Error 127
make[1]: Leaving directory '/home/charlesdai/Projects/appyter-catalog'
make[1]: Entering directory '/home/charlesdai/Projects/appyter-catalog'
cd app && npm i && npm run build && cd .. && docker-compose build app && touch app/.build
npm WARN [email protected] requires a peer of [email protected] - 3 but none is installed. You must install peer dependencies yourself.
npm WARN [email protected] requires a peer of popper.js@^1.16.0 but none is installed. You must install peer dependencies yourself.
npm WARN optional SKIPPING OPTIONAL DEPENDENCY: [email protected] (node_modules/fsevents):
npm WARN notsup SKIPPING OPTIONAL DEPENDENCY: Unsupported platform for [email protected]: wanted {"os":"darwin","arch":"any"} (current: {"os":"linux","arch":"x64"})
audited 828 packages in 4.252s
25 packages are looking for funding
run `npm fund` for details
found 0 vulnerabilities
> [email protected] build /home/charlesdai/Projects/appyter-catalog/app
> parcel build --no-cache public/index.html
๐จ /home/charlesdai/Projects/appyter-catalog/app/public/appyters.json: Unexpected token: punc (;)
at Z.get (/home/charlesdai/Projects/appyter-catalog/app/node_modules/terser/dist/bundle.min.js:1:525)
at Object.errorToJson (/home/charlesdai/Projects/appyter-catalog/app/node_modules/@parcel/utils/src/errorUtils.js:9:20)
at Pipeline.process (/home/charlesdai/Projects/appyter-catalog/app/node_modules/parcel-bundler/src/Pipeline.js:29:26)
at async Object.run (/home/charlesdai/Projects/appyter-catalog/app/node_modules/parcel-bundler/src/worker.js:15:12)
at async Bundler.loadAsset (/home/charlesdai/Projects/appyter-catalog/app/node_modules/parcel-bundler/src/Bundler.js:577:19)
at async /home/charlesdai/Projects/appyter-catalog/app/node_modules/parcel-bundler/src/Bundler.js:610:13
at async Promise.all (index 12)
at async Bundler.loadAsset (/home/charlesdai/Projects/appyter-catalog/app/node_modules/parcel-bundler/src/Bundler.js:599:21)
at async /home/charlesdai/Projects/appyter-catalog/app/node_modules/parcel-bundler/src/Bundler.js:610:13
at async Promise.all (index 3)
npm ERR! code ELIFECYCLE
npm ERR! errno 1
npm ERR! [email protected] build: `parcel build --no-cache public/index.html`
npm ERR! Exit status 1
npm ERR!
npm ERR! Failed at the [email protected] build script.
npm ERR! This is probably not a problem with npm. There is likely additional logging output above.
npm ERR! A complete log of this run can be found in:
npm ERR! /home/charlesdai/.npm/_logs/2020-06-24T20_37_35_688Z-debug.log
Makefile:20: recipe for target 'app/.build' failed
make[1]: *** [app/.build] Error 1
make[1]: Leaving directory '/home/charlesdai/Projects/appyter-catalog'
Makefile:46: recipe for target 'start' failed
make: *** [start] Error 2
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.