Skip to content

Latest commit

 

History

History
811 lines (633 loc) · 42.9 KB

README.md

File metadata and controls

811 lines (633 loc) · 42.9 KB

AGR Curation API

This repo holds the code for the API and the UI of the AGR curation system.

Getting Started

These instructions will get you a copy of the project and the API up and running locally. This includes Postgres, OpenSearch and Cerebro.

Contents

Developing

Before you start coding for a new feature or making a bugfix, it is important to know the intended goal of your code:

  • Developing fixes, new features and major changes to be released in some future release (standard) => Must be developed and tested on alpha
  • Smaller (bug)fixes to a version currently being previewed on beta, in preparation of a release to production => Must be developed and tested on beta
  • Hotfixes, urgent bugfixes to the current production version which cannot wait on the next full release (bundled with other updates currently under development) to be deployed => Must be developed off the latest production release.

For more details about the intended use of and deployment to each of the environments, see deployment environments.

Branching

The three permanent branches in this repository represent the code to be deployed to their respective environments: alpha, beta and production.

  • To create code that must solely go on the alpha environment (most day-to-day development):

    • Create a feature branch to work on that starts off alpha and have the name contain the ticket number and a short description of it.

      > git checkout alpha
      > git pull
      > git checkout -b feature/${JIRA-TICKET-NR}_short-description
    • Do your coding and testing, and push the branch to github

      #Coding and testing here
      > git push origin feature/${JIRA-TICKET-NR}_short-description
    • Once coding and testing completed, submit a pull request in github to merge back to alpha.
      If this PR does not require deployment (because it only concerns files like the README, or to batch multiple fast-consecutive PRs into a single deployment), apply the no-deploy GitHub label to the PR (on creation) to skip automatic deployment. Otherwise, deployment to alpha will automatically trigger once the PR is approved and merged.

  • To make fixes to the version currently deployed on the beta environment:

    • Create a new release/vx.y.z-rca off beta (replacing x, y, z and a as appropriate), with as version the current beta release candidate version number increased by one (a+1).

      > git checkout beta
      > git pull
      > git checkout -b release/vx.y.z-rca
      > git push -u origin release/vx.y.z-rca

      Alternatively, you can use a betafix/* branch instead of a release/vx.y.z-rca for changes that need merging in the beta code branch but no release creation or deployment (such as certain github actions changes to fix broken workflows). Use these only when creating a new release is not an option.

    • If one or a few specific things need fixing then work directly on this release branch, otherwise create feature branches starting from this release branch as approriate (including the ticket number in the branch name) and create PRs to merge those back into the release branch.

    • Once coding and testing completed, ensure you've pushed the latest changes and submit a pull request in github to merge the release/vx.y.z-rca branch back to beta. Before merging, do the pre-deployment steps when appropriate. Then merge, submitting a short description of the changes as commit message body, to trigger release creation and deployment.

    • Once deployment completed successfully, do the require post-deployment steps.

  • To make fixes to the version currently deployed on the production environment:

    • Create a new patch release branch release/vx.y.z off production (replacing x, y, z and a as appropriate), with as version the current production patch version number increased by one (z+1).

      > git checkout production
      > git pull
      > git checkout -b release/vx.y.z
      > git push -u origin release/vx.y.z

      Alternatively, you can use a prodfix branch instead of a release/vx.y.z for changes that need merging in the beta code branch but no release creation or deployment (such as certain github actions changes to fix broken workflows). Use these only when creating a new release is not an option.

    • If one or a few specific things need fixing then work directly on this release branch, otherwise create feature branches starting from this release branch as approriate (including the ticket number in the branch name) and create PRs to merge those back into the release branch.

    • Once coding and testing completed, ensure you've pushed the latest changes and submit a pull request in github to merge the release/vx.y.z branch back to production. Make sure all approriate testing has occured, as merging this PR will deploy directly to production.

      Before merging, do the pre-deployment steps when appropriate. Then merge, submitting a short description of the changes as commit message body, to trigger release creation and deployment.

Data structure changes

When making code changes which change any of the data structures, the supporting database schema needs to reflect these changes to successfully be able to deploy and run this new code. Hibernate takes care of creating new tables and columns, but changing existing columns, removing columns and moving or transforming data is not handled by hibernate, as this usually requires human judgment to correctly assess the required steps to maintain data integrity. To automate and correctly replicate the deployment of these sort of data migrations accross environments, we use flyway.

Flyway works by manually defining and storing SQL files to handle the data migrations in src/main/resources/db/migration/, with a filename formatted as v${VERSION}__${DESCRIPTION}.sql.

  • ${DESCRIPTION} can be a small free-form description of the changes made and is optional (along with the __ separator)
  • ${VERSION} must be defined. The agreed format for version is x.y.z.a, where x.y.z represents the curation application version on which this migration file will be applied (excluding any release-candidate portion of it), and the .a part represents an incremental number for all migrations applied to this release. As an example, the first migration to be applied on top of v0.4.0 (so while developing what could become v0.5.0) could be named v0.4.0.1__updated-da-foreign-keys.sql, the second one v0.4.0.2__renamed-reference-field.sql etc.

Installing

Currently there are several scripts in the docker directory that can be used to start the various servers, locally if there is not a dev server available.

Maven

Maven version 3.8.1 should be installed in order to compile all the code and libraries.

> brew install maven # Mac
> apt-get install maven # Linux ubuntu
> # Windows: Use WSL to run Ubuntu and run entire setup there

Docker setup

Docker desktop can be downloaded via the following link: Docker Download

Once docker is setup and installed you will need to issue the following docker command to get a docker network created:

> docker network create curation
549127add...

Postgres

Run Postgres Script which will launch an empty postgres instance locally from the official postgres docker image.

If you are needing to connect to the production version of the database, use your personal postgres credentials that have been provided to you, or contact a system admin to recover them when you lost them.

When you need admin access into the production database, the following command can be used to get the credentials (but will only work once you have been setup with an AWS user and configured your local aws environment).

docker run --rm -it -v ~/.aws:/root/.aws amazon/aws-cli secretsmanager get-secret-value --secret-id curation-db-admin --region us-east-1 --query SecretString --output text

The output will look something like the following:

	{
		"username":"********",
		"password":"********",
		"host":"********",
		"engine":"postgres",
		"port":5432,
		"dbname":"curation",
		"dbInstanceIdentifier":"curation-db"
	}

OpenSearch (Elastic Search alternative)

Run OpenSearch script which will launch OpenSearch from the official OpenSearch docker image.

If you point a browser over to http://localhost:9200 you should see something like the following:

	{
		"name": "8e063394038e",
		"cluster_name": "docker-cluster",
		"cluster_uuid": "M50gdkzjRFeDbu0XBZq85g",
		"version": {
			"number": "7.9.0",
			"build_flavor": "default",
			"build_type": "docker",
			"build_hash": "a479a2a7fce0389512d6a9361301708b92dff667",
			"build_date": "2020-08-11T21:36:48.204330Z",
			"build_snapshot": false,
			"lucene_version": "8.6.0",
			"minimum_wire_compatibility_version": "6.8.0",
			"minimum_index_compatibility_version": "6.0.0-beta1"
		},
		"tagline": "You Know, for Search"
	}
	

Cerebro

Run Cerebro Script which will launch cerebro, an ES control plane, from the appropriate docker image.

Connect to this by browsing to http://localhost:9000, this is used to connect to the ES server, which can be accessed by entering http://opensearch:9200 (the internal docker address) in the Node address field, while opensearch is directly accessible at http://localhost:9200 on the local machine.

Up and Running Servers

After running all the previous commands issue the docker command: docker ps -a

This will show all the services that are up and running.

CONTAINER ID   IMAGE                                   COMMAND                  CREATED          STATUS          PORTS                                                                                                                                                                          NAMES
4ee5641a9481   yannart/cerebro                         "./bin/cerebro"          26 minutes ago   Up 26 minutes   0.0.0.0:9000->9000/tcp, :::9000->9000/tcp                                                                         cerebro
7811ffe3c92d   opensearchproject/opensearch:1.2.4   "./opensearch-docker…"   24 minutes ago   Up 24 minutes   0.0.0.0:9200->9200/tcp, :::9200->9200/tcp, 9300/tcp, 0.0.0.0:9600->9600/tcp, :::9600->9600/tcp, 9650/tcp          opensearch
7de3cf028e9c   postgres:13                             "docker-entrypoint.s…"   3 days ago       Up 3 days       0.0.0.0:5432->5432/tcp, :::5432->5432/tcp                                                                         postgres

Code Style

This repository uses checkstyle to validate formatting of code to a coding standard. Checkstyle is a development tool to help programmers write Java code that adheres to a coding standard. It automates the process of checking Java code to spare humans of this boring (but important) task. This makes it ideal for projects that want to enforce a coding standard.

This repository also uses Prettier to validate and update the formatting of code in the frontend codebase.

Running Checkstyle

To run checkstyle locally in order to check your code before committing run the following:

> mvn checkstyle:check

Or the make target:

> make checkstyle

This will check all the style of the code and report any issues that it finds with the code. Here is an example output:

> make checkstyle
[INFO] Scanning for projects...
[INFO] Inspecting build with total of 1 modules...
[INFO] Installing Nexus Staging features:
[INFO]   ... total of 1 executions of maven-deploy-plugin replaced with nexus-staging-maven-plugin
[INFO] 
[INFO] ----------------< org.alliancegenome:agr_curation_api >-----------------
[INFO] Building AGR Curation :: API 0.0.0
[INFO]   from pom.xml
[INFO] --------------------------------[ jar ]---------------------------------
[INFO] 
[INFO] --- checkstyle:3.3.1:check (default-cli) @ agr_curation_api ---
[INFO] There is 1 error reported by Checkstyle 10.17.0 with checkstyle.xml ruleset.
[ERROR] src/main/java/org/alliancegenome/curation_api/main/Main.java:[9] (regexp) RegexpSinglelineJava: Indent must use tab characters
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time:  6.107 s
[INFO] Finished at: 2024-05-29T05:23:33-04:00
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal org.apache.maven.plugins:maven-checkstyle-plugin:3.3.1:check (default-cli) on project agr_curation_api: You have 1 Checkstyle violation. -> [Help 1]
[ERROR] 
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR] 
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/MojoFailureException
make: *** [checkstyle] Error 1

Checkstyle can be configured via the checkstyle.xml file in the root of the project, this file can also be used by IDE's to keep code in line with the ruleset.

Running Prettier

To validate formatting on JavaScript files locally with Prettier:

> npx prettier --check **/*.js

Or the make target:

> make prettier-check

To update formatting on JavaScript files locally with Prettier:

> npx prettier --write **/*.js

Or the make target:

> make prettier-write

Prettier can be configured via the .prettierrc.yaml file in the src/main/cliapp director of this project, this file can also be used by IDE's to keep code in line with the ruleset.

Checkstyle Links

Prettier Links

Building

Before building locally, make sure and create a copy of the application.properties file

> cp src/main/resources/application.properties.defaults src/main/resources/application.properties

this way custom configuration changes can be used without having to commit them into the repo.

Both the UI or API can be run locally without needing a separate build step. For instructions on how to do so, see Running.

Building the application uberjar is done as part of the Docker image creation, instructions on how to use this process for local image building can be found below.

Building Docker Image

When needed, the following command can be issued to build a complete runnable application as a docker image:

> make docker

Running

Depending on which part of the application you would like to run find the following commands:

Running API

The API does not have to be build in order to run it in dev mode

> make apirun
mvn compile quarkus:dev
...
Listening for transport dt_socket at address: 5005
__  ____  __  _____   ___  __ ____  ______ 
 --/ __ \/ / / / _ | / _ \/ //_/ / / / __/ 
 -/ /_/ / /_/ / __ |/ , _/ ,< / /_/ /\ \   
--\___\_\____/_/ |_/_/|_/_/|_|\____/___/   
...

This will run just the API and you can find it running at: http://localhost:8080, however if the API was built without the UI there will be nothing to see at that url. There is a swagger interface running at: http://localhost:8080/swagger-ui.
Windows note: If you're running the curation software through WSL (on Windows), Java might serve this interface through IPv6, in which case localhost:8080 will be accessible within WSL, but not from your windows browser. Instead, you can then access the API and the swagger interface in your browser at it's IPv6 counterpart: http://[::1]:8080.

Following the prompts at the bottom of the screen after the server is up and running.

--
Tests paused
Press [r] to resume testing, [o] Toggle test output, [h] for more options>

Some notable commands are s for restarting the server to pick up java changes. j change the debugging level of the server, q or ctrl-c in order to stop the server.

When running into issues when running/building the API, try a clean build by running make clean first.

Running UI

To build and run the UI, issue the following command:

> make uirun
...

This will start the local UI and it will be running at: http://localhost:3000 this is in develop mode so any changes made to code will automatically restart the UI. The UI also proxies all /api requests over to the API.

If need be the API_URL can be changed by setting it on the command line before running the UI:

> export API_URL=https://alpha-curation.alliancegenome.org  # to send proxied requests to the alpha server.

Additionally, there are two convenience commands that will proxy /api requests to either the alpha or beta environment

> make uirunalpha
> make uirunbeta

Running the docker image

To run the complete application as the locally built docker image, execute the following command:

> make docker-run

Releasing and deploying

Deployment environments

There are three environments to which code automatically gets deployed at different stages during development:

  • The alpha environment should be regarded as the developers environment, where active code development can be happening at any moment, and things are expected to break every now and then. This environment receives new deployments on every push made to the alpha branch.
  • The beta environment should be regarded as the testers environment, where (external) testers can have a first look at newly developed functionality before it is ready to be pushed to production. It is more stable than alpha, but is still subject to regular change as feedback is collected and the final kinks are being ironed out. This environment receives new deployments for every pre-release created on Github.
  • The production environment is the final stage of deployment, where users can rely on full-tested and ready-to-use features and functionality at any moment. This is the most stable and reliable environment. This environment receives new deployments for every full (stable) release created on Github.

All deployments are fully automated through Github actions, for which the configuration files can be found in the .github/workflows/ directory. Deployments to the alpha environment happen automatically as code gets pushed to the alpha branch (after merging a PR), but for deployments to beta and production, a small number of steps needs to be taken in order to create a release and trigger deployment, which are described below.

Promoting code versions

The general flow from coding to a production release goes something like this:

  1. Coding and testing for most tickets happens on alpha (throughout the sprint)
  2. When a stable state of alpha gets reached (usually at sprint review), this stable state gets promoted to beta as a release candidate, available for external user testing and testing against the latest production data and for testsubmissions of new production-ready data.
  3. On approval from the external users, and in agreement with the product manager and the team, this release candidate gets promoted to a new stable release on production.

As the code goes through the different stages, it becomes more and more stable as it gets closer to production.

Promoting code from alpha to beta

  1. Look at the alpha-environment dashboard and ensure that for each Entity and Ontology, the number in the Database * count column matches the number in the Search index * count column (ignore the values for Disease Annotations and Literature References). If there is a mismatch for any row, investigate what caused this and fix the issue first before continuing the deployment.
    It may be that the scheduled mass reindexing failed (run every Sunday 00:00 UTC), in which case a successful mass-reindexing must first be performed on the alpha environment, before continuing the deployment to the beta environment.

  2. Decide on the exact commit to promote. As active development constantly happens on the alpha branch, a stable point must be chosen to promote in order to keep the amount of bugs present in the (pre)release to be created to a minimum. Usually this means choosing the commit which was on alpha at the time of presenting at the past sprint review meeting.

    The commit history can be visualised by browsing the github repository's network graph page.

  3. Decide on a proper release version number to be used for the new prerelease that will be created as a result of this promotion (see Release versioning).
    Generally speaking, for beta (pre)releases that means either

    • Incrementing the release candidate version on the latest release candidate when it was decided the previous release candidate will not be promoted to a full release and the latest changes need to be added in order to be able to create a full release. or
    • Incrementing the MAJOR or MINOR release numbers as appropriate when the previous release candidate was (or will be) promoted to a full release and reset to rc1 for the next release.
    • When choosing an alpha release make sure to include the Pull Requests that may have fixed bugs shortly after the sprint review, if they needed to be included in beta release.
  4. Create a release/vx.y.z-rca branch from the commit on the alpha branch that was chosen earlier to be promoted.

    git pull
    git checkout -b release/vx.y.z-rca <commit-sha>
    git push origin release/vx.y.z-rca
  5. Create a pull request to merge this release branch into the beta branch. Do not merge yet and do not enable auto-merge.

  6. Await PR approval, then do the necessary pre-deployment steps to prepare the beta environment for a successful deployement.

  7. In the above PR, click the Merge pull request button and

    • Leave the autogenerated Merge pull request #... commit title
    • Provide a short description of the changes introduced in this release in the commit body textarea. When deploying code as demoed during sprint review, use As sprint review MMM DDth YYYY as the commit message body (e.g. As sprint review Jan 19th 2022). Then hit the confirm merge button. This will trigger a chain of github actions workflows which will a. Create a git tag (using the commit message body provided as tag message) b. Create a github release (prerelease when merged in beta). c. Build and deploy the application as a container to AWS
  8. In addition to the above decribed actions, merging a release-branch PR in beta will also automatically create a mergeback PR to alpha to make the (pre)release tag reachable from alpha (and report the correct version number in the alpha UI and version API endpoint). Assign one or a few reviewers to this PR, and merge it on approval.

Promoting code from beta to production

  1. Look at the beta-environment dashboard and ensure that for each Entity and Ontology, the number in the Database * count column matches the number in the Search index * count column (ignore the values for Disease Annotations and Literature References). If there is a mismatch for any row, investigate what caused this and fix the issue first before continuing the deployment.
    It may be that the scheduled mass reindexing failed (run every Sunday 00:00 UTC), in which case a successful mass-reindexing must first be performed on the beta environment, before continuing the deployment of new features to the beta environment.
  2. Decide on a proper release version number to be used for the new release that will be created as a result of this promotion (see Release versioning).
    Generally speaking, for production (full) releases that means removing the release-candidate extension from the release number used by the latest release candidate, which will be promoted to a full release here.
  3. Create a release/vx.y.z branch from beta, or use a specific commit from history to promote if a new release canidate for the next release was already added to beta.
    git pull
    git checkout -b release/vx.y.z beta
  4. Update the RELEASE-NOTES.md file to contain a section describing all (noteworthy) changes made since the last full release, and include JIRA ticket and/or PR references where possible. Commit after update.
  5. Push the release branch to github
    git push origin release/vx.y.z
  6. Create a pull request to merge this release branch into the production branch. Do not merge yet and do not enable auto-merge.
  7. Await PR approval, then do the necessary pre-deployment steps to prepare the production environment for a successful deployement.
  8. In the above PR, click the Merge pull request button and
    • Leave the autogenerated Merge pull request #... commit title
    • Provide a short description of the changes introduced in this release in the commit body textarea. When deploying code as demoed during sprint review, use As sprint review MMM DDth YYYY as the commit message body (e.g. As sprint review Jan 19th 2022). Then hit the confirm merge button. This will trigger a chain of github actions workflows which will a. Create a git tag (using the commit message body provided as tag message) b. Create a github release (full latest release when merged in production). c. Build and deploy the application as a container to AWS
  9. In addition to the above decribed actions, merging a release-branch PR in production will also automatically create a mergeback PR to alpha to make the release tag reachable from alpha (and report the correct version number in the alpha UI and version API endpoint). Assign one or a few reviewers to this PR, and merge it on approval.

Additional deployment steps

In order to successfully deploy to the beta or production environment, as few additional steps need to be taken before and after merging into the respective branch and triggering deployment to ensure the new version of the application can function in a consistent state upon and after deployment.

Pre-deployment steps

  1. Compare the environment variables set in the Elastic Beanstalk environment between the environment you want to deploy to and from (e.g. compare curation-beta to curation-alpha for deployment to beta, or curation-production to curation-beta for deployment to production). This can be done through the EB console, or by using the eb printenv CLI. Scan for relevant change:
    • New variables should be added to the environment to be deployed to, before initiating the deployment
    • ENV-specific value changes should be ignored (for example, datasource host will be different for each)
    • Other variable value changes should be propagated as appropriate, before initiating the deployment
    • Removed variables should be cleaned up after successfull deployment
  2. When wanting to deploy a prerelease to the beta environment, reset the beta postgres DB and roll down the latest production DB backup.
    This must be done to catch any potentially problems that could be caused by new data available only on the production environment, before the code causing it would get deployed to the production environment.
    The restore function automatically prevents users from writing to the database while it is being reloaded, by temporarily making the target database read-only and restoring in a separated database before renaming.
    1. For instructions on how to trigger the restore, see the agr_db_backups README
    2. After the restore completed, restart the beta environment app-server to re-apply all flyway migrations that were not yet present in the restored (production) database.
      > aws elasticbeanstalk restart-app-server --environment-name curation-beta
    3. Check the logs for errors after app-server restart, which could indicate a DB restore failure and troubleshoot accordingly if necessary to fix any errors.
  3. Connect to the Environment's search domain and delete all indexes. A link to the Cerebro view into each environment's search indexes is available in the curation interface under Developer Links > Search index UI (cerebro) (VPN connection required). Alternatively, you can reach this UI manually by browsing to the AGR Cerebro interface and entering the environment's domain endpoint manually. The domain endpoint URL can be found through the Amazon OpenSearch console.

Post-deployment steps

  1. Check the logs for the environment that you're releasing to and ensure the application started successfully and that all flyway migrations completed successfully. If errors occured, troubleshoot and fix accordingly.
    • If the application failed to start due to incompatible indexes (the indexes were not deleted before deployment), delete the indexes and restart the app-server by running
      > aws elasticbeanstalk restart-app-server --environment-name curation-<beta|production>
    • If any of the migrations failed, try restarting the application once more, or create a new release to fix the failing migration if the second restart did not resolve the issue:
      1. Create a new release/v* branch from the beta or production branch as appropriate.
        • Update to the next release candidate (a in vx.y.z-rca) for fixes to beta, using the beta branch as starting commit
        • Update to the next patch release (z in vx.y.z) for fixes to production, using the production branch as starting commit
      2. Fix the required migration files
      3. Create a PR to merge the new release/v* branch back into beta/production
      4. repeat the post-deployment steps for this new release (candidate).
  2. Reindex all data types by calling the system/reindexeverything endpoint with default arguments (found in the System Endpoints section in the swagger UI) and follow-up through the log server to check for progress and errors.
  3. Once reindexing completed, look at the dashboard page (where you deployed to) and ensure that for each Entity and Ontology, the number in the Database * count column matches the number in the Search index * count column (ignore the values for Disease Annotations and Literature References). If there is a mismatch for any row, investigate what caused this and fix the issue first before continuing the deployment.
  4. If code to support new ontologies was added, create the respective new data loads through the Data loads page and load the new file.
  5. After completing all above steps successfully, return to the code promoting section to complete the last step(s) (alpha to beta or beta to production)

Release versioning

For our release versioning, we apply Semantic Versioning whereby

  • MAJOR.MINOR.PATCH is used for full releases
  • Release-candidate extensions are used for prereleases in the format MAJOR.MINOR.PATCH-rcx where x is an increment starting at one.
  • PATCH version upgrades should only be used for hotfixing bugs on the production environment (e.g. a bugfix applied to v0.1.0 becomes v0.1.1).

Note: For the time being, the MAJOR release number is kept at 0 to indicate the early development phase, until the product is ready for active usage by users external to the development team, in a stable production environment.

Manual Release Creation

Release creation is generally automated through github actions upon merging appropriately named release/* branches into either the beta branch (for prereleases) or the production branch (for full releases).

To manually create a new (pre-)release and deploy to beta or production, do the following steps:

  1. Ensure you're on the branch you intend to create a new release for and pull the latest code.

    git checkout beta
    git pull
  2. Determine the approriate release number to be created (see Release versioning).

  3. Ensure you have a clean working directory before continuing, you can save any local changes for later using git stash if needed.

  4. Tag the release and push the tag to github. Prefix the release nr with v as tagname, and provide a short description for the release. Beta releases should contain a reference to the date they were demoed at sprint review when applicable.

    # Tag name should be the release nr to be released, prefixed with v (eg. v0.2.0-rc1)
    # Tag message for beta releases can be something like 'As sprint review Jan 19th 2022'
    git tag -a vx.y.z-rca -m 'As sprint review MMM DDth YYYY'
    git push origin vx.y.z-rca
  5. Check the pre-deployment steps and perform them as appropriate.

  6. Go to the AGR curation release page on github, create a new release by clicking the "Draft a new release" button at the top.

    • In the "Choose a tag" selection box, select the git tag you created in the previous step
    • Give the release a proper title including the sofware name and the release number ("AGR Curation release tag-name")
    • Let GitHub autogenerate a summary of all changes made in this release by clicking the "Auto-generate rease notes" button.
    • Ensure the "Set as a pre-release" checkbox at the bottom is checked appropriately.
      • Checking this box creates a prerelease, which only get deployed to the beta environment.
      • Leaving the box unchecked (the default) creates a full release which gets deployed to the production environment.
  7. Confirm all entered details are correct and publish the release.

Once published, github actions kicks in and the release will get deployed to the appropriate environments. Completion of these deployments is reported in the #a-team-code slack channel. After receiving a successful deployment notification, perform the appropriate post-deployment steps.

Database

The Curation application connects to the postgres DB using a user called curation_app, which is member of a role called curation_admins. This role and user are used to ensure the database can be closed for all but admin users on initiating a DB restore, such that no accidential writes can happen to the postgres database during the restore process to ensure data integrity. When restoring to or creating a new postgres server, ensure the curation_admins role exists and has a member user called curation_app, to ensure the database can be restored with the correct ownerships, and give the curation_admins role all permissions to the curation database, the public schema, and all tables and sequences in it, and change the default privileges to allow the curation_admins role all permissions on all newly created tables and sequences in the public schema.

This can be achieved by connecting to the curation database using the admin (postgres) user (using psql) and executing the following queries:

-- Create the role
CREATE ROLE curation_admins;

-- Create user (change the password)
CREATE USER curation_app WITH PASSWORD '...';
-- Grant role to user
GRANT curation_admins TO curation_app;

-- Grant required privileges to group (role)
GRANT ALL ON DATABASE curation TO GROUP curation_admins;
GRANT ALL ON SCHEMA public TO GROUP curation_admins;
GRANT ALL ON ALL TABLES IN SCHEMA public TO GROUP curation_admins;
GRANT ALL ON ALL SEQUENCES IN SCHEMA public TO GROUP curation_admins;
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON TABLES TO GROUP curation_admins;
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON SEQUENCES TO GROUP curation_admins;

Loading Data

Assuming the API is running on localhost, check the Swagger interface for the correct endpoint to post data to. Here is an example of loading the Gene BGI file from the Alliance.

> curl -vX POST http://localhost:8080/api/gene/bulk/bgifile -d @1.0.1.4_BGI_ZFIN_4.json --header "Content-Type: application/json"

Here is an example of loading Do ontology:

> curl -vX POST http://localhost:8080/api/doterm/bulk/owl -d @doid.owl --header "Content-Type: application/xml"

Example of loading Allele's

> curl -vX POST http://localhost:8080/api/allele/bulk/allelefile -d @1.0.1.4_ALLELE_MGI_0.json --header "Content-Type: application/json"

Submitting Data

Here is an example using curl:

curl \
   -H "Authorization: Bearer 2C07D715..." \
   -X POST "https://${Curation_System}.alliancegenome.org/api/data/submit" \
   -F "LoadType_SubType=@/full/path/to/file1.json" \
   -F "LoadType_SubType=@/full/path/to/file2.json"

Valid values for LoadType, and SubType can be found in the examples below.

Curation System

This value can be one of the following list.

Name Description
alpha-curation Alpha Curation Site
beta-curation Beta Curation Site
curation Production Curation Site

DQMs can find information regarding the LinkML version and data classes currently supported by the curation system at the following link: https://${Curation_System}.alliancegenome.org/api/version

API Access Token

This will be a key that is generated via logging into the curation website for the DQM's to use for uploading files.

Load Type

Load type corresponds with the type of file that load needs.

Name Description
GENE LinkML Gene
ALLELE LinkML Allele
AGM LinkML AGM
DISEASE_ANNOTATION LinkML Disease Annotations

Sub Type

This is a grouping mechanism to group files together

Name Description
FB Fly Base
HUMAN Human Supplied by RGD
MGI Mouse Genome Database
RGD Rat Genome Database
SGD Saccharomyces Genome Database
WB Worm Base
ZFIN Zebrafish Information Network

Including corresponding LinkML version in the JSON file submission header

The LinkML version for which the file is being submitted now needs to be added to the JSON file header, for example:

{
  "linkml_version" : "v1.3.2",
  "disease_agm_ingest_set" : [ {

Querying Data

Search

Documentation to use the API search endpoints.

Find

Documentation to use the API find endpoints.

EB Deployment (beta)

This section is WIP.

CLI deployment to Elastic Beanstalk requires the EB CLI to be installed on your local machine. Follow the instructions provided by AWS to do so. The EB CLI uses the same config files and environment variables as the standard AWS CLI to locate the appropriate authentication.

A EB cli configuration file is stored in the .elasticbeanstalk repo subdirectory. If you would encounter (local) problems running eb commands, remove this directory and reinitialise it by executing make eb-init.

All EB environment configurations are stored in the .ebextensions repo subdirectory.

Before attempting a sandbox environment creation, inspect the eb-create Makefile target and export all variables defined in it with the appropriate values in your local shell environment to allow storage of all required configurations in the EB environment configurations.

Then you can run make eb-create to create a new environment and make eb-terminate to terminate it.

Maintainers

Current maintainers: