Feature/fix aws invoke

.github/workflows/release.yml

@@ -0,0 +1,30 @@
+name: Release Packages
+
+on:
+  release:
+    types: [released]
+
+jobs:
+  publish-to-npm-registry:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v3
+
+
+      - uses: actions/setup-node@v4
+        with:
+          node-version: '20.x'
+
+      - run: npm install
+
+      # build and package
+
+      - run: npm run build
+
+      - run: npm run package
+
+      - run: npx publib-npm
+        env:
+          NPM_TOKEN: ${{secrets.NPM_PUBLISH_TOKEN}}
+          NPM_ACCESS_LEVEL: public

.gitignore

@@ -17,3 +17,12 @@ TODO.*md
 NOTES.md
 prev/
 temp/
+/.idea/
+
+# Generated by jsii - safe to delete, and ideally should be in .gitignore
+tsconfig.json
+
+dist/
+
+.jsii
+*.tsbuildinfo

.npmignore

@@ -1,6 +1,20 @@
-*.ts
-!*.d.ts
-
 # CDK asset staging directory
 .cdk.staging
 cdk.out
+
+# Exclude typescript source and config
+*.ts
+!*.d.ts
+tsconfig.json
+*.tsbuildinfo
+
+# Include javascript files and typescript declarations
+!*.js
+!*.d.ts
+
+# Exclude jsii outdir
+dist
+
+# Include .jsii and .jsii.gz
+!.jsii
+!.jsii.gz

README.md

@@ -1,122 +1,3 @@
-# nextflow-stack
+# Oncoanalyser AWS
 
-An AWS stack for running Nextflow pipelines on Batch using shared resources.
-
-🚧
-
-Highlight key aspects
-
-* Precise resource allocation requests under SPOT pricing model
-* No EBS costs and local SSD discounted via SPOT pricing
-* No duplication of Batch queues
-* Fusion, Wave
-* Improved resume experience compared to use of ephermal workdir disk
-* Ability to highly optimise instances for individual processes
-
-Future work
-
-* Dynamic queue selection
-* Retry on SPOT pre-emption X times
-* BYO bucket
-* CodePipeline for deployment
-* Improve configuration and data handling
-
-## Table of contents
-
-🚧
-
-## Deployment
-
-🚧
-
-* Detail application stacks that must be deployed aplus additional set up (i.e. ECR, Docker images)
-
-### Development
-
-> `deployment/development-stack.ts`
-
-### CodePipeline CI/CD
-
-> `deployment/codepipeline-stack.ts`
-
-## Pipelines
-
-🚧
-
-### oncoanalyser
-
-🚧
-
-### star-align-nf
-
-🚧
-
-> `umccr/star-align-nf`
-
-### UMCCR post-processing
-
-🚧
-
-#### Design
-
-🚧
-
-Diagram (avoid overlap with Overview diagram) including reference data etc
-
-Section detailing current compromises
-
-* GDS token access
-* Migrating data from GDS to S3 for execution
-* Not fully optimised for speed; show timeline or similar
-* Many Docker images on DockerHub, ideally would be on ECR
-  * Must resolve 502 errors in Wave when pulling from grch.io or ECR
-* Only broad control over processes run currently
-* Passing run configuration by CLI args is somewhat clumsy
-  * Alternative: JSON on remote (S3, API call); could be extended to general config
-
-Items that need to be addressed
-
-* **Important**: Isofox takes an expected count file that is dependent of read length
-  * So we **must** be sure that we're using expected counts for the correct read length
-* Application still tied to UMCCR VPC and other resources
-* Repetition between pipeline stacks and associated code (run.sh, Dockerfile, etc)
-* Cannot parallelise workflow stack deployment in CodePipeline with waves in current set up
-* Extra arguments for `run.sh` are ignored, error should be raised
-* Job cancellation is difficult when the pipeline crashes
-* I have observed a rare issue with unexpected Fusion shutdown that interrupts processing
-* Staging data from GDS to S3 suffers significant slow down after ~one hour (bursting related?)
-  * Could spin out multiple instances or high capacity instance improve transfer speed
-
-Other notes
-
-* Nextflow config `nextflow_aws.config` could be split into processes and input
-* Include other parameters in config: HMF refdata path, VBE path, genome version, workdir
-* To discuss staged data location, sorting, structure, retention
-* Lifecycle of data in Nextflow S3 workdir; single workdir or per run/sample/etc
-
-#### Usage
-
-🚧
-
-Diagram describing common run modes with correponding commands
-
-Run modes (relative to CUPPA)
-
-* WGTS
-* WGS only
-* WTS only
-* WGTS with existing WGS
-* WGTS with existing WTS
-* WGTS with existing WGS and WTS
-
-> include run resuming and use of this as an alternative to providing existing data
-
-> note how to run any individual process/stage with the appropriate inputs
-
-#### Notes
-
-🚧
-
-Other important items to note
-
-* Fusion usually gives much better performance but not always
+A CDK package that deploys Oncoanalyser.