update to viash 0.9

.github/workflows/build.yml

@@ -1,120 +1,21 @@
-name: build
+name: Build
 
 on:
   push:
     branches: [ 'main' ]
   workflow_dispatch:
     inputs:
-      target_branch:
-        description: 'Branch to deploy to. If not specified, `build-${BRANCH_NAME}` will be used.'
-        required: false
       version:
-        description: 'Version name to use for the build. If not specified, `build-${BRANCH_NAME}` will be used.'
+        description: |
+          The version of the project to build. Example: `1.0.3`.
+          
+          If not provided, a development build with a version name
+          based on the branch name will be built. Otherwise, a release
+          build with the provided version will be built.
         required: false
 
 jobs:
-  # phase 1
-  list:
-    runs-on: ubuntu-latest
-
-    outputs:
-      target_branch: ${{ steps.defaults.outputs.target_branch }}
-      version: ${{ steps.defaults.outputs.version }}
-      component_matrix: ${{ steps.set_matrix.outputs.matrix }}
-
-    steps:
-    - uses: actions/checkout@v4
-
-    - uses: viash-io/viash-actions/setup@v5
-
-    - name: Determine version tag from branch name
-      id: defaults
-      run: |
-        BRANCH_NAME=$(echo $GITHUB_REF | sed 's/refs\/heads\///')
-
-        VERSION=${{ github.event.inputs.version }}
-        if [ -z "$VERSION" ]; then
-          VERSION="build-$BRANCH_NAME"
-        fi
-        echo "version=$VERSION" >> $GITHUB_OUTPUT
-
-        TARGET_BRANCH=${{ github.event.inputs.target_branch }}
-        if [ -z "$TARGET_BRANCH" ]; then
-          TARGET_BRANCH="build-$BRANCH_NAME"
-        fi
-        echo "target_branch=$TARGET_BRANCH" >> $GITHUB_OUTPUT
-
-    - name: Remove target folder from .gitignore
-      run: |
-        # allow publishing the target folder
-        sed -i '/^target.*/d' .gitignore
-
-    - uses: viash-io/viash-actions/ns-build@v5
-      with:
-        config_mod: .functionality.version := '${{ steps.defaults.outputs.version }}'
-        parallel: true
-
-    - name: Deploy to target branch
-      uses: peaceiris/actions-gh-pages@v4
-      with:
-        github_token: ${{ secrets.GITHUB_TOKEN }}
-        publish_dir: .
-        publish_branch: ${{ steps.defaults.outputs.target_branch }}
-
-    - id: ns_list
-      uses: viash-io/viash-actions/ns-list@v5
-      with:
-        platform: docker
-        src: src
-        format: json
-
-    - id: set_matrix
-      run: |
-        echo "matrix=$(jq -c '[ .[] | 
-          { 
-            "name": (.functionality.namespace + "/" + .functionality.name),
-            "dir": .info.config | capture("^(?<dir>.*\/)").dir
-          }
-        ]' ${{ steps.ns_list.outputs.output_file }} )" >> $GITHUB_OUTPUT
-
-  # phase 2
   build:
-    needs: list
-
-    runs-on: ubuntu-latest
-
-    strategy:
-      fail-fast: false
-      matrix:
-        component: ${{ fromJson(needs.list.outputs.component_matrix) }}
-
-    steps:
-    # Remove unnecessary files to free up space. Otherwise, we get 'no space left on device.'
-    - uses: data-intuitive/reclaim-the-bytes@v2
-
-    - uses: actions/checkout@v4
-
-    - uses: viash-io/viash-actions/setup@v5
-
-    - name: Build container
-      uses: viash-io/viash-actions/ns-build@v5
-      with:
-        config_mod: .functionality.version := '${{ needs.list.outputs.version }}'
-        platform: docker
-        src: ${{ matrix.component.dir }}
-        setup: build
-
-    - name: Login to container registry
-      uses: docker/login-action@v3
-      with:
-        registry: ghcr.io
-        username: ${{ secrets.GTHB_USER }}
-        password: ${{ secrets.GTHB_PAT }}
-
-    - name: Push container
-      uses: viash-io/viash-actions/ns-build@v5
-      with:
-        config_mod: .functionality.version := '${{ needs.list.outputs.version }}'
-        platform: docker
-        src: ${{ matrix.component.dir }}
-        setup: push
+    uses: openproblems-bio/actions/.github/workflows/build.yml@main
+    with:
+      version: ${{ github.event.inputs.version }}

.github/workflows/test.yml

@@ -1,107 +1,9 @@
-name: test
+name: Test
 
 on:
-  pull_request:
   push:
-    branches: [ '**' ]
+  pull_request:
 
 jobs:
-  run_ci_check_job:
-    runs-on: ubuntu-latest
-    outputs:
-      run_ci: ${{ steps.github_cli.outputs.check }}
-    steps:
-      - name: 'Check if branch has an existing pull request and the trigger was a push'
-        id: github_cli
-        run: |
-          pull_request=$(gh pr list -R ${{ github.repository }} -H ${{ github.ref_name }} --json url --state open --limit 1 | jq '.[0].url')
-          # If the branch has a PR and this run was triggered by a push event, do not run
-          if [[ "$pull_request" != "null" && "$GITHUB_REF_NAME" != "main" && "${{ github.event_name == 'push' }}" == "true" && "${{ !contains(github.event.head_commit.message, 'ci force') }}" == "true" ]]; then
-            echo "check=false" >> $GITHUB_OUTPUT
-          else
-            echo "check=true" >> $GITHUB_OUTPUT
-          fi
-        env:
-          GITHUB_TOKEN: ${{ secrets.GTHB_PAT }}
-
-  # phase 1
-  list:
-    needs: run_ci_check_job
-    env:
-      s3_bucket: s3://openproblems-bio/public/neurips-2023-competition/workflow-resources/
-    runs-on: ubuntu-latest
-    if: ${{ needs.run_ci_check_job.outputs.run_ci == 'true' }}
-
-    outputs:
-      matrix: ${{ steps.set_matrix.outputs.matrix }}
-      cache_key: ${{ steps.cache.outputs.cache_key }}
-
-    steps:
-    - uses: actions/checkout@v4
-      with:
-        fetch-depth: 0 
-
-    - uses: viash-io/viash-actions/setup@v5
-
-    - uses: viash-io/viash-actions/project/sync-and-cache-s3@v5
-      id: cache
-      with:
-        s3_bucket: $s3_bucket
-        dest_path: resources
-        cache_key_prefix: resources__
-
-    - id: ns_list
-      uses: viash-io/viash-actions/ns-list@v5
-      with:
-        platform: docker
-        format: json
-
-    - id: ns_list_filtered
-      uses: viash-io/viash-actions/project/detect-changed-components@v5
-      with:
-        input_file: "${{ steps.ns_list.outputs.output_file }}"
-
-    - id: set_matrix
-      run: |
-        echo "matrix=$(jq -c '[ .[] | 
-          { 
-            "name": (.functionality.namespace + "/" + .functionality.name),
-            "config": .info.config
-          }
-        ]' ${{ steps.ns_list_filtered.outputs.output_file }} )" >> $GITHUB_OUTPUT
-
-  # phase 2
-  viash_test:
-    needs: list
-    if: ${{ needs.list.outputs.matrix != '[]' && needs.list.outputs.matrix != '' }}
-    runs-on: ubuntu-latest
-
-    strategy:
-      fail-fast: false
-      matrix:
-        component: ${{ fromJson(needs.list.outputs.matrix) }}
-
-    steps:
-    # Remove unnecessary files to free up space. Otherwise, we get 'no space left on device.'
-    - uses: data-intuitive/reclaim-the-bytes@v2
-
-    - uses: actions/checkout@v4
-
-    - uses: viash-io/viash-actions/setup@v5
-
-    # use cache
-    - name: Cache resources data
-      uses: actions/cache@v4
-      timeout-minutes: 10
-      with:
-        path: resources
-        key: ${{ needs.list.outputs.cache_key }}
-
-    - name: Run test
-      timeout-minutes: 30
-      run: |
-        VIASH_TEMP=$RUNNER_TEMP/viash viash test \
-          "${{ matrix.component.config }}" \
-          --cpus 2 \
-          --memory "5gb"
-
+  build:
+    uses: openproblems-bio/actions/.github/workflows/test.yml@main

_viash.yaml

@@ -1,15 +1,25 @@
-viash_version: 0.8.6
+name: task_perturbation_prediction
+version: 1.1.0
 
-source: src
-target: target
+# package metadata
+description: |
+  Predicting how small molecules change gene expression in different cell types.
+license: MIT
+keywords: [single-cell, perturbation prediction, perturbation, openproblems, benchmark]
+links:
+  issue_tracker: https://github.com/openproblems-bio/task_perturbation_prediction/issues
+  repository: https://github.com/openproblems-bio/task_perturbation_prediction
+  docker_registry: ghcr.io
 
+# technical settings
+organization: openproblems-bio
+viash_version: 0.9.0-RC6
+info:
+  test_resources:
+    - type: s3
+      path: s3://openproblems-data/resources/perturbation_prediction
+      dest: resources
+
+# set default labels
 config_mods: |
-  .functionality.version := 'dev'
-  .functionality.arguments[.multiple == true].multiple_sep := ';'
-  .platforms[.type == 'docker'].target_registry := 'ghcr.io'
-  .platforms[.type == 'docker'].target_organization := 'openproblems-bio/task_perturbation_prediction'
-  .platforms[.type == 'docker'].target_image_source := 'https://github.com/openproblems-bio/task_perturbation_prediction'
-  .platforms[.type == "nextflow"].directives.tag := "$id"
-  .platforms[.type == "nextflow"].auto.simplifyOutput := false
-  .platforms[.type == "nextflow"].config.labels := { lowmem : "memory = 20.Gb", midmem : "memory = 50.Gb", highmem : "memory = 100.Gb", lowcpu : "cpus = 5", midcpu : "cpus = 15", highcpu : "cpus = 30", lowtime : "time = 1.h", midtime : "time = 4.h", hightime : "time = 8.h", veryhightime : "time = 24.h" }
-  .platforms[.type == "nextflow"].config.script := "process.errorStrategy = 'ignore'"
+  .runners[.type == "nextflow"].config.labels := { lowmem : "memory = 20.Gb", midmem : "memory = 50.Gb", highmem : "memory = 100.Gb", lowcpu : "cpus = 5", midcpu : "cpus = 15", highcpu : "cpus = 30", lowtime : "time = 1.h", midtime : "time = 4.h", hightime : "time = 8.h", veryhightime : "time = 24.h" }

nextflow.config

@@ -1 +1,3 @@
 process.container = 'nextflow/bash:latest'
+
+process.errorStrategy = "ignore"

scripts/script.R

@@ -0,0 +1,48 @@
+library(tidyverse)
+
+# aws s3 sync s3://openproblems-data/resources/perturbation_prediction/results output/benchmark_results
+
+df <- yaml::read_yaml("output/benchmark_results/kaggle_2024-06-02_22-27-09/score_uns.yaml") %>%
+  map_dfr(as.data.frame) %>%
+  as_tibble
+
+df %>% filter(metric_ids == "mean_rowwise_rmse") %>% arrange(metric_values) %>% select(method_id, metric_values)
+
+#    method_id                       metric_values
+#    <chr>                                   <dbl>
+#  1 ground_truth                             0   
+#  2 nn_retraining_with_pseudolabels          1.29
+#  3 scape                                    1.31
+#  4 pyboost                                  1.32
+#  5 jn_ap_op2                                1.34
+#  6 lgc_ensemble                             1.41
+#  7 mean_across_compounds                    1.47
+#  8 transformer_ensemble                     1.55
+#  9 zeros                                    1.57
+# 10 mean_outcome                             1.57
+# 11 mean_across_celltypes                    2.50
+# 12 sample                                   3.02
+
+#######
+
+df <- yaml::read_yaml("output/benchmark_results/run_2024-06-02_22-27-09/score_uns.yaml") %>%
+  map_dfr(as.data.frame) %>%
+  as_tibble
+
+df %>% filter(metric_ids == "mean_rowwise_rmse") %>% arrange(metric_values) %>% select(method_id, metric_values)
+
+# # A tibble: 12 × 2
+#    method_id                       metric_values
+#    <chr>                                   <dbl>
+#  1 ground_truth                            0    
+#  2 nn_retraining_with_pseudolabels         0.757
+#  3 scape                                   0.775
+#  4 pyboost                                 0.795
+#  5 lgc_ensemble                            0.802
+#  6 mean_across_celltypes                   0.892
+#  7 jn_ap_op2                               0.894
+#  8 transformer_ensemble                    0.897
+#  9 mean_outcome                            0.899
+# 10 zeros                                   0.918
+# 11 mean_across_compounds                   0.943
+# 12 sample                                  1.36 

src/common/create_component/config.vsh.yaml

@@ -44,13 +44,15 @@ functionality:
     - type: python_script
       path: script.py
     - path: read_and_merge_yaml.py
-platforms:
+engines:
   - type: docker
     image: python:3.10-slim
     setup:
       - type: python
         pypi: ruamel.yaml
   - type: native
+runners:
+  - type: executable
   - type: nextflow
 
 

src/common/create_component/script.py

@@ -60,7 +60,9 @@ def create_config(par, component_type, pretty_name, script_path) -> str:
     |  # This platform allows running the component natively
     |  - type: native
     |  # Allows turning the component into a Nextflow module / pipeline.
-    |  - type: nextflow
+    |runners:
+  - type: executable
+  - type: nextflow
     |    directives:
     |      label: [midtime,midmem,midcpu]
     |'''

src/common/sync_test_resources/config.vsh.yaml

@@ -34,8 +34,10 @@ functionality:
   resources:
     - type: bash_script
       path: script.sh
-platforms:
+engines:
   - type: docker
     image: "amazon/aws-cli:2.7.12"
   - type: native
+runners:
+  - type: executable
   - type: nextflow