Changed metadata for supplied algorithms

.github/actions/respro-build-qc/action.yml

@@ -0,0 +1,132 @@
+name: respro build QC
+description: |
+  Install ResistanceProfiler from a GitHub release tag (no conda, no autobump
+  of a pinned version) and run the build quality-control step against a
+  converted database's output artifacts. Used by both the autobump workflows
+  and the PR-triggered build-check workflow so every build is validated
+  against the same respro version.
+
+inputs:
+  source-name:
+    description: Source database name (e.g. stanford-hiv). Passed to respro as the project name.
+    required: true
+  output-dir:
+    description: Directory containing rules.tsv / metadata.json / formula-rules.tsv.
+    required: true
+  work-dir:
+    description: Working directory for GenBank fetch + built .db + logs.
+    required: false
+    default: /tmp/respro_qc
+  respro-version:
+    description: |
+      ResistanceProfiler version to build from. Defaults to "latest", which
+      resolves the newest tag on the the-foxlab/ResistanceProfiler repository
+      at run time. Set to a concrete tag (e.g. "v0.1.3") to pin a specific
+      release.
+    required: false
+    default: latest
+
+outputs:
+  qc_failed:
+    description: '"true" when the respro build QC reported a hard failure, otherwise "false".'
+    value: ${{ steps.qc_summary.outputs.qc_failed }}
+  summary:
+    description: 'Multi-line markdown summary of the build QC, suitable for PR bodies.'
+    value: ${{ steps.qc_summary.outputs.block }}
+
+runs:
+  using: composite
+
+  steps:
+    - name: Set up Python
+      uses: actions/setup-python@v5
+      with:
+        python-version: "3.12"
+
+    - name: Resolve latest respro release tag
+      if: inputs.respro-version == 'latest'
+      id: respro_tag
+      shell: bash
+      run: |
+        set -euo pipefail
+        # Fetch the newest non-web tag from the ResistanceProfiler repo.
+        # Tags are prefixed with "v" (e.g. v0.1.3); web releases use "web-v".
+        tag=$(curl -fsSL \
+          -H "Accept: application/vnd.github+json" \
+          "https://api.github.com/repos/the-foxlab/ResistanceProfiler/tags" \
+          | python -c '
+        import json, sys
+        tags = json.load(sys.stdin)
+        for t in tags:
+            name = t.get("name", "")
+            if name.startswith("v") and not name.startswith("web-"):
+                print(name)
+                break
+        else:
+            sys.exit("No non-web respro release tag found")
+        ')
+        echo "tag=${tag}" >> "$GITHUB_OUTPUT"
+        echo "Resolved latest respro release tag: ${tag}"
+
+    - name: Determine respro version
+      id: respro_version
+      shell: bash
+      run: |
+        if [ "${{ inputs.respro-version }}" = "latest" ]; then
+          echo "version=${{ steps.respro_tag.outputs.tag }}" >> "$GITHUB_OUTPUT"
+        else
+          echo "version=${{ inputs.respro-version }}" >> "$GITHUB_OUTPUT"
+        fi
+
+    - name: Install respro from release tag
+      shell: bash
+      run: |
+        set -euo pipefail
+        python -m pip install --upgrade pip
+        # Build from the tagged source tarball. The runner has a C compiler
+        # so the mappy native dependency builds from source without conda.
+        version="${{ steps.respro_version.outputs.version }}"
+        echo "Installing ResistanceProfiler ${version} from source"
+        python -m pip install \
+          "respro @ git+https://github.com/the-foxlab/ResistanceProfiler.git@${version}"
+        respro --version
+
+    - name: respro build QC
+      id: respro_qc
+      shell: bash
+      continue-on-error: true
+      run: |
+        python scripts/respro_build_qc.py \
+          --source-name "${{ inputs.source-name }}" \
+          --output-dir "${{ inputs.output-dir }}" \
+          --work-dir "${{ inputs.work-dir }}"
+        echo "qc_exit=$?" >> "$GITHUB_OUTPUT"
+
+    - name: Read respro QC summary
+      id: qc_summary
+      shell: bash
+      run: |
+        summary_path="${{ inputs.work-dir }}/qc_summary.md"
+        EOF=$(dd if=/dev/urandom bs=15 count=1 status=none | base64)
+        echo "block<<${EOF}" >> "$GITHUB_OUTPUT"
+        qc_failed="false"
+        if [ ! -f "$summary_path" ]; then
+          # The QC script exited before writing a summary (infrastructure
+          # error: crash, missing respro, import failure). Emit a synthetic
+          # failure block so the PR comment still renders and the workflow
+          # fails cleanly via qc_failed=true.
+          echo "## respro build QC" >> "$GITHUB_OUTPUT"
+          echo "" >> "$GITHUB_OUTPUT"
+          echo "**Source:** \`${{ inputs.source-name }}\`" >> "$GITHUB_OUTPUT"
+          echo "**Status:** ⚠️ **\`respro init\` FAILED** — the QC step did not produce a summary (infrastructure error before QC ran)." >> "$GITHUB_OUTPUT"
+          echo "**Exit code:** \`unknown\`" >> "$GITHUB_OUTPUT"
+          qc_failed="true"
+        else
+          cat "$summary_path" >> "$GITHUB_OUTPUT"
+          if [ -f "${{ inputs.work-dir }}/qc_summary.json" ]; then
+            status=$(python -c "import json;print(json.load(open('${{ inputs.work-dir }}/qc_summary.json'))['status'])" 2>/dev/null || echo "")
+            if [ "$status" = "failed" ]; then qc_failed="true"; fi
+          fi
+        fi
+        echo "${EOF}" >> "$GITHUB_OUTPUT"
+        echo "qc_failed=${qc_failed}" >> "$GITHUB_OUTPUT"

.github/workflows/herpesdrg-autobump.yml

@@ -100,6 +100,15 @@ jobs:
         run: |
           python scripts/build_metadata_manifest.py
 
+      - name: respro build QC
+        id: qc
+        if: steps.update_check.outputs.should_update == 'true'
+        uses: ./.github/actions/respro-build-qc
+        with:
+          source-name: herpesdrg
+          output-dir: databases/herpesdrg/output
+          work-dir: /tmp/respro_qc
+
       - name: No update detected
         if: steps.update_check.outputs.should_update != 'true'
         run: |
@@ -122,6 +131,8 @@ jobs:
             - Source commit SHA: ${{ steps.source.outputs.source_commit_sha }}
             - Source URL: ${{ steps.source.outputs.source_url }}
 
+            ${{ steps.qc.outputs.summary }}
+
             This PR was generated by the autobump workflow.
           labels: autobump
           add-paths: |
@@ -130,3 +141,9 @@ jobs:
             databases/herpesdrg/output/metadata.json
             databases/herpesdrg/output/non-migrated-rules.txt
             databases/manifest.json
+
+      - name: Mark workflow failed on respro build error
+        if: steps.update_check.outputs.should_update == 'true' && steps.qc.outputs.qc_failed == 'true'
+        run: |
+          echo "respro build QC failed for herpesdrg; see the PR body and the 'respro build QC' step log."
+          exit 1

.github/workflows/hsv-daehne-jaki-autobump.yml

@@ -117,6 +117,15 @@ jobs:
         run: |
           python scripts/build_metadata_manifest.py
 
+      - name: respro build QC
+        id: qc
+        if: steps.update_check.outputs.should_update == 'true'
+        uses: ./.github/actions/respro-build-qc
+        with:
+          source-name: hsv_daehne_jaki
+          output-dir: databases/hsv_daehne_jaki/output
+          work-dir: /tmp/respro_qc
+
       - name: No update detected
         if: steps.update_check.outputs.should_update != 'true'
         run: |
@@ -139,6 +148,8 @@ jobs:
             - Zenodo updated timestamp: ${{ steps.zenodo.outputs.updated }}
             - Source file: ${{ steps.zenodo.outputs.source_url }}
 
+            ${{ steps.qc.outputs.summary }}
+
             This PR was generated by the autobump workflow.
           labels: autobump
           add-paths: |
@@ -147,3 +158,9 @@ jobs:
             databases/hsv_daehne_jaki/output/metadata.json
             databases/hsv_daehne_jaki/output/non-migrated-rules.txt
             databases/manifest.json
+
+      - name: Mark workflow failed on respro build error
+        if: steps.update_check.outputs.should_update == 'true' && steps.qc.outputs.qc_failed == 'true'
+        run: |
+          echo "respro build QC failed for hsv_daehne_jaki; see the PR body and the 'respro build QC' step log."
+          exit 1

.github/workflows/pr-build-check.yml

@@ -0,0 +1,97 @@
+name: PR database build check
+
+on:
+  pull_request:
+    paths:
+      - "databases/**"
+      - "scripts/respro_build_qc.py"
+      - ".github/actions/respro-build-qc/**"
+      - ".github/workflows/pr-build-check.yml"
+
+permissions:
+  contents: read
+  pull-requests: write
+
+jobs:
+  detect-changed-databases:
+    runs-on: ubuntu-latest
+    outputs:
+      herpesdrg: ${{ steps.filter.outputs.herpesdrg }}
+      hsv_daehne_jaki: ${{ steps.filter.outputs.hsv_daehne_jaki }}
+      stanford-hiv: ${{ steps.filter.outputs.stanford-hiv }}
+      any: ${{ steps.filter.outputs.any }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Detect changed databases
+        id: filter
+        uses: dorny/paths-filter@v3
+        with:
+          # `any` covers shared infrastructure (QC script, composite action,
+          # this workflow) so a change there triggers a full build check.
+          filters: |
+            herpesdrg:
+              - 'databases/herpesdrg/**'
+            hsv_daehne_jaki:
+              - 'databases/hsv_daehne_jaki/**'
+            stanford-hiv:
+              - 'databases/stanford-hiv/**'
+            any:
+              - 'databases/herpesdrg/**'
+              - 'databases/hsv_daehne_jaki/**'
+              - 'databases/stanford-hiv/**'
+              - 'scripts/respro_build_qc.py'
+              - '.github/actions/respro-build-qc/**'
+              - '.github/workflows/pr-build-check.yml'
+
+  build-check:
+    needs: detect-changed-databases
+    if: needs.detect-changed-databases.outputs.any == 'true'
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        # One job per database; the `if` guard on each step skips databases
+        # that were not touched by the PR.
+        database:
+          - name: herpesdrg
+            dir: databases/herpesdrg/output
+            changed: ${{ needs.detect-changed-databases.outputs.herpesdrg }}
+          - name: hsv_daehne_jaki
+            dir: databases/hsv_daehne_jaki/output
+            changed: ${{ needs.detect-changed-databases.outputs.hsv_daehne_jaki }}
+          - name: stanford-hiv
+            dir: databases/stanford-hiv/output
+            changed: ${{ needs.detect-changed-databases.outputs.stanford-hiv }}
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Run respro build QC
+        if: matrix.database.changed == 'true'
+        id: qc
+        uses: ./.github/actions/respro-build-qc
+        with:
+          source-name: ${{ matrix.database.name }}
+          output-dir: ${{ matrix.database.dir }}
+          work-dir: /tmp/respro_qc_${{ matrix.database.name }}
+
+      - name: Comment QC summary on PR
+        if: matrix.database.changed == 'true' && always()
+        uses: peter-evans/create-or-update-comment@v4
+        with:
+          issue-number: ${{ github.event.pull_request.number }}
+          body: |
+            ### respro build QC — `${{ matrix.database.name }}`
+
+            ${{ steps.qc.outputs.summary }}
+
+            _Failed builds must be fixed before merge._
+
+      - name: Fail on respro build error
+        if: matrix.database.changed == 'true' && steps.qc.outputs.qc_failed == 'true'
+        run: |
+          echo "respro build QC failed for ${{ matrix.database.name }}; see the PR comment and the build logs."
+          exit 1

.github/workflows/stanford-hiv-autobump.yml

@@ -120,6 +120,15 @@ jobs:
         run: |
           python scripts/build_metadata_manifest.py
 
+      - name: respro build QC
+        id: qc
+        if: steps.update_check.outputs.should_update == 'true'
+        uses: ./.github/actions/respro-build-qc
+        with:
+          source-name: stanford-hiv
+          output-dir: databases/stanford-hiv/output
+          work-dir: /tmp/respro_qc
+
       - name: No update detected
         if: steps.update_check.outputs.should_update != 'true'
         run: |
@@ -142,11 +151,19 @@ jobs:
             - Source version: ${{ steps.source.outputs.source_version }}
             - Source URL: ${{ steps.source.outputs.source_url }}
 
+            ${{ steps.qc.outputs.summary }}
+
             This PR was generated by the autobump workflow.
           labels: autobump
           add-paths: |
             databases/stanford-hiv/output/rules.tsv
             databases/stanford-hiv/output/formula-rules.tsv
             databases/stanford-hiv/output/metadata.json
             databases/stanford-hiv/output/non-migrated-rules.txt
-            databases/manifest.json
+            databases/manifest.json
+
+      - name: Mark workflow failed on respro build error
+        if: steps.update_check.outputs.should_update == 'true' && steps.qc.outputs.qc_failed == 'true'
+        run: |
+          echo "respro build QC failed for stanford-hiv; see the PR body and the 'respro build QC' step log."
+          exit 1

.gitignore

@@ -6,7 +6,7 @@ __pycache__/
 
 .pytest_cache/
 .mypy_cache/
-
+charmd/
 .act/
 
 .DS_Store

README.md

@@ -22,6 +22,8 @@ databases/
   manifest.json                 # generated discovery file, see below
 
   <source_name>/
+    example/
+      example.fasta
     scripts/
       convert.py
       requirements.txt
@@ -45,6 +47,13 @@ It's built from all `databases/*/output/metadata.json` files, and each entry con
 name, relative paths to `metadata.json`, `rules.tsv`, and `formula-rules.tsv` (empty when absent),
 and the embedded metadata content.
 
+The manifest's top-level `respro_version` field is a PEP 440 version specifier (e.g. `">=0.1.3"`)
+stating the minimum ResPro release required to build the listed databases. ResPro clients compare
+their own installed version against this specifier and prompt the user to upgrade when too old,
+so breaking changes to the TSV contract or metadata schema can be gated on a known ResPro release.
+Bump this value in `scripts/build_metadata_manifest.py` (`RESPRO_VERSION_REQUIREMENT`) whenever a
+newer ResPro feature becomes a hard requirement.
+
 This is the primariy file that is used to display supported databases via the ResPro command:
 
 ```bash

databases/herpesdrg/output/metadata.json

@@ -44,18 +44,16 @@
     },
     {
       "name": "drug_interpretation",
-      "method": "by_phenotype",
-      "thresholds": {
-        "resistant": 1,
-        "intermediate": 1
-      }
+      "method": "by_phenotype"
     },
     {
       "name": "drug_interpretation",
       "method": "by_fold_ic50",
       "thresholds": {
+        "susceptible": 0.0,
+        "low-level resistance": 2.0,
         "intermediate": 5.0,
-        "resistant": 15.0
+        "high-level resistance": 15.0
       }
     },
     {