refactor: move minimizerReferences from files to buildFiles

.github/workflows/build-and-deploy.yml

@@ -2,18 +2,14 @@ name: build-and-deploy
 
 on:
   push:
-    branches:
-      - master
-      - staging
-      - release
   pull_request:
   repository_dispatch:
     types: build-and-deploy
   workflow_dispatch:
   workflow_call:
 
 concurrency:
-  group: cli-${{ github.workflow }}-${{ github.ref_type }}-${{ github.event.pull_request.number || github.ref || github.run_id }}
+  group: cli-${{ github.workflow }}-${{ github.ref_type }}-${{ github.head_ref || github.ref_name || github.run_id }}
   cancel-in-progress: true
 
 defaults:

.github/workflows/validate-pr.yml

@@ -0,0 +1,56 @@
+name: "validate-pr"
+
+on:
+  pull_request:
+
+permissions:
+  contents: "read"
+
+concurrency:
+  group: "validate-pr-${{ github.event.pull_request.number || github.run_id }}"
+  cancel-in-progress: true
+
+defaults:
+  run:
+    shell: "bash -euxo pipefail {0}"
+
+env:
+  PYTHONUNBUFFERED: "1"
+
+jobs:
+  validate:
+    runs-on: "ubuntu-24.04"
+
+    steps:
+      - name: "Checkout PR code"
+        uses: "actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683" # v4.2.2
+        with:
+          repository: "${{ github.event.pull_request.head.repo.full_name }}"
+          ref: "${{ github.event.pull_request.head.ref }}"
+          fetch-depth: 0
+          fetch-tags: true
+
+      - name: "Install system dependencies"
+        run: "sudo apt-get install genometools python3 --yes -qq >/dev/null"
+
+      - name: "Install Python dependencies"
+        run: "pip3 install -r requirements.txt"
+
+      - name: "Validate GFF files"
+        run: "./scripts/validate-gff 'data/'"
+
+      - name: "Rebuild (validation only)"
+        id: "rebuild"
+        continue-on-error: true
+        run: "./scripts/rebuild --input-dir 'data/' --output-dir 'data_output/' --no-pull --allow-dirty"
+
+      - name: "Upload validation artifacts"
+        if: "always()"
+        uses: "actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02" # v4.6.2
+        with:
+          name: "validation-output"
+          path: "data_output/"
+
+      - name: "Fail if rebuild or validation found errors"
+        if: "steps.rebuild.outcome == 'failure'"
+        run: "exit 1"

data/community/collection.json

@@ -37,6 +37,7 @@
     "community/pathoplexus/cchfv/M",
     "community/v-gen-lab/chikV/genotypes",
     "community/masphl-bioinformatics/hav/vp1-2b-junction",
-    "community/masphl-bioinformatics/hav/whole-genome"
+    "community/masphl-bioinformatics/hav/whole-genome",
+    "community/itps/zikav"
   ]
 }

data/community/isuvdl/mazeller/prrsv1/orf5/yimim2025/CHANGELOG.md

@@ -1,3 +1,7 @@
+## Unreleased
+
+- Remove invalid `qc.missingData.scoreWeight` and `qc.mixedSites.scoreWeight`
+
 ## 2025-09-09T12:13:13Z
 
 Add schema definition url to `pathogen.json`. This is a purely technical change, for convenience of dataset authors. The data itself is not modified.

data/community/isuvdl/mazeller/prrsv1/orf5/yimim2025/pathogen.json

@@ -27,8 +27,7 @@
     "missingData": {
       "enabled": true,
       "missingDataThreshold": 2000,
-      "scoreBias": 500,
-      "scoreWeight": 50
+      "scoreBias": 500
     },
     "snpClusters": {
       "enabled": true,
@@ -38,8 +37,7 @@
     },
     "mixedSites": {
       "enabled": true,
-      "mixedSitesThreshold": 15,
-      "scoreWeight": 80
+      "mixedSitesThreshold": 15
     },
     "frameShifts": {
       "enabled": true,

data/community/isuvdl/mazeller/prrsv2/orf5/yimim2023/CHANGELOG.md

@@ -1,3 +1,7 @@
+## Unreleased
+
+- Unify attributes per schema update
+
 ## 2025-09-09T12:13:13Z
 
 Add schema definition url to `pathogen.json`. This is a purely technical change, for convenience of dataset authors. The data itself is not modified.

data/community/isuvdl/mazeller/prrsv2/orf5/yimim2023/pathogen.json

@@ -3,15 +3,13 @@
   "attributes": {
     "name": "PRRSV-2 ORF5 Lineages, Yim-im & Zhang 2023 Microbiol Spectr",
     "reference accession": "DQ478308.1",
-    "reference name": "PRRSV0004437"
+    "reference name": "PRRSV0004437",
+    "experimental": true
   },
   "compatibility": {
     "cli": "3.0.0-alpha.0",
     "web": "3.0.0-alpha.0"
   },
-  "deprecated": false,
-  "enabled": true,
-  "experimental": true,
   "files": {
     "changelog": "CHANGELOG.md",
     "examples": "sequences.fasta",

data/community/itps/zikav/CHANGELOG.md

@@ -0,0 +1,10 @@
+## Unreleased
+
+- Move `placementMaskRanges` to tree.json
+- Unify attributes per schema update
+
+## 2026-03-04T12:40:26Z
+
+Initial release of a Zika Virus (zikav) dataset for genotype classification!
+
+Read more about Nextclade datasets in the documentation: https://docs.nextstrain.org/projects/nextclade/en/stable/user/datasets.htmls

data/community/itps/zikav/README.md

@@ -0,0 +1,19 @@
+# Nextclade Dataset for Zika virus Based on Reference "Natal RGN"
+
+## Dataset Attributes
+
+| Attribute            | Value                                    |
+| -------------------- | ---------------------------------------- |
+| Name                 | Zika virus                               |
+| RefName              | Natal RGN                                |
+| RefAccession         | NC_035889.1                              |
+
+## Scope of This Dataset
+
+The dataset aims to enable rapid and accurate identification of the two main Zika virus genotypes, improving the resolution and speed of phylogenetic classification in routine surveillance networks.
+
+The source code is available at [InstitutoTodosPelaSaude/nextclade-datasets-workflows](https://github.com/InstitutoTodosPelaSaude/nextclade-datasets-workflows/tree/main/zikav).
+
+For bugs, please open an [issue](https://github.com/InstitutoTodosPelaSaude/nextclade-datasets-workflows/issues).
+
+Read more about Nextclade datasets in the Nextclade documentation: [Nextclade Datasets](https://docs.nextstrain.org/projects/nextclade/en/stable/user/datasets.html).

data/community/itps/zikav/genome_annotation.gff3

@@ -0,0 +1,21 @@
+##gff-version 3
+#!gff-spec-version 1.21
+#!processor NCBI annotwriter
+##sequence-region NC_035889.1 1 10808
+##species https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?id=64320
+NC_035889.1	RefSeq	region	1	10808	.	+	.	ID=NC_035889.1:1..10808;Dbxref=taxon:64320;collection-date=2015;country=Brazil: Rio Grande do Norte%2C Natal;gbkey=Src;genome=genomic;isolation-source=fetus' brain autopsy;mol_type=genomic RNA;nat-host=Homo sapiens;strain=Natal RGN
+NC_035889.1	RefSeq	region	1	10808	.	+	.	ID=id-NC_035889.1:1..10808;Note=Mature peptides were annotated by RefSeq staff using homoloy to NC_012532.1 and the cleavage sites reported in Kuno and Chang%2C 2007 (PMID 17195954). Questions about the annotation of this sequence should be directed to info@ncbi.nlm.nih.gov.;gbkey=Comment
+NC_035889.1	GenBank	CDS	108	473	.	+	.	gene_name=ancC
+NC_035889.1	GenBank	CDS	108	419	.	+	.	gene_name=C
+NC_035889.1	GenBank	CDS	474	977	.	+	.	gene_name=prM
+NC_035889.1	GenBank	CDS	474	752	.	+	.	gene_name=pr
+NC_035889.1	GenBank	CDS	753	977	.	+	.	gene_name=M
+NC_035889.1	GenBank	CDS	978	2489	.	+	.	gene_name=E
+NC_035889.1	GenBank	CDS	2490	3545	.	+	.	gene_name=NS1
+NC_035889.1	GenBank	CDS	3546	4223	.	+	.	gene_name=NS2A
+NC_035889.1	GenBank	CDS	4224	4613	.	+	.	gene_name=NS2B
+NC_035889.1	GenBank	CDS	4614	6464	.	+	.	gene_name=NS3
+NC_035889.1	GenBank	CDS	6465	6845	.	+	.	gene_name=NS4A
+NC_035889.1	GenBank	CDS	6846	6914	.	+	.	gene_name=2K
+NC_035889.1	GenBank	CDS	6915	7667	.	+	.	gene_name=NS4B
+NC_035889.1	GenBank	CDS	7668	10376	.	+	.	gene_name=NS5

data/community/itps/zikav/pathogen.json

@@ -0,0 +1,63 @@
+{
+  "$schema": "https://raw.githubusercontent.com/nextstrain/nextclade/refs/heads/release/packages/nextclade-schemas/input-pathogen-json.schema.json",
+  "alignmentParams": {
+    "retryReverseComplement": true,
+    "alignmentPreset": "high-diversity"
+  },
+  "attributes": {
+    "name": "Zika virus",
+    "reference accession": "NC_035889.1",
+    "reference name": "Natal RGN",
+    "experimental": true
+  },
+  "compatibility": {
+    "cli": "3.0.0-alpha.0",
+    "web": "3.0.0-alpha.0"
+  },
+  "files": {
+    "changelog": "CHANGELOG.md",
+    "examples": "sequences.fasta",
+    "genomeAnnotation": "genome_annotation.gff3",
+    "pathogenJson": "pathogen.json",
+    "readme": "README.md",
+    "reference": "reference.fasta",
+    "treeJson": "tree.json"
+  },
+  "meta": {
+    "bugs": "https://github.com/InstitutoTodosPelaSaude/nextclade-datasets-workflows/issues",
+    "source code": "https://github.com/InstitutoTodosPelaSaude/nextclade-datasets-workflows/tree/main/zikav"
+  },
+  "qc": {
+    "frameShifts": {
+      "enabled": true
+    },
+    "missingData": {
+      "enabled": true,
+      "missingDataThreshold": 1625,
+      "scoreBias": 536
+    },
+    "mixedSites": {
+      "enabled": false,
+      "mixedSitesThreshold": 11
+    },
+    "privateMutations": {
+      "cutoff": 35,
+      "enabled": true,
+      "typical": 5,
+      "weightLabeledSubstitutions": 2,
+      "weightReversionSubstitutions": 1,
+      "weightUnlabeledSubstitutions": 1
+    },
+    "snpClusters": {
+      "enabled": false
+    },
+    "stopCodons": {
+      "enabled": true
+    }
+  },
+  "schemaVersion": "3.0.0",
+  "version": {
+    "tag": "unreleased"
+  },
+  "defaultCds": "E"
+}