[feature] [newvir] Add RSV-B

group_vars/loculus/main.yml

@@ -1,6 +1,9 @@
 name: 'W-ASAP'
 host: 'db.wasap.genspectrum.org'
 
+# Pin the Loculus version
+#loculus_commit_sha: 4948577ed
+
 logo:
   url: 'https://db.wasap.genspectrum.org/static/images/wasap-logo.png'
   width: 200
@@ -39,7 +42,7 @@ s3:
     bucket: wasap
 
 organisms:
-   covid:
+  covid:
     schema:
       organismName: 'SARS-CoV-2'
       image: 'https://db.wasap.genspectrum.org/static/images/covid-virus.png'
@@ -111,11 +114,8 @@ organisms:
         args:
           - "--watch"
           - "--disableConsensusSequences"
-    referenceGenomes:
-      singleReference:
-        nucleotideSequences: []
-        genes: []
-   rsva:
+    referenceGenomes: []
+  rsva:
     schema:
       organismName: 'RSV-A'
       image: 'https://db.wasap.genspectrum.org/static/images/rsv-virus.png'
@@ -187,10 +187,80 @@ organisms:
         args:
           - "--watch"
           - "--disableConsensusSequences"
-    referenceGenomes:
-      singleReference:
-        nucleotideSequences: []
-        genes: []
+    referenceGenomes: []
+  rsvb:
+    schema:
+      organismName: 'RSV-B'
+      image: 'https://db.wasap.genspectrum.org/static/images/rsv-virus.png'
+      metadata:
+        - name: date
+          type: date
+          header: Sample details
+          displayName: Submission Date
+        - name: sampleId
+          type: string
+          header: Sample details
+          displayName: Sample ID
+          initiallyVisible: true
+        - name: batchId
+          type: string
+          definition: "Sequencing batch"
+          header: Sample details
+          displayName: Batch ID
+        - name: locationCode
+          type: string
+          header: Sample details
+          displayName: Location Code
+        - name: locationName
+          type: string
+          generateIndex: true
+          autocomplete: true
+          header: Sample details
+          displayName: Location
+          initiallyVisible: true
+        - name: samplingDate
+          type: date
+          header: Sample details
+          displayName: Sampling Date
+          initiallyVisible: true
+        - name: sr2siloVersion
+          type: string
+          header: srSILO info
+          definition: "Version of sr2silo used for preprocessing"
+          displayName: sr2silo Version
+          hideOnSequenceDetailsPage: true
+          hideInSearchResultsTable: true
+        - name: countSiloReads
+          type: string
+          header: srSILO info
+          hideOnSequenceDetailsPage: true
+          hideInSearchResultsTable: true
+          definition: "Number of reads in subsampled srSILO alignments"
+          displayName: srSILO Read Count
+      website:
+        tableColumns:
+          - date
+          - locationName
+          - samplingDate
+        defaultOrder: ascending
+        defaultOrderBy: date
+      submissionDataTypes:
+        consensusSequences: false
+        files:
+          enabled: true
+          categories:
+            - name: nucleotideAlignment
+            - name: siloReads
+      files:
+        - name: nucleotideAlignment
+        - name: siloReads
+    preprocessing:
+      - version: 1
+        image: ghcr.io/loculus-project/preprocessing-dummy
+        args:
+          - "--watch"
+          - "--disableConsensusSequences"
+    referenceGenomes: []
 
 dataUseTerms:
   enabled: false

group_vars/srsilo/main.yml

@@ -22,6 +22,7 @@ srsilo_group: srsilo
 srsilo_enabled_viruses:
   - covid
   - rsva
+  - rsvb
 
 # Per-virus configuration overrides
 # Use this to override global defaults for specific viruses
@@ -36,6 +37,11 @@ srsilo_virus_config:
     fetch_max_reads: 50000000     # 50M reads - lower volume expected
     chunk_size: 500000            # Smaller chunks
     docker_memory_limit: 50g
+  rsvb:
+    fetch_days: 120
+    fetch_max_reads: 50000000     # 50M reads - lower volume expected
+    chunk_size: 500000            # Smaller chunks
+    docker_memory_limit: 50g
 
 # =============================================================================
 # Repository Path

playbooks/srsilo/vars/test_vars.yml

@@ -25,6 +25,11 @@ srsilo_virus_config:
     fetch_max_reads: 2000000        # Limit to 2M reads instead of 50M
     chunk_size: 30000               # Reduced for 8GB RAM test environment
     docker_memory_limit: 7g         # 90% of 8GB RAM
+  rsvb:
+    fetch_days: 45                  # Fetch only 45 days instead of 90
+    fetch_max_reads: 2000000        # Limit to 2M reads instead of 50M
+    chunk_size: 30000               # Reduced for 8GB RAM test environment
+    docker_memory_limit: 7g         # 90% of 8GB RAM
 
 # Shared configuration (same as production)
 srsilo_api_base_url: https://api.db.wasap.genspectrum.org

roles/loculus/templates/values.yaml.j2

@@ -1,5 +1,8 @@
 name: '{{ name }}'
 host: '{{ host }}'
+{% if loculus_commit_sha is defined %}
+sha: '{{ loculus_commit_sha | truncate(7, True, "") }}'
+{% endif %}
 environment: '{{ loculus_environment }}' # Do not change this: this setup uses a "local" environment as the public traffic is proxied through nginx.
 
 runDevelopmentMainDatabase: {{ runDevelopmentMainDatabase | lower }}
@@ -138,13 +141,11 @@ organisms:
 {% endfor %}
 {% endfor %}
     referenceGenomes:
-      singleReference:
-        nucleotideSequences: {{ organism_config.referenceGenomes.singleReference.nucleotideSequences }}
-        genes: {{ organism_config.referenceGenomes.singleReference.genes }}
+      {{ organism_config.referenceGenomes }}
 {% endfor %}
 
 dataUseTerms:
   enabled: {{ dataUseTerms.enabled | lower }}
 disableIngest: {{ disableIngest | lower }}
 disableEnaSubmission: {{ disableEnaSubmission | lower }}
-accessionPrefix: "{{ accessionPrefix }}"
+accessionPrefix: "{{ accessionPrefix }}"

roles/nginx/defaults/main.yml

@@ -17,8 +17,10 @@ nginx_loculus_ssl_certificate_key_path: "/etc/letsencrypt/live/db.{{ nginx_domai
 # Upstream Service Ports - Per-virus LAPIS/SILO instances
 nginx_lapis_covid_port: 8083
 nginx_lapis_rsva_port: 8084
+nginx_lapis_rsvb_port: 8085
 nginx_silo_covid_port: 8081
 nginx_silo_rsva_port: 8082
+nginx_silo_rsvb_port: 8086
 nginx_wasap_scout_port: 9001
 nginx_loculus_port: 9000
 nginx_loculus_auth_port: 9083

roles/nginx/templates/conf.d/wasap.conf.j2

@@ -11,6 +11,10 @@ upstream lapis-rsva {
     server 127.0.0.1:{{ nginx_lapis_rsva_port }};
 }
 
+upstream lapis-rsvb {
+    server 127.0.0.1:{{ nginx_lapis_rsvb_port }};
+}
+
 # SILO instances per virus
 upstream silo-covid {
     server 127.0.0.1:{{ nginx_silo_covid_port }};
@@ -20,6 +24,10 @@ upstream silo-rsva {
     server 127.0.0.1:{{ nginx_silo_rsva_port }};
 }
 
+upstream silo-rsvb {
+    server 127.0.0.1:{{ nginx_silo_rsvb_port }};
+}
+
 # Default HTTP to HTTPS redirect
 server {
     listen 80 default_server;

roles/nginx/templates/sites-available/silo.j2

@@ -31,6 +31,21 @@ server {
         proxy_set_header X-Forwarded-Prefix /rsva/;
     }
 
+    # Redirect /rsvb without trailing slash
+    location = /rsvb {
+        return 301 /rsvb/;
+    }
+
+    # RSVB instance - proxy with path stripped via trailing slash
+    location /rsvb/ {
+        proxy_pass http://silo-rsvb/;
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+        proxy_set_header X-Forwarded-Prefix /rsvb/;
+    }
+
     # Backward compatibility: proxy all other requests to COVID backend
     # This catches /, /info, etc. and proxies directly (no redirect)
     # Using proxy instead of redirect to preserve CORS preflight behavior

roles/srsilo/defaults/main.yml

@@ -43,6 +43,11 @@ srsilo_viruses:
     instance_name: wise-rsva
     lapis_port: 8084
     silo_port: 8082
+  rsvb:
+    organism: rsvb
+    instance_name: wise-rsvb
+    lapis_port: 8085
+    silo_port: 8086
 
 # List of enabled viruses (for multi-virus playbooks)
 # This controls which viruses are actually processed - override in group_vars

roles/srsilo/files/viruses/rsvb/database_config.yaml

@@ -0,0 +1,28 @@
+schema:
+  metadata:
+    - name: sampleId
+      type: string
+      generateIndex: true
+    - name: batchId
+      type: string
+      generateIndex: true
+    - name: locationCode
+      type: string
+      generateIndex: true
+    - name: samplingDate
+      type: date
+      generateIndex: false
+    - name: locationName
+      type: string
+      generateIndex: true
+    - name: readId
+      type: string
+    - name: sr2siloVersion
+      type: string
+      generateIndex: true
+  opennessLevel: OPEN
+  instanceName: wise-rsvb
+  features:
+    - name: generalizedAdvancedQuery
+  primaryKey: readId
+siloClientThreadCount: 96

roles/srsilo/files/viruses/rsvb/preprocessing_config.yaml

@@ -0,0 +1,2 @@
+ndjsonInputFilename: sorted.ndjson.zst
+withoutUnalignedSequences: true