feat(pipeline): VEP database mode, and a pipeline-specific database URL

Makes a real annotation runnable locally without the 25 GB VEP cache, which is what
the demo needs and what a reviewer can reproduce in minutes.

- params.vep_database (VEP_DATABASE=true) queries Ensembl's public database instead of
  a local cache. Slower per variant and fewer fields, so --everything is swapped for the
  flags the loader actually stores. Its cache placeholder is NO_CACHE, not NO_FILE:
  Nextflow rejects two staged inputs sharing a filename.
- PIPELINE_DATABASE_URL is handed to the pipeline when set. The loader runs inside a
  container, where the API's own localhost URL would point at the container itself.
- README: how to run the UI's annotate button locally against host Nextflow + Docker.

Verified end to end on pipeline/tests/data/tiny.vcf: bcftools norm split the multiallelic
record, VEP 113 annotated 4 variants live, the loader wrote them and marked the job
succeeded, and the UI shows them. The deletion came back as 22:42126611 CT>C with exact
VCF alleles, which is the case the audit's ID-tagging fix exists for.

Tests: api 51, loader 16, stub run 3/3; ruff, mypy clean.

pipeline/main.nf

@@ -14,7 +14,10 @@ workflow {
     def must_exist = !workflow.stubRun
 
     vcf_ch  = Channel.fromPath(params.vcf, checkIfExists: true)
-    cache   = file(params.vep_cache, checkIfExists: must_exist)
+    // In database mode there is no cache to stage. Its placeholder differs from the plugin one:
+    // Nextflow rejects two staged inputs that share a filename.
+    cache   = file(params.vep_database ? "${projectDir}/assets/NO_CACHE" : params.vep_cache,
+                   checkIfExists: must_exist && !params.vep_database)
     plugins = file(params.vep_plugin_data ?: "${projectDir}/assets/NO_FILE", checkIfExists: must_exist)
 
     NORMALISE(vcf_ch)

pipeline/modules/vep.nf

@@ -17,10 +17,14 @@ process VEP {
         "--plugin CADD,snv=${plugin_data}/${params.cadd_snv},indels=${plugin_data}/${params.cadd_indels}",
         "--plugin AlphaMissense,file=${plugin_data}/${params.alphamissense}",
     ].join(' ')
+    // --everything needs the cache (it implies --af_gnomade and friends); the database offers a
+    // smaller set, but still the consequence, gene, HGVS and ClinVar fields the loader stores.
+    def source = params.vep_database ? "--database" : "--cache --offline --dir_cache ${cache}"
+    def fields = params.vep_database ? "--symbol --hgvs --canonical --biotype --variant_class --check_existing" : "--everything"
     """
     vep -i $vcf -o ${vcf.simpleName}.vep.tsv --tab \\
-        --assembly ${params.assembly} --cache --offline --dir_cache ${cache} \\
-        --everything --pick ${plugins} \\
+        --assembly ${params.assembly} ${source} \\
+        ${fields} --pick ${plugins} \\
         --stats_file ${vcf.simpleName}.vep_summary.html --fork ${task.cpus}
     """
 

pipeline/nextflow.config

@@ -5,6 +5,9 @@ params {
     assembly        = "GRCh38"
     vep_cache       = "${projectDir}/cache/vep"   // INSTALL.pl -a cf -s homo_sapiens -y GRCh38 -c <dir>
     vep_plugin_data = null        // CADD + AlphaMissense modules and data; plugins skipped when null
+    // Query Ensembl's public database instead of a local cache: no 25 GB download, but slow
+    // per variant and fewer fields. Fine for a handful of variants, wrong for a whole genome.
+    vep_database    = (System.getenv('VEP_DATABASE') ?: 'false').toBoolean()
     cadd_snv        = "whole_genome_SNVs.tsv.gz"
     cadd_indels     = "gnomad.genomes.r4.0.indel.tsv.gz"
     alphamissense   = "AlphaMissense_hg38.tsv.gz"