feat: add support for functional elements (#14)

Cargo.toml

@@ -13,7 +13,7 @@ name = "scarus"
 path = "src/main.rs"
 
 [dependencies]
-annonars = "0.24.5"
+annonars = "0.25.0"
 anyhow = "1.0"
 bio = "1.4"
 biocommons-bioutils = "0.1.4"

src/strucvars/eval/del/section2.rs

@@ -1,6 +1,6 @@
 //! Implementation of evaluation of copy number loss section 2.
 
-use annonars::{clinvar_minimal::pbs::ClinicalSignificance, clinvar_minimal::pbs::ReviewStatus};
+use annonars::pbs::annonars::clinvar::v1::minimal::{ClinicalSignificance, ReviewStatus};
 use bio::bio_types::genome::AbstractInterval;
 use hgvs::data::interface::{Provider, TxExonsRecord, TxInfoRecord};
 

src/strucvars/eval/mod.rs

@@ -7,6 +7,7 @@ pub mod result;
 
 use std::{path::Path, sync::Arc};
 
+use annonars::pbs::annonars::clinvar::v1::minimal::Record;
 use hgvs::data::interface::Provider;
 use itertools::Itertools as _;
 use prost::Message as _;
@@ -102,6 +103,7 @@ impl Evaluator {
             path_clinvar_minimal,
             "clinvar",
             "meta",
+            "clinvar_by_accession",
         )
         .map_err(|e| anyhow::anyhow!("failed to open 'minimal' ClinVar RocksDB: {}", e))?;
 
@@ -187,7 +189,7 @@ impl Evaluator {
         chrom: &str,
         start: u32,
         stop: u32,
-    ) -> Result<Vec<annonars::clinvar_minimal::pbs::Record>, anyhow::Error> {
+    ) -> Result<Vec<Record>, anyhow::Error> {
         tracing::trace!("starting clinvar query {}:{}-{}", &chrom, start, stop);
         let start = start as i32;
         let stop = stop as i32;
@@ -227,7 +229,7 @@ impl Evaluator {
                 }
 
                 // Otherwise, decode the value from the database.
-                let record = annonars::clinvar_minimal::pbs::Record::decode(value)?;
+                let record = Record::decode(value)?;
                 result.push(record);
 
                 // Proceed to the next database row.

tests/data/strucvars/hi_ts/bootstrap.sh

@@ -32,12 +32,16 @@ RERE
 LINCADL
 "
 
+wget -O /tmp/cdot-0.2.22.GCF_000001405.25_GRCh37.p13_genomic.105.20201022.gff.json.gz \
+    https://github.com/SACGF/cdot/releases/download/data_v0.2.22/cdot-0.2.22.GCF_000001405.25_GRCh37.p13_genomic.105.20201022.gff.json.gz
+
+rm -rf $SCRIPTPATH/txs_example_hi.bin.zst
 mehari db create \
     -vvv \
     --genome-release grch37 \
     --path-mane-txs-tsv $SCRIPTPATH/tx-mane.tsv \
     --path-out $SCRIPTPATH/txs_example_hi.bin.zst \
-    --path-cdot-json ~/mehari-data-tx/cdot-refseq-vep110~0+0.2.21.json.gz \
+    --path-cdot-json /tmp/cdot-0.2.22.GCF_000001405.25_GRCh37.p13_genomic.105.20201022.gff.json.gz \
     --path-seqrepo-instance ~/mehari-data-tx/seqrepo/main \
     $(for gene_symbol in $gene_symbols; do echo --gene-symbols $gene_symbol; done)