feat: implement VariantMapper (#12) (#13)

Cargo.toml

@@ -5,14 +5,17 @@ edition = "2021"
 
 [dependencies]
 anyhow = "1.0.69"
+base16ct = "0.2.0"
 bio = "1.1.0"
 chrono = "0.4.23"
 enum-map = "2.4.2"
 flate2 = "1.0.25"
 lazy_static = "1.4.0"
 linked-hash-map = "0.5.6"
 log = "0.4.17"
+md-5 = "0.10.5"
 nom = "7.1.3"
+phf = { version = "0.11.1", features = ["macros"] }
 postgres = { version = "0.19.4", features = ["with-chrono-0_4"] }
 pretty_assertions = "1.3.0"
 regex = "1.7.1"
@@ -21,5 +24,7 @@ serde = { version = "1.0.152", features = ["derive"] }
 serde_json = "1.0.93"
 
 [dev-dependencies]
+csv = "1.2.0"
 env_logger = "0.10.0"
+serde = { version = "1.0.152", features = ["derive"] }
 test-log = "0.2.11"

src/data/interface.rs

@@ -232,6 +232,12 @@ pub trait Provider {
         end: Option<usize>,
     ) -> Result<String, anyhow::Error>;
 
+    /// Returns a list of protein accessions for a given sequence.
+    ///
+    /// The list is guaranteed to contain at least one element with the MD5-based accession
+    /// (MD5_01234abc..def56789) at the end of the list.
+    fn get_acs_for_protein_seq(&self, seq: &str) -> Result<Vec<String>, anyhow::Error>;
+
     /// Return a list of transcripts that are similar to the given transcript, with relevant
     /// similarity criteria.
     ///

src/data/uta.rs

@@ -7,6 +7,7 @@ use postgres::{Client, NoTls, Row};
 use std::fmt::Debug;
 use std::sync::Mutex;
 
+use crate::sequences::seq_md5;
 use crate::static_data::{Assembly, ASSEMBLY_INFOS};
 
 use crate::data::{
@@ -276,6 +277,23 @@ impl ProviderInterface for Provider {
         Ok(seq[begin..end].into())
     }
 
+    fn get_acs_for_protein_seq(&self, seq: &str) -> Result<Vec<String>, anyhow::Error> {
+        let md5 = seq_md5(seq, true)?;
+        let sql = format!(
+            "SELECT ac FROM {}.seq_anno WHERE seq_id = $1",
+            self.config.db_schema
+        );
+
+        let mut result = Vec::new();
+        for row in self.conn.lock().unwrap().query(&sql, &[&md5])? {
+            result.push(row.get(0));
+        }
+
+        // Add sentinel sequence.
+        result.push(format!("MD5_{}", &md5));
+        Ok(result)
+    }
+
     fn get_similar_transcripts(
         &self,
         tx_ac: &str,

src/data/uta_sr.rs

@@ -124,6 +124,10 @@ impl ProviderInterface for Provider {
         self.seqrepo.fetch_sequence_part(&aos, begin, end)
     }
 
+    fn get_acs_for_protein_seq(&self, seq: &str) -> Result<Vec<String>, anyhow::Error> {
+        self.inner.get_acs_for_protein_seq(seq)
+    }
+
     fn get_similar_transcripts(
         &self,
         tx_ac: &str,

src/lib.rs

@@ -2,6 +2,6 @@ pub mod data;
 pub mod mapper;
 pub mod normalizer;
 pub mod parser;
+pub(crate) mod sequences;
 pub mod static_data;
-pub(crate) mod utils;
 pub mod validator;

src/mapper/alignment.rs

@@ -484,15 +484,12 @@ impl Mapper {
 
 #[cfg(test)]
 mod test {
-    use std::{rc::Rc, str::FromStr};
+    use std::str::FromStr;
 
     use pretty_assertions::assert_eq;
 
     use crate::{
-        data::{
-            interface::{Provider as Interface, TxExonsRecord},
-            uta::{Config, Provider},
-        },
+        data::{interface::TxExonsRecord, uta_sr::test_helpers::build_provider},
         parser::{CdsFrom, CdsInterval, CdsPos, GenomeInterval, Mu, TxInterval, TxPos},
     };
 
@@ -563,30 +560,19 @@ mod test {
         assert_eq!(none_if_default(-1i32), Some(-1i32));
     }
 
-    fn get_config() -> Config {
-        Config {
-            db_url: std::env::var("TEST_UTA_DATABASE_URL")
-                .expect("Environment variable TEST_UTA_DATABASE_URL undefined!"),
-            db_schema: std::env::var("TEST_UTA_DATABASE_SCHEMA")
-                .expect("Environment variable TEST_UTA_DATABASE_SCHEMA undefined!"),
-        }
-    }
-
     #[test]
     fn construction() -> Result<(), anyhow::Error> {
-        let config = get_config();
-        let provider = Provider::with_config(&config)?;
+        let provider = build_provider()?;
 
-        assert_eq!(provider.data_version(), config.db_schema);
+        assert_eq!(provider.data_version(), "uta_20210129");
         assert_eq!(provider.schema_version(), "1.1");
 
         Ok(())
     }
 
     #[test]
     fn failures() -> Result<(), anyhow::Error> {
-        let config = get_config();
-        let provider = Rc::new(Provider::with_config(&config)?);
+        let provider = build_provider()?;
 
         // unknown sequences
 
@@ -640,8 +626,7 @@ mod test {
         alt_ac: &str,
         cases: &Vec<(GenomeInterval, TxInterval, CdsInterval)>,
     ) -> Result<(), anyhow::Error> {
-        let config = get_config();
-        let provider = Rc::new(Provider::with_config(&config)?);
+        let provider = build_provider()?;
         let mapper = Mapper::new(provider, tx_ac, alt_ac, "splign")?;
 
         for (g_interval, n_interval, c_interval) in cases {