update and add more sssom mappings

EBISPOT · Sep 11, 2024 · 37effaa · 37effaa
1 parent be7dfc3
commit 37effaa
Show file tree

Hide file tree

Showing 8 changed files with 19 additions and 14 deletions.
diff --git a/00_fetch_data/sssom/fetch.sh b/00_fetch_data/sssom/fetch.sh
@@ -0,0 +1,8 @@
+#!/bin/bash
+
+rm -f *.sssom.tsv.gz
+
+curl -L https://data.monarchinitiative.org/mappings/latest/upheno_custom.sssom.tsv | gzip > upheno_custom.sssom.tsv.gz
+curl -L https://raw.githubusercontent.com/mapping-commons/mh_mapping_initiative/master/mappings/mp_hp_mgi_all.sssom.tsv | gzip > mp_hp_mgi_all.sssom.tsv.gz
+curl -L https://raw.githubusercontent.com/obophenotype/bio-attribute-ontology/master/src/mappings/oba-efo.sssom.tsv | gzip > oba-efo.sssom.tsv.gz
+curl -L https://raw.githubusercontent.com/obophenotype/bio-attribute-ontology/master/src/mappings/oba-vt.sssom.tsv | gzip > oba-vt.sssom.tsv.gz
diff --git a/01_ingest/grebi_ingest_sssom/src/main.rs b/01_ingest/grebi_ingest_sssom/src/main.rs
@@ -52,14 +52,15 @@ fn main() {
         return line.trim_start_matches("#").to_string();
     }).collect::<Vec<String>>().join("\n");
 
-    let yaml_header:serde_yaml::Value = serde_yaml::from_str::<serde_yaml::Value>(yaml.as_str()).unwrap();
-
-    let yaml_header_curie_map = yaml_header.get("curie_map").unwrap().as_mapping().unwrap();
-
+
     let expand:PrefixMap = {
         let mut builder = PrefixMapBuilder::new();
-        for (k, v) in yaml_header_curie_map {
-            builder.add_mapping(k.as_str().unwrap().to_string() + ":", v.as_str().unwrap().to_string());
+        let yaml_header:serde_yaml::Value = serde_yaml::from_str::<serde_yaml::Value>(yaml.as_str()).unwrap();
+        let yaml_header_curie_map = yaml_header.get("curie_map");
+        if yaml_header_curie_map.is_some() {
+            for (k, v) in yaml_header_curie_map.unwrap().as_mapping().unwrap() {
+                builder.add_mapping(k.as_str().unwrap().to_string() + ":", v.as_str().unwrap().to_string());
+            }
         }
         builder.build()
     };

diff --git a/configs/datasource_configs/mp_hp.json → configs/datasource_configs/sssom.json b/configs/datasource_configs/mp_hp.json → configs/datasource_configs/sssom.json
@@ -1,10 +1,10 @@
 {
-    "name": "MP_HP",
+    "name": "SSSOM",
     "enabled": true,
     "ingests": [
         {
             "ingest_files": [
-                "./00_fetch_data/sssom/mphp.sssom.tsv"
+                "./00_fetch_data/sssom/*.tsv.gz"
             ],
             "ingest_script": "./target/release/grebi_ingest_sssom",
             "ingest_args": []

diff --git a/configs/subgraph_configs/ebi_full_monarch.json b/configs/subgraph_configs/ebi_full_monarch.json
@@ -19,7 +19,6 @@
         "skos:exactMatch",
         "ncit:P368",
         "ncit:C98965",
-        "cheminf:000407",
         "dcterms:identifier",
         "oboinowl:hasAlternativeId",
         "semapv:crossSpeciesExactMatch"
@@ -65,7 +64,7 @@
         "./configs/datasource_configs/gwas.json",
         "./configs/datasource_configs/hgnc.json",
         "./configs/datasource_configs/impc.json",
-        "./configs/datasource_configs/mp_hp.json",
+        "./configs/datasource_configs/sssom.json",
         "./configs/datasource_configs/ols.json",
         "./configs/datasource_configs/reactome.json",
         "./configs/datasource_configs/ubergraph.json",

diff --git a/configs/subgraph_configs/hett.json b/configs/subgraph_configs/hett.json
@@ -20,7 +20,6 @@
         "skos:exactMatch",
         "ncit:P368",
         "ncit:C98965",
-	"cheminf:000407",
 	"dcterms:identifier",
         "oboinowl:hasAlternativeId"
     ],

diff --git a/configs/subgraph_configs/hra_kg.json b/configs/subgraph_configs/hra_kg.json
@@ -20,7 +20,6 @@
         "skos:exactMatch",
         "ncit:P368",
         "ncit:C98965",
-	"cheminf:000407",
 	"dcterms:identifier",
         "oboinowl:hasAlternativeId"
     ],

diff --git a/configs/subgraph_configs/monarch.json b/configs/subgraph_configs/monarch.json
@@ -20,7 +20,6 @@
         "skos:exactMatch",
         "ncit:P368",
         "ncit:C98965",
-	"cheminf:000407",
 	"dcterms:identifier",
         "oboinowl:hasAlternativeId"
     ],

diff --git a/grebi_shared/src/prefix_map.rs b/grebi_shared/src/prefix_map.rs
@@ -129,7 +129,7 @@ Node {
 #[inline(always)]
 fn reprefix_impl<'a>(subject:&[u8], buf:&[u8]) -> Option<Vec<u8>> {
 
-    if subject.len() == 0 {
+    if subject.len() == 0 || buf.len() == 0 {
         return None;
     }