Merge pull request #514 from thoth-pub/feature/513_further_doi_regex

Allow angle brackets in DOI validation (#513)

CHANGELOG.md

@@ -10,6 +10,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - Upgrade `actix-session` to v0.8.0
   - Upgrade `chrono` to v0.4.31
 
+### Fixed
+  - [#513](https://github.com/thoth-pub/thoth/issues/513) - Expand DOI regex to include angle brackets
+
 ## [[0.11.6]](https://github.com/thoth-pub/thoth/releases/tag/v0.11.6) - 2023-09-08
 ### Security
   - Upgrade `chrono` to v0.4.30

thoth-api/migrations/v0.11.7/down.sql

@@ -0,0 +1,11 @@
+ALTER TABLE work DROP CONSTRAINT work_doi_check;
+ALTER TABLE work ADD CONSTRAINT work_doi_check
+  CHECK (doi ~ '^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\[\]\/:a-zA-Z0-9]+$');
+
+ALTER TABLE reference DROP CONSTRAINT reference_doi_check;
+ALTER TABLE reference ADD CONSTRAINT reference_doi_check
+  CHECK (doi ~ '^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\[\]\/:a-zA-Z0-9]+$');
+
+ALTER TABLE institution DROP CONSTRAINT institution_institution_doi_check;
+ALTER TABLE institution ADD CONSTRAINT institution_institution_doi_check
+  CHECK (institution_doi ~ '^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\[\]\/:a-zA-Z0-9]+$');

thoth-api/migrations/v0.11.7/up.sql

@@ -0,0 +1,11 @@
+ALTER TABLE work DROP CONSTRAINT work_doi_check;
+ALTER TABLE work ADD CONSTRAINT work_doi_check
+  CHECK (doi ~ '^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\[\]<>\/:a-zA-Z0-9]+$');
+
+ALTER TABLE reference DROP CONSTRAINT reference_doi_check;
+ALTER TABLE reference ADD CONSTRAINT reference_doi_check
+  CHECK (doi ~ '^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\[\]<>\/:a-zA-Z0-9]+$');
+
+ALTER TABLE institution DROP CONSTRAINT institution_institution_doi_check;
+ALTER TABLE institution ADD CONSTRAINT institution_institution_doi_check
+  CHECK (institution_doi ~ '^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\[\]<>\/:a-zA-Z0-9]+$');

thoth-api/src/model/mod.rs

@@ -46,7 +46,7 @@ pub enum WeightUnit {
     feature = "backend",
     derive(DieselNewType, juniper::GraphQLScalarValue),
     graphql(
-        description = r#"Digital Object Identifier. Expressed as `^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\[\]\/:a-zA-Z0-9]+$`"#
+        description = r#"Digital Object Identifier. Expressed as `^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\[\]<>\/:a-zA-Z0-9]+$`"#
     )
 )]
 #[derive(Debug, Clone, Default, Serialize, Deserialize, PartialEq, Eq)]
@@ -144,7 +144,7 @@ impl FromStr for Doi {
             // and captures the identifier segment starting with the "10." directory indicator
             // Corresponds to database constraints although regex syntax differs slightly
             // (e.g. `;()/` do not need to be escaped here)
-            r"^(?i:(?:https?://)?(?:www\.)?(?:dx\.)?doi\.org/)?(10\.\d{4,9}/[-._;()\[\]/:a-zA-Z0-9]+$)").unwrap();
+            r"^(?i:(?:https?://)?(?:www\.)?(?:dx\.)?doi\.org/)?(10\.\d{4,9}/[-._;()\[\]<>/:a-zA-Z0-9]+$)").unwrap();
         }
         if input.is_empty() {
             Err(ThothError::DoiEmptyError)
@@ -703,6 +703,10 @@ fn test_doi_fromstr() {
     assert!(Doi::from_str("https://doi-org/10.12345/Test-Suffix.01").is_err());
     assert!(Doi::from_str("10.https://doi.org/12345/Test-Suffix.01").is_err());
     assert!(Doi::from_str("http://dx.doi.org/10.2990/1471-5457(2005)24[2:tmpwac]2.0.co;2").is_ok());
+    assert!(Doi::from_str(
+        "https://doi.org/10.1002/(SICI)1098-2736(199908)36:6<637::AID-TEA4>3.0.CO;2-9"
+    )
+    .is_ok());
 }
 
 #[test]