Add custom Doi and Orcid structs and use them to warn of parsing errors as user types

thoth-api/Cargo.toml

@@ -22,6 +22,7 @@ chrono = { version = "0.4", features = ["serde"] }
 csv = { version = "1.1.6", optional = true }
 diesel = { version = "1.4.0", features = ["postgres", "uuidv07", "chrono", "r2d2", "64-column-tables", "serde_json"], optional = true }
 diesel-derive-enum = { version = "1.1.0", features = ["postgres"], optional = true }
+diesel-derive-newtype = "0.1"
 diesel_migrations = { version = "1.4.0", features = ["postgres"], optional = true }
 dotenv = "0.9.0"
 failure = "0.1.6"

thoth-api/migrations/0.4.1/down.sql

@@ -0,0 +1,16 @@
+-- Reinstate earlier versions of ORCID and DOI validation
+
+ALTER TABLE contributor
+    DROP CONSTRAINT contributor_orcid_check,
+    ADD CONSTRAINT contributor_orcid_check
+    CHECK (orcid ~* '0000-000(1-[5-9]|2-[0-9]|3-[0-4])\d{3}-\d{3}[\dX]');
+
+ALTER TABLE work
+    DROP CONSTRAINT work_doi_check,
+    ADD CONSTRAINT work_doi_check
+    CHECK (doi ~* 'https:\/\/doi.org\/10.\d{4,9}\/[-._\;\(\)\/:a-zA-Z0-9]+$');
+
+ALTER TABLE funder
+    DROP CONSTRAINT funder_funder_doi_check,
+    ADD CONSTRAINT funder_funder_doi_check
+    CHECK (funder_doi ~* 'https:\/\/doi.org\/10.\d{4,9}\/[-._\;\(\)\/:a-zA-Z0-9]+$');

thoth-api/migrations/0.4.1/up.sql

@@ -0,0 +1,21 @@
+-- Improve validation of ORCID identifiers (include protocol/resource name, make case-sensitive)
+-- Should be kept in line with Orcid::FromStr, although regex syntax differs slightly
+
+ALTER TABLE contributor
+    DROP CONSTRAINT contributor_orcid_check,
+    ADD CONSTRAINT contributor_orcid_check
+    CHECK (orcid ~ '^https:\/\/orcid\.org\/0000-000(1-[5-9]|2-[0-9]|3-[0-4])\d{3}-\d{3}[\dX]$');
+
+-- Improve validation of DOI identifiers (add line start marker, escape periods, make case-sensitive)
+-- Should be kept in line with Orcid::FromStr, although regex syntax differs slightly
+-- (e.g. `;()/` need to be escaped here but not in Orcid::FromStr)
+
+ALTER TABLE work
+    DROP CONSTRAINT work_doi_check,
+    ADD CONSTRAINT work_doi_check
+    CHECK (doi ~ '^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\/:a-zA-Z0-9]+$');
+
+ALTER TABLE funder
+    DROP CONSTRAINT funder_funder_doi_check,
+    ADD CONSTRAINT funder_funder_doi_check
+    CHECK (funder_doi ~ '^https:\/\/doi\.org\/10\.\d{4,9}\/[-._\;\(\)\/:a-zA-Z0-9]+$');

thoth-api/src/account/model.rs

@@ -1,10 +1,9 @@
-use chrono::DateTime;
-use chrono::Utc;
 use serde::{Deserialize, Serialize};
 use uuid::Uuid;
 
 use crate::errors::ThothError;
 use crate::errors::ThothResult;
+use crate::model::Timestamp;
 #[cfg(feature = "backend")]
 use crate::schema::account;
 #[cfg(feature = "backend")]
@@ -22,8 +21,8 @@ pub struct Account {
     pub is_superuser: bool,
     pub is_bot: bool,
     pub is_active: bool,
-    pub created_at: DateTime<Utc>,
-    pub updated_at: DateTime<Utc>,
+    pub created_at: Timestamp,
+    pub updated_at: Timestamp,
     pub token: Option<String>,
 }
 
@@ -54,8 +53,8 @@ pub struct PublisherAccount {
     pub account_id: Uuid,
     pub publisher_id: Uuid,
     pub is_admin: bool,
-    pub created_at: DateTime<Utc>,
-    pub updated_at: DateTime<Utc>,
+    pub created_at: Timestamp,
+    pub updated_at: Timestamp,
 }
 
 #[cfg_attr(feature = "backend", derive(Insertable))]
@@ -99,8 +98,8 @@ pub struct AccountDetails {
     pub surname: String,
     pub email: String,
     pub token: Option<String>,
-    pub created_at: DateTime<Utc>,
-    pub updated_at: DateTime<Utc>,
+    pub created_at: Timestamp,
+    pub updated_at: Timestamp,
     pub resource_access: AccountAccess,
 }
 

thoth-api/src/contribution/model.rs

@@ -1,10 +1,9 @@
-use chrono::DateTime;
-use chrono::Utc;
 use serde::{Deserialize, Serialize};
 use strum::Display;
 use strum::EnumString;
 use uuid::Uuid;
 
+use crate::model::Timestamp;
 #[cfg(feature = "backend")]
 use crate::schema::contribution;
 #[cfg(feature = "backend")]
@@ -54,7 +53,7 @@ pub enum ContributionField {
 }
 
 #[cfg_attr(feature = "backend", derive(Queryable))]
-#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
+#[derive(Debug, Clone, Default, Serialize, Deserialize, PartialEq)]
 #[serde(rename_all = "camelCase")]
 pub struct Contribution {
     pub contribution_id: Uuid,
@@ -64,8 +63,8 @@ pub struct Contribution {
     pub main_contribution: bool,
     pub biography: Option<String>,
     pub institution: Option<String>,
-    pub created_at: DateTime<Utc>,
-    pub updated_at: DateTime<Utc>,
+    pub created_at: Timestamp,
+    pub updated_at: Timestamp,
     pub first_name: Option<String>,
     pub last_name: String,
     pub full_name: String,
@@ -113,7 +112,7 @@ pub struct ContributionHistory {
     pub contribution_id: Uuid,
     pub account_id: Uuid,
     pub data: serde_json::Value,
-    pub timestamp: DateTime<Utc>,
+    pub timestamp: Timestamp,
 }
 
 #[cfg_attr(
@@ -133,25 +132,6 @@ impl Default for ContributionType {
     }
 }
 
-impl Default for Contribution {
-    fn default() -> Contribution {
-        Contribution {
-            contribution_id: Default::default(),
-            work_id: Default::default(),
-            contributor_id: Default::default(),
-            contribution_type: Default::default(),
-            main_contribution: Default::default(),
-            biography: None,
-            institution: None,
-            created_at: chrono::TimeZone::timestamp(&Utc, 0, 0),
-            updated_at: chrono::TimeZone::timestamp(&Utc, 0, 0),
-            first_name: None,
-            last_name: Default::default(),
-            full_name: Default::default(),
-        }
-    }
-}
-
 #[test]
 fn test_contributiontype_default() {
     let contributiontype: ContributionType = Default::default();

thoth-api/src/contributor/model.rs

@@ -1,5 +1,3 @@
-use chrono::DateTime;
-use chrono::Utc;
 use serde::Deserialize;
 use serde::Serialize;
 use std::fmt;
@@ -8,6 +6,8 @@ use strum::EnumString;
 use uuid::Uuid;
 
 use crate::graphql::utils::Direction;
+use crate::model::Orcid;
+use crate::model::Timestamp;
 #[cfg(feature = "backend")]
 use crate::schema::contributor;
 #[cfg(feature = "backend")]
@@ -35,17 +35,17 @@ pub enum ContributorField {
 }
 
 #[cfg_attr(feature = "backend", derive(Queryable))]
-#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
+#[derive(Debug, Clone, Default, Serialize, Deserialize, PartialEq)]
 #[serde(rename_all = "camelCase")]
 pub struct Contributor {
     pub contributor_id: Uuid,
     pub first_name: Option<String>,
     pub last_name: String,
     pub full_name: String,
-    pub orcid: Option<String>,
+    pub orcid: Option<Orcid>,
     pub website: Option<String>,
-    pub created_at: DateTime<Utc>,
-    pub updated_at: DateTime<Utc>,
+    pub created_at: Timestamp,
+    pub updated_at: Timestamp,
 }
 
 #[cfg_attr(
@@ -57,7 +57,7 @@ pub struct NewContributor {
     pub first_name: Option<String>,
     pub last_name: String,
     pub full_name: String,
-    pub orcid: Option<String>,
+    pub orcid: Option<Orcid>,
     pub website: Option<String>,
 }
 
@@ -72,7 +72,7 @@ pub struct PatchContributor {
     pub first_name: Option<String>,
     pub last_name: String,
     pub full_name: String,
-    pub orcid: Option<String>,
+    pub orcid: Option<Orcid>,
     pub website: Option<String>,
 }
 
@@ -82,7 +82,7 @@ pub struct ContributorHistory {
     pub contributor_id: Uuid,
     pub account_id: Uuid,
     pub data: serde_json::Value,
-    pub timestamp: DateTime<Utc>,
+    pub timestamp: Timestamp,
 }
 
 #[cfg_attr(
@@ -101,7 +101,7 @@ pub struct NewContributorHistory {
     derive(juniper::GraphQLInputObject),
     graphql(description = "Field and order to use when sorting contributors list")
 )]
-#[derive(Default, Debug, Clone, Serialize, Deserialize, PartialEq)]
+#[derive(Debug, Clone, Default, Serialize, Deserialize, PartialEq)]
 pub struct ContributorOrderBy {
     pub field: ContributorField,
     pub direction: Direction,
@@ -113,21 +113,6 @@ impl Default for ContributorField {
     }
 }
 
-impl Default for Contributor {
-    fn default() -> Contributor {
-        Contributor {
-            contributor_id: Default::default(),
-            first_name: None,
-            last_name: "".to_string(),
-            full_name: "".to_string(),
-            orcid: None,
-            website: None,
-            created_at: chrono::TimeZone::timestamp(&Utc, 0, 0),
-            updated_at: chrono::TimeZone::timestamp(&Utc, 0, 0),
-        }
-    }
-}
-
 impl fmt::Display for Contributor {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         if let Some(orcid) = &self.orcid {

thoth-api/src/errors.rs

@@ -31,6 +31,20 @@ pub enum ThothError {
     CsvError(String),
     #[fail(display = "Could not generate {}: {}", _0, _1)]
     IncompleteMetadataRecord(String, String),
+    #[fail(
+        display = "{} is not a validly formatted ORCID and will not be saved",
+        _0
+    )]
+    OrcidParseError(String),
+    #[fail(
+        display = "{} is not a validly formatted DOI and will not be saved",
+        _0
+    )]
+    DoiParseError(String),
+    #[fail(display = "Cannot parse ORCID: no value provided")]
+    OrcidEmptyError,
+    #[fail(display = "Cannot parse DOI: no value provided")]
+    DoiEmptyError,
 }
 
 impl juniper::IntoFieldError for ThothError {