Merge pull request #53 from biothings/publication-refactor

Publication refactor
biothings · Aug 11, 2023 · d6723f5 · d6723f5
2 parents 85bba2a + ae3a5f4
commit d6723f5
Show file tree

Hide file tree

Showing 5 changed files with 103 additions and 27 deletions.
diff --git a/__test__/base_transformer.test.ts b/__test__/base_transformer.test.ts
@@ -57,41 +57,41 @@ describe("test base transformer", () => {
     test("Test _updatePublications function if pubmed id is prefixed", () => {
         const tf = new base_tf(input, {});
         const fake = {
-            pubmed: "PMID:1233"
+            ref_pmid: "PMID:1233"
         }
         const res = tf._updatePublications(fake);
-        expect(res).not.toHaveProperty('pubmed');
+        expect(res).not.toHaveProperty('ref_pmid');
         expect(res.publications).toEqual(["PMID:1233"]);
     })
 
     test("Test _updatePublications function if pubmed id is NOT prefixed", () => {
         const tf = new base_tf(input, {});
         const fake = {
-            pubmed: 1233
+            ref_pmid: 1233
         }
         const res = tf._updatePublications(fake);
-        expect(res).not.toHaveProperty('pubmed');
+        expect(res).not.toHaveProperty('ref_pmid');
         expect(res.publications).toEqual(["PMID:1233"])
     })
 
     test("Test _updatePublications function if pmc id is prefixed", () => {
         const tf = new base_tf(input, {});
         const fake = {
-            pmc: "PMC:1233"
+            ref_pmcid: "PMCID:1233"
         }
         const res = tf._updatePublications(fake);
-        expect(res).not.toHaveProperty('pmc');
-        expect(res.publications).toEqual(["PMC:1233"]);
+        expect(res).not.toHaveProperty('ref_pmcid');
+        expect(res.publications).toEqual(["PMCID:1233"]);
     })
 
     test("Test _updatePublications function if pmc id is NOT prefixed", () => {
         const tf = new base_tf(input, {});
         const fake = {
-            pmc: 123
+            ref_pmcid: 123
         }
         const res = tf._updatePublications(fake);
-        expect(res).not.toHaveProperty('pmc');
-        expect(res.publications).toEqual(["PMC:123"])
+        expect(res).not.toHaveProperty('ref_pmcid');
+        expect(res.publications).toEqual(["PMCID:123"])
     })
 
     test("Test extractObjectIDs function if output id type not in result", () => {

diff --git a/__test__/biothings_transformer.test.ts b/__test__/biothings_transformer.test.ts
@@ -58,7 +58,7 @@ describe("test biothings transformer", () => {
             let tf = new biothings_tf(input, {});
             let res = await tf.transform();
             expect(res).toHaveLength(27);
-            expect(res[0]).not.toHaveProperty('pubmed');
+            expect(res[0]).not.toHaveProperty('ref_pmid');
             expect(res[0]).toHaveProperty('publications', ["PMID:21873635"]);
         })
     })

diff --git a/src/transformers/biolink_transformer.ts b/src/transformers/biolink_transformer.ts
@@ -15,12 +15,18 @@ export default class BiolinkTransformer extends BaseTransformer {
                         rec['object'][prefix] = rec.object.id;
                     }
                 }
-                if (rec.publications === undefined || rec.publications.length === 0 || !(rec.publications[0]['id'].startsWith("PMID"))) {
+                if (rec.publications === undefined || rec.publications.length === 0) {
                     delete rec.publications
                 } else {
-                    rec.publications = rec.publications.map(pub => {
-                        return { "id": pub.id.split(':').slice(-1)[0] }
-                    })
+                    const oldPublications = rec.publications;
+                    rec.publications = [];
+                    for (let oldPub of oldPublications) {
+                        if (!oldPub?.id?.startsWith?.("PMID:")) {
+                            continue;
+                        }
+
+                        rec.publications.push({ id: oldPub.id.split(':').slice(-1)[0] });
+                    }
                 }
                 if (!("provided_by" in rec)) {
                     delete rec.provided_by

diff --git a/src/transformers/ctd_transformer.ts b/src/transformers/ctd_transformer.ts
@@ -7,9 +7,15 @@ export default class CTDTransformer extends BaseTransformer {
             if (typeof item.PubMedIDs === "string") {
                 item.PubMedIDs = item.PubMedIDs.split('|');
             }
+            if (typeof item.PubMedIds === "string") {
+                item.PubMedIds = item.PubMedIds.split('|');
+            }
             if (typeof item.DiseaseID === "string") {
                 item.DiseaseID = item.DiseaseID.split(':').slice(-1)[0];
             }
+            if (typeof item.DiseaseId === "string") {
+                item.DiseaseId = item.DiseaseId.split(':').slice(-1)[0];
+            }
             return item;
         });
         return { data: res };

diff --git a/src/transformers/transformer.ts b/src/transformers/transformer.ts
@@ -54,20 +54,84 @@ export default class BaseTransformer {
     }
 
     _updatePublications(mappedResponse: any) {
-        if ("pubmed" in mappedResponse) {
-            mappedResponse.pubmed = toArray(mappedResponse.pubmed);
-            mappedResponse.publications = mappedResponse.pubmed.map(item =>
-                typeof item === "string" && item.toUpperCase().startsWith("PMID:") ? item.toUpperCase() : "PMID:" + item,
-            );
-            delete mappedResponse.pubmed;
+        if (!Array.isArray(mappedResponse.publications)) {
+            mappedResponse.publications = [];
         }
-        if ("pmc" in mappedResponse) {
-            mappedResponse.pmc = toArray(mappedResponse.pmc);
-            mappedResponse.publications = mappedResponse.pmc.map(item =>
-                typeof item === "string" && item.toUpperCase().startsWith("PMC:") ? item.toUpperCase() : "PMC:" + item,
-            );
-            delete mappedResponse.pmc;
+
+        const publicationTypes = [
+            {prop: "ref_pmid", prefix: "PMID:", urls: ["http://www.ncbi.nlm.nih.gov/pubmed/", "http://europepmc.org/abstract/MED/", "https://www.ncbi.nlm.nih.gov/pubmed/"]},
+            {prop: "ref_pmcid", prefix: "PMCID:", urls: ["http://www.ncbi.nlm.nih.gov/pmc/articles/", "http://europepmc.org/articles/"]},
+            {prop: "ref_clinicaltrials", prefix: "clinicaltrials:", urls: ["https://clinicaltrials.gov/ct2/show/", "https://www.clinicaltrials.gov/ct2/show/"]},
+            {prop: "ref_doi", prefix: "doi:", urls: ["https://doi.org/", "http://www.nejm.org/doi/full/", "https://www.tandfonline.com/doi/abs/", "http://onlinelibrary.wiley.com/doi/"]},
+            {prop: "ref_isbn", prefix: "isbn:", urls: ["https://www.isbn-international.org/identifier/"]}
+        ]
+
+        // handle URLs (which could be CURIEs)
+        if ("ref_url" in mappedResponse) {
+            for (let publication of toArray(mappedResponse.ref_url)) {
+                if (typeof publication !== "string" || publication.length === 0) {
+                    continue;
+                }
+
+                let isCurie = false;
+                for (let publicationType of publicationTypes) {
+                    for (let url of publicationType.urls) {
+                        if (publication.startsWith(url)) {
+                            isCurie = true;
+
+                            if (!mappedResponse[publicationType.prop]) {
+                                mappedResponse[publicationType.prop] = [];
+                            }
+                            else if (!Array.isArray(mappedResponse[publicationType.prop])) {
+                                mappedResponse[publicationType.prop] = toArray(mappedResponse[publicationType.prop]);
+                            }
+
+                            mappedResponse[publicationType.prop].push(publication.slice(url.length));
+
+                            break;
+                        }
+                    }
+
+                    if (isCurie) {
+                        break;
+                    }
+                }
+
+                if (!isCurie) {
+                    mappedResponse.publications.push(publication);
+                }
+            }
         }
+        delete mappedResponse.ref_url;
+
+        for (let publicationType of publicationTypes) {
+            if (publicationType.prop in mappedResponse) {
+                for (let publication of toArray(mappedResponse[publicationType.prop])) {
+                    // handle numbers
+                    if (typeof publication === "number") {
+                        publication = publication.toString();
+                    }
+
+                    if (typeof publication !== "string" || publication.length === 0) {
+                        continue;
+                    }
+
+                    if (publication.toUpperCase().startsWith(publicationType.prefix.toUpperCase())) {
+                        mappedResponse.publications.push(publicationType.prefix + publication.slice(publicationType.prefix.length));
+                    }
+                    else {
+                        mappedResponse.publications.push(publicationType.prefix + publication);
+                    }
+                }
+
+                delete mappedResponse[publicationType.prop];
+            }
+        }
+
+        if (mappedResponse.publications.length === 0) {
+            delete mappedResponse.publications;
+        }
+
         return mappedResponse;
     }