feat(thesesfr): added PHD_THESIS match

ezpaarse-project · Jul 17, 2024 · 2998e80 · 2998e80
1 parent 1e9a588
commit 2998e80
Show file tree

Hide file tree

Showing 2 changed files with 21 additions and 2 deletions.
diff --git a/thesesfr/parser.js b/thesesfr/parser.js
@@ -21,7 +21,24 @@ module.exports = new Parser(function analyseEC(parsedUrl, ec) {
 
   let match;
 
-  if ((match = /^\/([0-9]{8}[0-9X])$/i.exec(path)) !== null) {
+  if ((match = /^\/api\/v1\/document\/([0-9a-z]+)$/i.exec(path)) !== null) {
+    // /api/v1/document/2019LYSE2053
+    result.rtype = 'PHD_THESIS';
+    result.unitid = match[1];
+    switch (ec.status) {
+    case 200:
+      result.mime = 'PDF';
+      break;
+    case 302:
+      result.mime = 'HTML';
+      break;
+
+    default:
+      result.mime = 'MISC';
+      break;
+    }
+
+  } else if ((match = /^\/([0-9]{8}[0-9X])$/i.exec(path)) !== null) {
     // /258987731
     result.rtype = 'BIO';
     result.mime = 'HTML';

diff --git a/thesesfr/test/thesesfr.2024-01-09.csv b/thesesfr/test/thesesfr.2024-01-09.csv
@@ -1,4 +1,6 @@
 out-unitid;out-rtype;out-mime;out-ppn;out-publication_date;out-institution_code;in-url
 2023UPASP097;ABS;HTML;;2023;UPAS;https://beta.theses.fr/2023UPASP097
 s366354;ABS;HTML;;;;https://beta.theses.fr/s366354
-258987731;BIO;HTML;258987731;;;https://beta.theses.fr/258987731
+258987731;BIO;HTML;258987731;;;https://beta.theses.fr/258987731
+2019LYSE2053;PHD_THESIS;MISC;;;;https://theses.fr/api/v1/document/2019LYSE2053
+2010AIX22039;PHD_THESIS;MISC;;;;https://theses.fr/api/v1/document/2010AIX22039