feat(biblio-ref): fixed errors + added docs + generate tests

services/biblio-ref/Dockerfile

@@ -1,7 +1,7 @@
 # syntax=docker/dockerfile:1.2
 FROM python:3.9-slim-bullseye as build1
 WORKDIR /dvc
-RUN apt update && apt -y install git curl
+RUN apt update && apt -y install git
 RUN pip install dvc[webdav]==3.39.0
 RUN --mount=type=secret,id=webdav_login \
     --mount=type=secret,id=webdav_password \

services/biblio-ref/examples.http

@@ -13,18 +13,24 @@ Content-Type: application/json
 
 [
   {
-    "value": "Zohuri, B. (2019). A Comparison of Molten Salt Reactors to Light Water Reactors: Pros and Cons. In Molten Salt Reactors and Thorium Energy (pp. 81-98). Woodhead Publishing. https://doi.org/10.1016/B978-0-08-102337-2.00006-9"
+    "value": "J. A. TORRES, A. GARCIA-FUENTES, C. SALAZAR, E. CANO & F. VALLE - Caracterizacion de los pinares de Pil1l1s halepellsis Mill. en el sur de la Peninsula Iberica"
   },
   {
-    "value": "Please see https://doi.org/10.1016\/b978-0-323-90638-8.00002-3"
+    "value":"Y. B. LINHART. L. CHAOUNI-BENABDALLAH, J.-M. PARRY & J. D. THOMPSON - Selective herbivory of thyme chemotypes by a mollusk and a grasshopper"
+  },
+  {
+    "value": "works with just a doi : https://doi.org/10.1016\/b978-0-323-90638-8.00002-3"
   },
   {
     "value": "Gerris Caucasicus, Primary Prevention of Cardiovascular Disease with a Mediterranean Diet, 10.1056/nejmoa1200303"
   },
+  {
+    "value": "RETRACTED > Zohuri, B. (2019). A Comparison of Molten Salt Reactors to Light Water Reactors: Pros and Cons. In Molten Salt Reactors and Thorium Energy (pp. 81-98). Woodhead Publishing. https://doi.org/10.1016/B978-0-08-102337-2.00006-9"
+  },
   {
     "value": ["bad","input","type"]
   },
   {
     "value": "nothing here"
   }
-]
+]

services/biblio-ref/tests.hurl

@@ -2,14 +2,20 @@ POST {{host}}/v1/validate?indent=true
 content-type: application/json
 [
   {
-    "value": "Zohuri, B. (2019). A Comparison of Molten Salt Reactors to Light Water Reactors: Pros and Cons. In Molten Salt Reactors and Thorium Energy (pp. 81-98). Woodhead Publishing. https://doi.org/10.1016/B978-0-08-102337-2.00006-9"
+    "value": "J. A. TORRES, A. GARCIA-FUENTES, C. SALAZAR, E. CANO & F. VALLE - Caracterizacion de los pinares de Pil1l1s halepellsis Mill. en el sur de la Peninsula Iberica"
   },
   {
-    "value": "Please see https://doi.org/10.1016\/b978-0-323-90638-8.00002-3"
+    "value":"Y. B. LINHART. L. CHAOUNI-BENABDALLAH, J.-M. PARRY & J. D. THOMPSON - Selective herbivory of thyme chemotypes by a mollusk and a grasshopper"
+  },
+  {
+    "value": "works with just a doi : https://doi.org/10.1016\/b978-0-323-90638-8.00002-3"
   },
   {
     "value": "Gerris Caucasicus, Primary Prevention of Cardiovascular Disease with a Mediterranean Diet, 10.1056/nejmoa1200303"
   },
+  {
+    "value": "RETRACTED > Zohuri, B. (2019). A Comparison of Molten Salt Reactors to Light Water Reactors: Pros and Cons. In Molten Salt Reactors and Thorium Energy (pp. 81-98). Woodhead Publishing. https://doi.org/10.1016/B978-0-08-102337-2.00006-9"
+  },
   {
     "value": ["bad","input","type"]
   },
@@ -18,11 +24,18 @@ content-type: application/json
   }
 ]
 
+
 HTTP 200
 [{
     "value": {
-        "doi": "",
-        "status": "not_found"
+        "doi": "10.3406/ecmed.1999.1878",
+        "status": "found"
+    }
+},
+{
+    "value": {
+        "doi": "10.3406/ecmed.1999.1879",
+        "status": "found"
     }
 },
 {
@@ -37,6 +50,12 @@ HTTP 200
         "status": "retracted"
     }
 },
+{
+    "value": {
+        "doi": "",
+        "status": "not_found"
+    }
+},
 {
     "value": {
         "doi": "",

services/biblio-ref/v1/validate.ini

@@ -3,17 +3,29 @@ mimeType = application/json
 
 post.operationId = post-v1-validate
 post.summary = Valide une référence bibliographique
-post.description = Si un DOI est trouvé dans la référence bibliographique, valide la référence et indique si elle est rétractée
+post.description = Utilise crossref pour valider la référence bibliographique ainsi que de donner le doi si il existe et enfin indique si elle est rétractée.
 post.tags.0 = biblio-ref
 post.requestBody.content.application/json.schema.$ref = #/components/schemas/JSONStream
 post.requestBody.required = true
 post.responses.default.content.application/json.schema.$ref =  #/components/schemas/JSONStream
-post.responses.default.description = Les champs value contiennent un json constitués des champs `is_found` e `is_retracted`. 
+post.responses.default.description = Les champs value contiennent un json constitués des champs `doi` et `status`. Status prend plusieurs valeur `found`si la référence est trouvée et valide, `not_found` si elle est halucinée ou non présente dans Crossref, `retracted` si elle est rétractée. Indique les erreurs possibles `error_service` ou `error_data`.
 post.parameters.0.description = Indenter le JSON résultant
 post.parameters.0.in = query
 post.parameters.0.name = indent
 post.parameters.0.schema.type = boolean
 
+post.requestBody.content.application/json.example.0.id = 1
+post.requestBody.content.application/json.example.0.value = Y. B. LINHART. L. CHAOUNI-BENABDALLAH, J.-M. PARRY & J. D. THOMPSON - Selective herbivory of thyme chemotypes by a mollusk and a grasshopper
+post.requestBody.content.application/json.example.0.id = 2
+post.requestBody.content.application/json.example.0.value = J. A. TORRES, A. GARCIA-FUENTES, C. SALAZAR, E. CANO & F. VALLE, Caracterizacion de los pinares de Pil1l1s halepellsis Mill. en el sur de la Peninsula Iberica
+post.requestBody.content.application/json.example.0.id = 3
+post.requestBody.content.application/json.example.0.value = works with just a doi : https://doi.org/10.1016\/b978-0-323-90638-8.00002-3
+post.requestBody.content.application/json.example.0.id = 4
+post.requestBody.content.application/json.example.0.value = RETRACTED > Zohuri, B. (2019). A Comparison of Molten Salt Reactors to Light Water Reactors: Pros and Cons. In Molten Salt Reactors and Thorium Energy (pp. 81-98). Woodhead Publishing. https://doi.org/10.1016/B978-0-08-102337-2.00006-9
+post.requestBody.content.application/json.example.0.id = 5
+post.requestBody.content.application/json.example.0.value = Gerris Caucasicus, Primary Prevention of Cardiovascular Disease with a Mediterranean Diet, 10.1056/nejmoa1200303
+post.requestBody.content.application/json.example.0.id = 6
+post.requestBody.content.application/json.example.0.value = ["bad","input","type"]
 
 [use]
 plugin = @ezs/spawn

services/biblio-ref/v1/validate.py

@@ -117,25 +117,6 @@ def get_title_authors_doi(message):
         first_author_given = ""
     return {'title': title, 'first_author_given': first_author_given, 'first_author_name': first_author_name, 'doi': doi}
 
-def match_title(title, ref_biblio):
-    """
-    Match the title of the publication with the title of the biblio reference.
-    
-    Args:
-        title (str): The title of the publication.
-        ref_biblio (str): The biblio reference.
-    
-    Returns:
-        bool: True if the title of the publication matches the title of the biblio reference, False otherwise.
-    """
-    title = uniformize(title)
-    ref_biblio = uniformize(ref_biblio)
-
-    distance = fuzz.partial_ratio(title, ref_biblio)
-
-    #thereshold here
-    return distance > 90
-
 def compare_pubinfo_refbiblio(item,ref_biblio):
     """
     Compare informations of one of the crossref publis with the biblio
@@ -148,9 +129,9 @@ def compare_pubinfo_refbiblio(item,ref_biblio):
         tuple (bool, str): True if it's match and whith the doi
     """
     # Check first author
-    if item['first_author_name'] not in ref_biblio:
+    if uniformize(item['first_author_name']) not in ref_biblio:
         return False, ""
-    if not match_title(item['title'], ref_biblio):
+    if fuzz.partial_ratio(uniformize(item['title']), ref_biblio)<90:
         return False, ""
     return True, item['doi']
 
@@ -165,6 +146,7 @@ def verify_biblio(ref_biblio, mail=mail_adress):
     Returns : 
         a confidence score about the existence + doi of the biblio ref
     """
+    ref_biblio = uniformize(ref_biblio)
     url = f'https://api.crossref.org/works?query.bibliographic="{ref_biblio}"&mailto={mail}&rows=5'
     try:
         response = session.get(url)
@@ -178,6 +160,8 @@ def verify_biblio(ref_biblio, mail=mail_adress):
             # compare pub_info with ref_biblio
             match_item, doi = compare_pubinfo_refbiblio(item_info,ref_biblio)
             if match_item:
+                if doi in retracted_doi:
+                    return "retracted",doi
                 return "found",doi
 
         return "not_found",""
@@ -207,7 +191,7 @@ def verify_biblio(ref_biblio, mail=mail_adress):
             json.dump(data, sys.stdout)
             sys.stdout.write("\n")
 
-        elif crossref_status_code==404:  # If request return code 404
+        elif crossref_status_code==404:  # If request return code 404, check the title dans author
             status,doi = verify_biblio(ref_biblio)
             data["value"] = {"doi":doi, "status": status}