is_compliant and check_file_formats

Adapted check_file formats for each specific journal, altered is_compliant to pass for every publication date before 2023 and added mark as compliant action.

scoap3/articles/admin.py

@@ -25,7 +25,6 @@ class ComplianceReportAdmin(admin.ModelAdmin):
         "compliant",
         "check_license",
         "check_required_file_formats",
-        "check_file_formats",
         "check_arxiv_category",
         "check_article_type",
         "check_doi_registration_time",
@@ -46,8 +45,6 @@ class ComplianceReportAdmin(admin.ModelAdmin):
         "check_license_description",
         "check_required_file_formats",
         "check_required_file_formats_description",
-        "check_file_formats",
-        "check_file_formats_description",
         "check_arxiv_category",
         "check_arxiv_category_description",
         "check_article_type",
@@ -63,8 +60,6 @@ class ComplianceReportAdmin(admin.ModelAdmin):
         "compliant",
         "check_license",
         "check_license_description",
-        "check_file_formats",
-        "check_file_formats_description",
         "check_arxiv_category",
         "check_arxiv_category_description",
         "check_article_type",
@@ -231,6 +226,18 @@ def get_identifiers(self, obj):
         )
 
 
+def make_compliant(obj):
+    article = obj
+
+    reports = article.report.all()
+
+    for report in reports:
+        report.compliant = True
+        report.save()
+
+    return f"Articles marked as compliant {article.id}"
+
+
 class ArticleAdmin(admin.ModelAdmin):
     list_display = [
         "id",
@@ -255,7 +262,7 @@ class ArticleAdmin(admin.ModelAdmin):
         "publication_info__journal_title",
         "article_identifiers__identifier_value",
     ]
-    actions = ["make_compliance_check"]
+    actions = ["make_compliance_check", "mark_as_compliant"]
     list_filter = [
         "report__compliant",
         "_updated_at",
@@ -314,6 +321,20 @@ def make_compliance_check(self, request, queryset):
             """,
         )
 
+    @admin.action(description="Mark as compliant")
+    def mark_as_compliant(self, request, queryset):
+        ids = []
+        for obj in queryset:
+            make_compliant(obj)
+            ids.append(str(obj.id))
+        messages.success(
+            request,
+            f"""
+            Selected articles are being processed, it might take some time before seeing
+            the results in the reports. {', '.join(ids)}.
+            """,
+        )
+
     @admin.display(boolean=True, description="Compliant")
     def check_compliance(self, obj):
         report = obj.report.first()

scoap3/articles/api/serializers.py

@@ -72,7 +72,7 @@ def get_metadata(self, obj):
         return {
             "_files": [
                 {
-                    "filetype": entry.file.type,
+                    "filetype": entry.filetype,
                     "size": entry.file.size,
                     "key": entry.file.name,
                 }

scoap3/articles/migrations/0017_remove_compliancereport_check_file_formats_and_more.py

@@ -0,0 +1,20 @@
+# Generated by Django 4.2.5 on 2024-09-30 09:02
+
+from django.db import migrations
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("articles", "0016_compliancereport_check_required_file_formats_and_more"),
+    ]
+
+    operations = [
+        migrations.RemoveField(
+            model_name="compliancereport",
+            name="check_file_formats",
+        ),
+        migrations.RemoveField(
+            model_name="compliancereport",
+            name="check_file_formats_description",
+        ),
+    ]

scoap3/articles/migrations/0018_articlefile_filetype.py

@@ -0,0 +1,17 @@
+# Generated by Django 4.2.5 on 2024-10-03 12:44
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("articles", "0017_remove_compliancereport_check_file_formats_and_more"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="articlefile",
+            name="filetype",
+            field=models.TextField(blank=True, default=""),
+        ),
+    ]

scoap3/articles/models.py

@@ -1,4 +1,6 @@
+import datetime
 import mimetypes
+from datetime import date
 
 from django.db import models
 from django.db.models.fields.files import FieldFile
@@ -71,6 +73,7 @@ class ArticleFile(models.Model):
     file = CustomFileField(upload_to=article_file_upload_path)
     created = models.DateTimeField(auto_now_add=True)
     updated = models.DateTimeField(auto_now=True)
+    filetype = models.TextField(blank=True, default="")
 
     def __str__(self) -> str:
         return self.file.name
@@ -105,8 +108,6 @@ class ComplianceReport(models.Model):
     check_license_description = models.TextField(blank=True, default="")
     check_required_file_formats = models.BooleanField(default=False)
     check_required_file_formats_description = models.TextField(blank=True, default="")
-    check_file_formats = models.BooleanField(default=False)
-    check_file_formats_description = models.TextField(blank=True, default="")
     check_arxiv_category = models.BooleanField(default=False)
     check_arxiv_category_description = models.TextField(blank=True, default="")
     check_article_type = models.BooleanField(default=False)
@@ -136,6 +137,11 @@ def is_compliant(self):
         ):
             _check_arxiv_category = self.check_arxiv_category
 
+        if isinstance(
+            self.article.publication_date, datetime.date
+        ) and self.article.publication_date < date(2023, 1, 1):
+            return True
+
         return all(
             [
                 self.check_license,

scoap3/articles/tasks.py

@@ -7,7 +7,7 @@
 
 from scoap3.articles.models import Article, ComplianceReport
 from scoap3.authors.models import Author
-from scoap3.misc.models import Affiliation
+from scoap3.misc.models import Affiliation, PublicationInfo
 from scoap3.misc.utils import fetch_doi_registration_date
 
 logger = logging.getLogger(__name__)
@@ -27,34 +27,39 @@ def check_license(obj):
 
 
 def check_required_file_formats(obj):
-    required_formats = ["pdf", "xml"]
-    available_formats = [
-        file.file.name.split(".")[-1] for file in obj.related_files.all()
-    ]
-
-    missing_formats = [f for f in required_formats if f not in available_formats]
-
-    if missing_formats:
-        return (
-            False,
-            f"Missing required file formats: {', '.join(missing_formats)}.",
-        )
-    return True, "All required file formats are present."
-
+    publication_info = PublicationInfo.objects.filter(article_id=obj).first()
+
+    if not publication_info:
+        return False, "No publication information found."
+
+    journal_format_mapping = {
+        "Acta Physica Polonica B": ["pdf", "pdf/a"],
+        "Adv. High Energy Phys.": ["pdf", "pdf/a", "xml"],
+        "Prog. of Theor. and Exp. Phys.": ["pdf", "pdf/a", "xml"],
+        "European Physical Journal C": ["pdf/a", "xml"],
+        "Journal of High Energy Physics": ["pdf/a", "xml"],
+        "Nuclear Physics B": ["pdf", "pdf/a", "xml"],
+        "Physics Letters B": ["pdf", "pdf/a", "xml"],
+        "Physical review letters": ["pdf", "xml"],
+        "Physical review C": ["pdf", "xml"],
+        "Physical review D": ["pdf", "xml"],
+        "Chinese Physics C": ["pdf", "xml"],
+    }
+
+    default_required_formats = ["pdf", "pdf/a", "xml"]
+
+    required_formats = journal_format_mapping.get(
+        publication_info.journal_title, default_required_formats
+    )
 
-def check_file_formats(obj):
-    required_formats = ["pdf", "pdf_a", "xml"]
-    available_formats = [
-        file.file.name.split(".")[-1] for file in obj.related_files.all()
-    ]
+    available_formats = []
+    for file in obj.related_files.all():
+        available_formats.append(file.filetype)
 
     missing_formats = [f for f in required_formats if f not in available_formats]
 
     if missing_formats:
-        return (
-            False,
-            f"Missing required file formats: {', '.join(missing_formats)}.",
-        )
+        return False, f"Missing required file formats: {', '.join(missing_formats)}."
     return True, "All required file formats are present."
 
 
@@ -172,9 +177,6 @@ def compliance_checks(article_id):
         check_required_file_formats_compliance,
         check_required_file_formats_description,
     ) = check_required_file_formats(article)
-    check_file_formats_compliance, check_file_formats_description = check_file_formats(
-        article
-    )
     check_license_compliance, check_license_description = check_license(article)
     (
         check_affiliations_compliance,
@@ -193,8 +195,6 @@ def compliance_checks(article_id):
         check_doi_registration_time_description=check_doi_registration_description,
         check_required_file_formats=check_required_file_formats_compliance,
         check_required_file_formats_description=check_required_file_formats_description,
-        check_file_formats=check_file_formats_compliance,
-        check_file_formats_description=check_file_formats_description,
         check_license=check_license_compliance,
         check_license_description=check_license_description,
         check_authors_affiliation=check_affiliations_compliance,