#318 Added article compliance check for Funded by SCOAP3

pyproject.toml

@@ -46,6 +46,7 @@ django-lifecycle = "^1.1.2"
 country-converter = "^1.2"
 numpy = "<2"
 django-select2 = "^8.2.1"
+pymupdf = "^1.24.10"
 
 [tool.poetry.dev-dependencies]
 Werkzeug = {extras = ["watchdog"], version = "^2.3.4"}

scoap3/articles/admin.py

@@ -28,6 +28,7 @@ class ComplianceReportAdmin(admin.ModelAdmin):
         "check_article_type",
         "check_doi_registration_time",
         "check_authors_affiliation",
+        "check_contains_funded_by_scoap3",
         "get_is_compliant",
         "report_date",
     ]
@@ -53,6 +54,8 @@ class ComplianceReportAdmin(admin.ModelAdmin):
         "check_doi_registration_time_description",
         "check_authors_affiliation",
         "check_authors_affiliation_description",
+        "check_contains_funded_by_scoap3",
+        "check_contains_funded_by_scoap3_description",
     ]
     readonly_fields = [
         "article",
@@ -70,6 +73,8 @@ class ComplianceReportAdmin(admin.ModelAdmin):
         "check_doi_registration_time_description",
         "check_authors_affiliation",
         "check_authors_affiliation_description",
+        "check_contains_funded_by_scoap3",
+        "check_contains_funded_by_scoap3_description",
     ]
 
     list_filter = [
@@ -84,6 +89,7 @@ class ComplianceReportAdmin(admin.ModelAdmin):
         "article_id__report__check_article_type",
         "article_id__report__check_doi_registration_time",
         "article_id__report__check_authors_affiliation",
+        "article_id__report__check_contains_funded_by_scoap3",
     ]
 
     actions = ["export_as_csv"]
@@ -125,6 +131,7 @@ def export_as_csv(self, request, queryset):
             "Check DOI Registration": "check_doi_registration_time_description",
         }
 
+
         response = HttpResponse(content_type="text/csv")
         response["Content-Disposition"] = f"attachment; filename={filename}"
         writer = csv.writer(response)
@@ -164,6 +171,8 @@ class ArticleComplianceReportInline(admin.StackedInline):
         "check_doi_registration_time_description",
         "check_authors_affiliation",
         "check_authors_affiliation_description",
+        "check_contains_funded_by_scoap3",
+        "check_contains_funded_by_scoap3_description",
     ]
     can_delete = False
     can_create = False
@@ -186,6 +195,10 @@ class ArticleComplianceReportInline(admin.StackedInline):
                         "check_authors_affiliation",
                         "check_authors_affiliation_description",
                     ),
+                    (
+                        "check_contains_funded_by_scoap3",
+                        "check_contains_funded_by_scoap3_description",
+                    ),
                 ]
             },
         ),

scoap3/articles/migrations/0015_compliancereport_check_contains_funded_by_scoap3_and_more.py

@@ -0,0 +1,22 @@
+# Generated by Django 4.2.5 on 2024-09-17 11:31
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("articles", "0014_compliancereport_check_authors_affiliation_and_more"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="compliancereport",
+            name="check_contains_funded_by_scoap3",
+            field=models.BooleanField(default=False),
+        ),
+        migrations.AddField(
+            model_name="compliancereport",
+            name="check_contains_funded_by_scoap3_description",
+            field=models.TextField(blank=True, default=""),
+        ),
+    ]

scoap3/articles/models.py

@@ -103,6 +103,10 @@ class ComplianceReport(models.Model):
     check_doi_registration_time_description = models.TextField(blank=True, default="")
     check_authors_affiliation = models.BooleanField(default=False)
     check_authors_affiliation_description = models.TextField(blank=True, default="")
+    check_contains_funded_by_scoap3 = models.BooleanField(default=False)
+    check_contains_funded_by_scoap3_description = models.TextField(
+        blank=True, default=""
+    )
 
     def __str__(self):
         return f"Compliance Report for {self.article.title} on {self.report_date.strftime('%Y-%m-%d')}"

scoap3/articles/tasks.py

@@ -5,7 +5,8 @@
 from django.core.paginator import Paginator
 from django_opensearch_dsl.registries import registry
 
-from scoap3.articles.models import Article, ComplianceReport
+from scoap3.articles.models import Article, ArticleFile, ComplianceReport
+from scoap3.articles.util import is_string_in_pdf
 from scoap3.authors.models import Author
 from scoap3.misc.models import Affiliation
 from scoap3.misc.utils import fetch_doi_registration_date
@@ -133,6 +134,29 @@ def check_authors_affiliation(article):
     return True, "Authors' affiliations are compliant"
 
 
+def check_contains_funded_by_scoap3(article):
+    try:
+        article_files = ArticleFile.objects.filter(article_id=article)
+
+        if not article_files.exists():
+            return False, "No files found for the given article."
+
+        for article_file in article_files:
+            file_path = article_file.file.path
+            try:
+                if is_string_in_pdf(file_path, "Funded by SCOAP3"):
+                    return (
+                        True,
+                        f"Files contain the required text: 'Funded by SCOAP3'. File: {file_path}",
+                    )
+            except FileNotFoundError:
+                return False, f"File not found: {file_path}"
+
+        return False, "Files do not contain the required text: 'Funded by SCOAP3'"
+    except Exception as e:
+        return False, f"An unexpected error occurred: {str(e)}"
+
+
 @shared_task(name="compliance_checks", acks_late=True)
 def compliance_checks(article_id):
     try:
@@ -160,6 +184,10 @@ def compliance_checks(article_id):
         check_affiliations_compliance,
         check_affiliations_description,
     ) = check_authors_affiliation(article)
+    (
+        check_funded_by_scoap3_compliance,
+        check_funded_by_scoap3_description,
+    ) = check_contains_funded_by_scoap3(article)
 
     article.report.all().delete()
 
@@ -177,6 +205,8 @@ def compliance_checks(article_id):
         check_license_description=check_license_description,
         check_authors_affiliation=check_affiliations_compliance,
         check_authors_affiliation_description=check_affiliations_description,
+        check_contains_funded_by_scoap3=check_funded_by_scoap3_compliance,
+        check_contains_funded_by_scoap3_description=check_funded_by_scoap3_description,
     )
     report.save()
     logger.info("Compliance checks completed for article %s", article_id)

scoap3/articles/tests/test_funded_by_scoap3.py

@@ -0,0 +1,126 @@
+from io import BytesIO
+
+import fitz  # PyMuPDF
+import pytest
+from django.core.files import File
+from django.core.files.base import ContentFile
+
+from scoap3.articles.models import Article, ArticleFile
+from scoap3.articles.tasks import check_contains_funded_by_scoap3
+
+pytestmark = pytest.mark.django_db
+
+
+@pytest.fixture
+def create_article(db):
+    def _create_article():
+        return Article.objects.create(
+            title="Test Article",
+            subtitle="Subtitle",
+            abstract="Abstract",
+        )
+
+    return _create_article
+
+
+@pytest.fixture
+def create_pdf_with_text():
+    def _create_pdf_with_text(text):
+        pdf_bytes = BytesIO()
+        doc = fitz.open()
+        page = doc.new_page()
+        page.insert_text((72, 72), text)
+        doc.save(pdf_bytes)
+        doc.close()
+        pdf_bytes.seek(0)
+        return ContentFile(pdf_bytes.read(), "test_file.pdf")
+
+    return _create_pdf_with_text
+
+
+@pytest.fixture
+def attach_file_to_article(db):
+    def _attach_file_to_article(article, content, file_name):
+        file = File(content, name=file_name)
+        ArticleFile.objects.create(article_id=article, file=file)
+
+    return _attach_file_to_article
+
+
+class TestCheckContainsFundedBySCOAP3:
+    def test_contains_funded_by(
+        self, create_article, create_pdf_with_text, attach_file_to_article
+    ):
+        article = create_article()
+
+        file_with_text = create_pdf_with_text("Funded by SCOAP3")
+        attach_file_to_article(article, file_with_text, "file_with_text.pdf")
+
+        result, message = check_contains_funded_by_scoap3(article)
+        assert result is True
+        assert message.startswith("Files contain the required text: 'Funded by SCOAP3'")
+
+    def test_contains_funded_by_multiple_files(
+        self, create_article, create_pdf_with_text, attach_file_to_article
+    ):
+        article = create_article()
+
+        file_with_text_1 = create_pdf_with_text("Funded by SCOAP3")
+        file_with_text_2 = create_pdf_with_text(
+            "Some other content. Funded by SCOAP3 again."
+        )
+
+        attach_file_to_article(article, file_with_text_1, "file_with_text_1.pdf")
+        attach_file_to_article(article, file_with_text_2, "file_with_text_2.pdf")
+
+        result, message = check_contains_funded_by_scoap3(article)
+        assert result is True
+        assert message.startswith("Files contain the required text: 'Funded by SCOAP3'")
+
+    def test_does_not_contain_funded_by(
+        self, create_article, create_pdf_with_text, attach_file_to_article
+    ):
+        article = create_article()
+
+        file_without_text = create_pdf_with_text("Other text")
+        attach_file_to_article(article, file_without_text, "file_without_text.pdf")
+
+        result, message = check_contains_funded_by_scoap3(article)
+        assert result is False
+        assert message == "Files do not contain the required text: 'Funded by SCOAP3'"
+
+    def test_does_not_contain_funded_by_multiple_files(
+        self, create_article, create_pdf_with_text, attach_file_to_article
+    ):
+        article = create_article()
+
+        file_without_text_1 = create_pdf_with_text("This is some random text.")
+        file_without_text_2 = create_pdf_with_text("Some other random content.")
+
+        attach_file_to_article(article, file_without_text_1, "file_without_text_1.pdf")
+        attach_file_to_article(article, file_without_text_2, "file_without_text_2.pdf")
+
+        result, message = check_contains_funded_by_scoap3(article)
+        assert result is False
+        assert message == "Files do not contain the required text: 'Funded by SCOAP3'"
+
+    def test_mixed_files(
+        self, create_article, create_pdf_with_text, attach_file_to_article
+    ):
+        article = create_article()
+
+        file_with_text = create_pdf_with_text("Funded by SCOAP3")
+        file_without_text = create_pdf_with_text("Other text")
+        attach_file_to_article(article, file_with_text, "file_with_text.pdf")
+        attach_file_to_article(article, file_without_text, "file_without_text.pdf")
+
+        result, message = check_contains_funded_by_scoap3(article)
+        assert result is True
+        assert message.startswith("Files contain the required text: 'Funded by SCOAP3'")
+
+    def test_no_files(self, create_article):
+        article = create_article()
+
+        result, message = check_contains_funded_by_scoap3(article)
+        assert result is False
+        assert message == "No files found for the given article."

scoap3/articles/util.py

@@ -1,5 +1,7 @@
 from datetime import datetime
 
+import fitz
+
 from scoap3.articles.models import ArticleIdentifierType
 
 
@@ -25,3 +27,23 @@ def get_arxiv_primary_category(article_document):
 
 def parse_string_to_date_object(date_string):
     return datetime.fromisoformat(date_string.replace("Z", "+00:00"))
+
+
+def is_string_in_pdf(pdf_path, search_string):
+    try:
+        document = fitz.open(pdf_path)
+        search_string_lower = search_string.lower()
+
+        for page_num in range(document.page_count):
+            page = document[page_num]
+            page_text = page.get_text().lower()
+            if search_string_lower in page_text:
+                document.close()
+                return True
+
+        document.close()
+        return False
+    except FileNotFoundError:
+        raise FileNotFoundError(f"File not found: {pdf_path}")
+    except Exception as e:
+        raise Exception(f"An error occurred while reading the PDF: {str(e)}")

scoap3/authors/migrations/0005_alter_author_article_id.py

@@ -0,0 +1,23 @@
+# Generated by Django 4.2.5 on 2024-09-17 11:31
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("articles", "0015_compliancereport_check_contains_funded_by_scoap3_and_more"),
+        ("authors", "0004_alter_authoridentifier_author_id"),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name="author",
+            name="article_id",
+            field=models.ForeignKey(
+                on_delete=django.db.models.deletion.CASCADE,
+                related_name="authors",
+                to="articles.article",
+            ),
+        ),
+    ]