Merge pull request #351 from gitnnolabs/add_op_journals

Adiciona o modelo e coleta de periódico a partir do OpenAlex

article/wagtail_hooks.py

@@ -1,19 +1,20 @@
 from django.utils.translation import gettext as _
-
 from wagtail.contrib.modeladmin.options import (
     ModelAdmin,
-    modeladmin_register,
     ModelAdminGroup,
+    modeladmin_register,
 )
 
+from journal.wagtail_hooks import SourceJournalAdmin
+
 from .models import (
+    Affiliation,
     Article,
-    SourceArticle,
+    Concepts,
     Contributor,
-    Affiliation,
     Journal,
     License,
-    Concepts
+    SourceArticle,
 )
 
 
@@ -223,6 +224,7 @@ class ArticleAdminGroup(ModelAdminGroup):
         LicenseAdmin,
         ConceptsAdmin,
         SourceArticleAdmin,
+        SourceJournalAdmin
     )
 
 

config/settings/base.py

@@ -130,6 +130,7 @@
     "wagtailautocomplete",
     # "provided_data",
     "article",
+    "journal",
 ]
 
 # https://docs.djangoproject.com/en/dev/ref/settings/#installed-apps
@@ -374,6 +375,7 @@
 URL_API_CROSSREF = env.str("API_CROSSREF", default="https://api.crossref.org/works")
 URL_API_OPENALEX = env.str("API_OPENALEX", default="https://api.openalex.org/works")
 URL_API_OPENALEX_INSTITUTIONS = env.str("URL_API_OPENALEX_INSTITUTIONS", default="https://api.openalex.org/institutions")
+URL_API_OPENALEX_JOURNALS = env.str("URL_API_OPENALEX_JOURNALS", default="https://api.openalex.org/sources")
 
 LANGUAGES = [
     ("pt-BR", "Portuguese"),

journal/admin.py

@@ -0,0 +1,3 @@
+from django.contrib import admin
+
+# Register your models here.

journal/apps.py

@@ -0,0 +1,6 @@
+from django.apps import AppConfig
+
+
+class JournalConfig(AppConfig):
+    default_auto_field = "django.db.models.BigAutoField"
+    name = "journal"

journal/migrations/0001_initial.py

@@ -0,0 +1,83 @@
+# Generated by Django 4.1.6 on 2024-10-14 07:33
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    initial = True
+
+    dependencies = [
+        ("core", "0002_language"),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name="SourceJournal",
+            fields=[
+                (
+                    "id",
+                    models.BigAutoField(
+                        auto_created=True,
+                        primary_key=True,
+                        serialize=False,
+                        verbose_name="ID",
+                    ),
+                ),
+                (
+                    "specific_id",
+                    models.CharField(max_length=255, verbose_name="Specific Id"),
+                ),
+                (
+                    "issn",
+                    models.CharField(max_length=255, null=True, verbose_name="DOI"),
+                ),
+                (
+                    "title",
+                    models.CharField(max_length=1024, null=True, verbose_name="Título"),
+                ),
+                (
+                    "updated",
+                    models.CharField(
+                        max_length=50, null=True, verbose_name="Source updated date"
+                    ),
+                ),
+                (
+                    "created",
+                    models.CharField(
+                        max_length=50, null=True, verbose_name="Source created date"
+                    ),
+                ),
+                (
+                    "raw",
+                    models.JSONField(
+                        blank=True, null=True, verbose_name="Arquivo JSON"
+                    ),
+                ),
+                (
+                    "source",
+                    models.ForeignKey(
+                        null=True,
+                        on_delete=django.db.models.deletion.CASCADE,
+                        to="core.source",
+                        verbose_name="Origem",
+                    ),
+                ),
+            ],
+        ),
+        migrations.AddIndex(
+            model_name="sourcejournal",
+            index=models.Index(fields=["issn"], name="journal_sou_issn_b43f7e_idx"),
+        ),
+        migrations.AddIndex(
+            model_name="sourcejournal",
+            index=models.Index(
+                fields=["specific_id"], name="journal_sou_specifi_206501_idx"
+            ),
+        ),
+        migrations.AddIndex(
+            model_name="sourcejournal",
+            index=models.Index(fields=["title"], name="journal_sou_title_cd0c69_idx"),
+        ),
+    ]

journal/migrations/0002_remove_sourcejournal_journal_sou_issn_b43f7e_idx_and_more.py

@@ -0,0 +1,46 @@
+# Generated by Django 4.1.6 on 2024-10-14 08:16
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("journal", "0001_initial"),
+    ]
+
+    operations = [
+        migrations.RemoveIndex(
+            model_name="sourcejournal",
+            name="journal_sou_issn_b43f7e_idx",
+        ),
+        migrations.RemoveField(
+            model_name="sourcejournal",
+            name="issn",
+        ),
+        migrations.AddField(
+            model_name="sourcejournal",
+            name="country_code",
+            field=models.CharField(
+                max_length=255, null=True, verbose_name="Country Code"
+            ),
+        ),
+        migrations.AddField(
+            model_name="sourcejournal",
+            name="issn_l",
+            field=models.CharField(max_length=255, null=True, verbose_name="ISSN_L"),
+        ),
+        migrations.AddField(
+            model_name="sourcejournal",
+            name="issns",
+            field=models.CharField(max_length=255, null=True, verbose_name="ISSN"),
+        ),
+        migrations.AddIndex(
+            model_name="sourcejournal",
+            index=models.Index(fields=["issns"], name="journal_sou_issns_971d00_idx"),
+        ),
+        migrations.AddIndex(
+            model_name="sourcejournal",
+            index=models.Index(fields=["issn_l"], name="journal_sou_issn_l_971c0c_idx"),
+        ),
+    ]

journal/models.py

@@ -0,0 +1,154 @@
+from django.db import models
+from django.utils.translation import gettext as _
+
+from core.models import Source
+
+
+class SourceJournal(models.Model):
+    specific_id = models.CharField(
+        _("Specific Id"), max_length=255, null=False, blank=False
+    )
+    issns = models.CharField(_("ISSN"), max_length=255, null=True, blank=False)
+    issn_l = models.CharField(_("ISSN_L"), max_length=255, null=True, blank=False)
+    country_code = models.CharField(_("Country Code"), max_length=255, null=True, blank=False)
+    title = models.CharField(_("Title"), max_length=1024, null=True, blank=False)
+    updated = models.CharField(
+        _("Source updated date"), max_length=50, null=True, blank=False
+    )
+    created = models.CharField(
+        _("Source created date"), max_length=50, null=True, blank=False
+    )
+    raw = models.JSONField(_("JSON File"), null=True, blank=True)
+    source = models.ForeignKey(
+        Source,
+        verbose_name=_("Source"),
+        null=True,
+        on_delete=models.CASCADE,
+    )
+
+    class Meta:
+        indexes = [
+            models.Index(
+                fields=[
+                    "issns",
+                ]
+            ),
+            models.Index(
+                fields=[
+                    "issn_l",
+                ]
+            ),
+            models.Index(
+                fields=[
+                    "specific_id",
+                ]
+            ),
+            models.Index(
+                fields=[
+                    "title",
+                ]
+            ),
+        ]
+
+    def __unicode__(self):
+        return str("%s") % (self.issns or self.specific_id)
+
+    def __str__(self):
+        return str("%s") % (self.issns or self.specific_id)
+
+    @property
+    def has_specific_id(self):
+        return bool(self.specific_id)
+
+    @classmethod
+    def get(cls, **kwargs):
+        """
+        This function will try to get the source journal by attributes:
+
+            * issn
+            * specific_id
+
+        The kwargs must be a dict, something like this:
+
+            {
+                "specific_id": "https://openalex.org/sources/S183843087",
+                "issn": "1234-5678",
+                "source": OPENALEX,
+            }
+
+        return source journal|None
+
+        This function can raise:
+            ValueError
+            SourceJournal.DoesNotExist
+            SourceJournal.MultipleObjectsReturned
+        """
+
+        filters = {}
+
+        if (
+            not kwargs.get("issn")
+            and not kwargs.get("specific_id")
+            and not kwargs.get("source")
+        ):
+            raise ValueError("Param issn or specific_id is required")
+
+        if kwargs.get("specific_id"):
+            filters = {
+                "specific_id": kwargs.get("specific_id"),
+                "source": kwargs.get("source"),
+            }
+        elif kwargs.get("issn"):
+            filters = {"issn": kwargs.get("issn"), "source": kwargs.get("source")}
+
+        return cls.objects.get(**filters)
+
+
+    @classmethod
+    def create_or_update(cls, **kwargs):
+        """
+        This function will try to get the journal by issn.
+
+        If the journal exists update, otherwise create.
+
+        The kwargs must be a dict, something like this:
+
+            {
+                "issns": "1234-5678, 0987-6543",
+                "issn_l": "1987-9373",
+                "country_code": "BR",
+                "title": "Update the record",
+                "number": "999",
+                "volume": "9",
+                "sources": list of <sources> [<source>, <source>]
+            }
+
+        return journal(object), 0|1
+
+        0 = updated
+        1 = created
+
+        """
+
+        try:
+            journal = cls.get(**kwargs)
+            created = 0
+        except SourceJournal.DoesNotExist:
+            journal = cls.objects.create()
+            created = 1
+        except SourceJournal.MultipleObjectsReturned as e:
+            print(_("The source journal table have duplicity...."))
+            raise (SourceJournal.MultipleObjectsReturned)
+
+        journal.issns = kwargs.get("issns")
+        journal.issn_l = kwargs.get("issn_l")
+        journal.title = kwargs.get("title")
+        journal.country_code = kwargs.get("country_code")
+        journal.specific_id = kwargs.get("specific_id")
+        journal.updated = kwargs.get("updated")
+        journal.created = kwargs.get("created")
+        journal.raw = kwargs.get("raw")
+        journal.source = kwargs.get("source")
+        journal.save()
+
+        return journal, created

journal/scripts/load_journal_openalex.py

@@ -0,0 +1,23 @@
+from django.utils.translation import gettext as _
+
+from journal.tasks import load_journal_openalex
+
+from journal.models import SourceJournal
+
+
+def run(user_id, country="BR", delete=0, length=None):
+    """
+    Load the journal from OpenAlex to SourceJournal model.
+
+    About the OpenAlex API see: https://docs.openalex.org/
+    """
+
+    if int(delete):
+        SourceJournal.objects.all().delete()
+
+    if user_id and country and length:
+        load_journal_openalex.apply_async(args=(int(user_id), str(country), int(length)))
+    elif user_id and country:
+        load_journal_openalex.apply_async(args=(int(user_id), str(country)))
+    else:
+        print(_("Param user_id is required."))