Implementar verificação de disponibilidade de artigos em scielo.br

article/choices.py

@@ -1,4 +1,5 @@
 from django.utils.translation import gettext as _
+from core.utils.requester import NonRetryableError, RetryableError
 
 # Model RequestArticleChange, Field change_type
 RCT_UPDATE = "update"
@@ -124,3 +125,13 @@
     (AS_SCHEDULED_TO_PUBLISH, _("Scheduled to publish")),
     (AS_PUBLISHED, _("Published")),
 )
+
+VERIFY_ARTICLE_TYPE = [
+    ("TEXT", _("Text")),
+    ("PDF", _("pdf")),
+]
+
+VERIFY_HTTP_ERROR_CODE = [
+    (RetryableError, _("Excessively long response time. Retry later")),
+    (NonRetryableError, _("Site not found.")),
+]

article/migrations/0002_scielositestatus_checkarticleavailability.py

@@ -0,0 +1,141 @@
+# Generated by Django 5.0.3 on 2024-05-23 15:09
+
+import django.db.models.deletion
+from django.conf import settings
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("article", "0001_initial"),
+        migrations.swappable_dependency(settings.AUTH_USER_MODEL),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name="ScieloSiteStatus",
+            fields=[
+                (
+                    "id",
+                    models.BigAutoField(
+                        auto_created=True,
+                        primary_key=True,
+                        serialize=False,
+                        verbose_name="ID",
+                    ),
+                ),
+                (
+                    "created",
+                    models.DateTimeField(
+                        auto_now_add=True, verbose_name="Creation date"
+                    ),
+                ),
+                (
+                    "updated",
+                    models.DateTimeField(
+                        auto_now=True, verbose_name="Last update date"
+                    ),
+                ),
+                ("check_date", models.DateTimeField(blank=True, null=True)),
+                ("url_site_scielo", models.SlugField(max_length=500, unique=True)),
+                ("status", models.CharField(blank=True, max_length=80, null=True)),
+                (
+                    "type",
+                    models.CharField(
+                        blank=True,
+                        choices=[("TEXT", "Texto"), ("PDF", "pdf")],
+                        max_length=10,
+                        null=True,
+                    ),
+                ),
+                ("available", models.BooleanField(default=False)),
+                (
+                    "creator",
+                    models.ForeignKey(
+                        editable=False,
+                        on_delete=django.db.models.deletion.CASCADE,
+                        related_name="%(class)s_creator",
+                        to=settings.AUTH_USER_MODEL,
+                        verbose_name="Creator",
+                    ),
+                ),
+                (
+                    "updated_by",
+                    models.ForeignKey(
+                        blank=True,
+                        editable=False,
+                        null=True,
+                        on_delete=django.db.models.deletion.CASCADE,
+                        related_name="%(class)s_last_mod_user",
+                        to=settings.AUTH_USER_MODEL,
+                        verbose_name="Updater",
+                    ),
+                ),
+            ],
+            options={
+                "verbose_name": "Scielo Site Status",
+                "verbose_name_plural": "Scielo Site Status",
+            },
+        ),
+        migrations.CreateModel(
+            name="CheckArticleAvailability",
+            fields=[
+                (
+                    "id",
+                    models.BigAutoField(
+                        auto_created=True,
+                        primary_key=True,
+                        serialize=False,
+                        verbose_name="ID",
+                    ),
+                ),
+                (
+                    "created",
+                    models.DateTimeField(
+                        auto_now_add=True, verbose_name="Creation date"
+                    ),
+                ),
+                (
+                    "updated",
+                    models.DateTimeField(
+                        auto_now=True, verbose_name="Last update date"
+                    ),
+                ),
+                (
+                    "article",
+                    models.ForeignKey(
+                        null=True,
+                        on_delete=django.db.models.deletion.SET_NULL,
+                        to="article.article",
+                        unique=True,
+                    ),
+                ),
+                (
+                    "creator",
+                    models.ForeignKey(
+                        editable=False,
+                        on_delete=django.db.models.deletion.CASCADE,
+                        related_name="%(class)s_creator",
+                        to=settings.AUTH_USER_MODEL,
+                        verbose_name="Creator",
+                    ),
+                ),
+                (
+                    "updated_by",
+                    models.ForeignKey(
+                        blank=True,
+                        editable=False,
+                        null=True,
+                        on_delete=django.db.models.deletion.CASCADE,
+                        related_name="%(class)s_last_mod_user",
+                        to=settings.AUTH_USER_MODEL,
+                        verbose_name="Updater",
+                    ),
+                ),
+                ("site_status", models.ManyToManyField(to="article.scielositestatus")),
+            ],
+            options={
+                "abstract": False,
+            },
+        ),
+    ]

article/models.py

@@ -1,7 +1,8 @@
 import logging
+import datetime
 
 from django.contrib.auth import get_user_model
-from django.db import models
+from django.db import models, IntegrityError
 from django.utils.translation import gettext_lazy as _
 from modelcluster.fields import ParentalKey
 from modelcluster.models import ClusterableModel
@@ -29,6 +30,10 @@
 User = get_user_model()
 
 
+def verify_type_of_url(type):
+    return dict(choices.VERIFY_ARTICLE_TYPE).get("PDF") if type else dict(choices.VERIFY_ARTICLE_TYPE).get("TEXT")
+
+
 class Article(ClusterableModel, CommonControlField):
     """
     No contexto de Upload, Article deve conter o mínimo de campos,
@@ -314,3 +319,198 @@ def __str__(self) -> str:
         return f"{self.article or self.pid_v3} - {self.deadline}"
 
     base_form_class = RequestArticleChangeForm
+
+
+class CheckArticleAvailability(CommonControlField):
+    """
+        Modelo para armazenar o status de disponibilidade nos sites,
+        tanto na nova versao, quanto na antiga, do scielo.br.
+    """
+    article = models.ForeignKey(
+        Article, 
+        on_delete=models.SET_NULL,
+        null=True,
+        unique=True,
+    )
+    site_status = models.ManyToManyField(
+        "ScieloSiteStatus"
+    )
+
+    def __str__(self):
+        return f"{self.article.pid_v3}"
+
+    @classmethod
+    def get(cls, article):
+        return cls.objects.get(article=article)
+
+    def create_or_update_scielo_site_status(
+        self,
+        url,
+        status,
+        type,
+        available,
+        user,
+        date=None,
+    ):
+        obj = ScieloSiteStatus.create_or_update(
+            url=url,
+            status=status,
+            type=type,
+            available=available,
+            date=date,
+            user=user,
+        )
+        self.site_status.add(obj) 
+        self.save()
+
+
+    @classmethod
+    def create(
+        cls,
+        article,
+        status,
+        available,
+        url,
+        type,
+        user,
+        date=None,
+    ):
+        try:
+            obj = cls(
+                article=article,
+                creator=user,
+            )
+            obj.save()
+        except IntegrityError:
+            obj = cls.get(article=article)
+        obj.create_or_update_scielo_site_status(
+            url=url,
+            status=status,
+            type=type,
+            available=available,
+            user=user,
+            date=date,
+        )
+        return obj
+
+    @classmethod
+    def create_or_update(cls,
+        article,
+        status,
+        available,
+        url,
+        type,
+        user,
+        date=None,
+    ):
+        try:
+            obj = cls.get(article=article)
+            obj.create_or_update_scielo_site_status(
+            url=url,
+            status=status,
+            type=type,
+            available=available,
+            date=date,
+            user=user,
+        )
+            return obj
+        except cls.DoesNotExist:
+            cls.create(
+                article=article,
+                status=status,
+                available=available,
+                url=url,
+                type=type,
+                date=date,
+                user=user
+            )
+
+class ScieloSiteStatus(CommonControlField):
+    check_date = models.DateTimeField(null=True, blank=True)
+    url_site_scielo = models.SlugField(max_length=500, unique=True)
+    status = models.CharField(
+        max_length=80, 
+        null=True, 
+        blank=True
+    )
+    type = models.CharField(
+        max_length=10, 
+        choices=choices.VERIFY_ARTICLE_TYPE, 
+        null=True, 
+        blank=True,
+    )
+    available = models.BooleanField(default=False)
+
+    def update(
+        self,
+        status,
+        type,
+        available,
+        date=None,
+    ):
+        self.check_date = date or datetime.datetime.now()
+        self.status = status
+        self.available = available
+        self.type = verify_type_of_url(type)
+        self.save()
+        return self
+
+    class Meta:
+        verbose_name = "Scielo Site Status"
+        verbose_name_plural = "Scielo Site Status"
+
+    @classmethod
+    def get(cls, url):
+        return cls.objects.get(url_site_scielo=url)
+
+
+    @classmethod
+    def create(
+        cls,
+        url,
+        status,
+        type,
+        available,
+        user,
+        date=None,
+    ):
+        date = date or datetime.datetime.now()
+        obj = cls(
+            check_date=date,
+            url_site_scielo=url,
+            status=status,
+            type=verify_type_of_url(type),
+            available=available,
+            creator=user
+        )
+        obj.save()
+        return obj
+
+    @classmethod
+    def create_or_update(
+        cls,
+        url,
+        status,
+        type,
+        available,
+        user,
+        date=None,
+    ):
+        try:
+            obj = cls.get(url=url)
+            obj.update(
+                status=status,
+                type=type,
+                available=available,
+                date=date
+                )
+            return obj
+        except cls.DoesNotExist:
+            return cls.create(
+                url=url,
+                status=status,
+                type=type,
+                available=available,
+                user=user,
+                date=date
+            )

article/scripts/load_check_article.py

@@ -0,0 +1,11 @@
+from article.tasks import initiate_article_availability_check
+
+
+def run(pid_v3, username=None, user_id=None):
+    initiate_article_availability_check.apply_async(
+        kwargs=dict(
+            username=username,
+            user_id=user_id,
+            article_pid_v3=pid_v3,
+        )
+    )