Implementar verificação de disponibilidade de artigos no site QA e public (opac_5)

article/migrations/0002_scielositestatus_checkarticleavailability.py

@@ -0,0 +1,129 @@
+# Generated by Django 5.0.3 on 2024-05-21 00:43
+
+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)),
+                ("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={
+                "abstract": False,
+            },
+        ),
+        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",
+                    ),
+                ),
+                (
+                    "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

@@ -1,4 +1,5 @@
 import logging
+import datetime
 
 from django.contrib.auth import get_user_model
 from django.db import models
@@ -314,3 +315,152 @@ 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,
+    )
+    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,
+        user,
+        date=None,
+    ):
+        obj = ScieloSiteStatus.create_or_update(
+            url=url,
+            status=status,
+            date=date,
+            user=user,
+        )
+        self.site_status.add(obj) 
+        self.save()
+
+
+    @classmethod
+    def create(
+        cls,
+        article,
+        status,
+        url,
+        user,
+        date=None,
+    ):
+        obj = cls(
+            article=article,
+            creator=user,
+        )
+        obj.save()
+        obj.create_or_update_scielo_site_status(
+            url=url,
+            status=status,
+            user=user,
+            date=date,
+        )
+        return obj
+
+    @classmethod
+    def create_or_update(cls,
+        article,
+        status,
+        url,
+        user,
+        date=None,
+    ):
+        try:
+            obj = cls.get(article=article)
+            obj.create_or_update_scielo_site_status(
+            url=url,
+            status=status,
+            date=date,
+            user=user,
+        )
+            return obj
+        except cls.DoesNotExist:
+            cls.create(
+                article=article,
+                status=status,
+                url=url,
+                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)
+    available = models.BooleanField(default=False)
+
+    def update(
+        self,
+        status,
+        date=None,
+    ):
+        self.check_date = date or datetime.datetime.now()
+        self.available = status
+        self.save()
+        return self
+
+
+    @classmethod
+    def get(cls, url):
+        return cls.objects.get(url_site_scielo=url)
+
+
+    @classmethod
+    def create(
+        cls,
+        url,
+        status,
+        user,
+        date=None,
+    ):
+        date = date or datetime.datetime.now()
+        obj = cls(
+            check_date=date,
+            url_site_scielo=url,
+            available=status,
+            creator=user
+        )
+        obj.save()
+        return obj
+
+    @classmethod
+    def create_or_update(
+        cls,
+        url,
+        status,
+        user,
+        date=None,
+    ):
+        try:
+            obj = cls.get(url=url)
+            obj.update(
+                status=status,
+                date=date
+                )
+            return obj
+        except cls.DoesNotExist:
+            return cls.create(
+                url=url,
+                status=status,
+                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,
+        )
+    )

article/tasks.py

@@ -0,0 +1,89 @@
+import sys
+from django.db.models import Q
+
+from config import celery_app
+from core.utils.get_user import _get_user
+from core.utils.requester import fetch_data
+from article.models import CheckArticleAvailability, Article
+from collection.models import Collection
+from tracker.models import UnexpectedEvent
+
+@celery_app.task(bind=True)
+def initiate_article_availability_check(
+    self,
+    username,
+    user_id,
+    issn_print=None,
+    issn_electronic=None,
+    publication_year=None,
+    updated=None,
+    article_pid_v3=None,
+    collection_acron=None,
+):
+    if collection_acron:
+        collection = Collection.objects.filter(acron=collection_acron)
+    else:
+        collection = Collection.objects.all()
+
+    query = Q(journal__journalproc__collection__in=collection)
+    if not updated:
+        if article_pid_v3:
+            query |= Q(pid_v3=article_pid_v3)
+        if issn_print:
+            query |= Q(journal__official_journal__issn_print=issn_print)
+        if issn_electronic:
+            query |= Q(journal__official_journal__issn_electronic=issn_electronic)
+        if publication_year:
+            query |= Q(issue__publication_year=publication_year)
+
+    articles = Article.objects.filter(query)
+
+    for article in articles.iterator():
+        for article_per_lang in article.doi_with_lang.lang:
+            process_article_availability.apply_async(
+                kwargs=dict(
+                    user_id=user_id,
+                    username=username,
+                    pid_v3=article.pid_v3,
+                    journal_acron=article.journal.journal_acron,
+                    lang=article_per_lang,
+                )
+            )
+
+
+@celery_app.task(bind=True)
+def process_article_availability(self, user_id, username, pid_v3, journal_acron, lang):
+    urls = [
+        f"https://www.scielo.br/scielo.php?script=sci_arttext&pid={pid_v3}&lng={lang}&nrm=iso", 
+        f"https://www.scielo.br/j/{journal_acron}/a/{pid_v3}/?lang={lang}"
+    ]
+    try:
+        user = _get_user(self.request, user_id=user_id, username=username)
+        article = Article.objects.get(pid_v3=pid_v3)
+
+        for url in urls:
+            try:
+                response = fetch_data(url, timeout=2, verify=True)
+                CheckArticleAvailability.create_or_update(
+                    article=article,
+                    status=True,
+                    url=url,
+                    user=user,
+                )
+            except Exception as e:
+                CheckArticleAvailability.create_or_update(
+                    article=article,
+                    status=False,
+                    url=url,
+                    user=user,
+                )
+    except Exception as e:
+        exc_type, exc_value, exc_traceback = sys.exc_info()
+        UnexpectedEvent.create(
+            e=e,
+            exc_traceback=exc_traceback,
+            detail={
+                "function": "article.tasks.process_article_availability",
+                "urls": urls
+            },
+        )