From bc627afaffa97e2450ce6489a22bf4ad26e2d73f Mon Sep 17 00:00:00 2001
From: Vinicius Aguiar <vncsna@gmail.com>
Date: Sun, 4 Feb 2024 13:15:52 -0300
Subject: [PATCH] feat: add page views metadata (#557)

---
 ...027_dataset_page_views_table_page_views.py | 28 ++++++++++
 bd_api/apps/api/v1/models.py                  | 10 +++-
 bd_api/apps/api/v1/tasks.py                   | 55 ++++++++++++++++++-
 3 files changed, 91 insertions(+), 2 deletions(-)
 create mode 100644 bd_api/apps/api/v1/migrations/0027_dataset_page_views_table_page_views.py

diff --git a/bd_api/apps/api/v1/migrations/0027_dataset_page_views_table_page_views.py b/bd_api/apps/api/v1/migrations/0027_dataset_page_views_table_page_views.py
new file mode 100644
index 00000000..5518edb7
--- /dev/null
+++ b/bd_api/apps/api/v1/migrations/0027_dataset_page_views_table_page_views.py
@@ -0,0 +1,28 @@
+# -*- coding: utf-8 -*-
+# Generated by Django 4.2.6 on 2024-02-04 16:08
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("v1", "0026_alter_table_source_bucket_name"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="dataset",
+            name="page_views",
+            field=models.BigIntegerField(
+                default=0, help_text="Number of page views by Google Analytics"
+            ),
+        ),
+        migrations.AddField(
+            model_name="table",
+            name="page_views",
+            field=models.BigIntegerField(
+                default=0, help_text="Number of page views by Google Analytics"
+            ),
+        ),
+    ]
diff --git a/bd_api/apps/api/v1/models.py b/bd_api/apps/api/v1/models.py
index 8eb835a4..24326f62 100644
--- a/bd_api/apps/api/v1/models.py
+++ b/bd_api/apps/api/v1/models.py
@@ -551,6 +551,10 @@ class Dataset(BaseModel):
     is_closed = models.BooleanField(
         default=False, help_text="Dataset is for BD Pro subscribers only"
     )
+    page_views = models.BigIntegerField(
+        default=0,
+        help_text="Number of page views by Google Analytics",
+    )
 
     graphql_nested_filter_fields_whitelist = ["id", "slug"]
 
@@ -946,8 +950,12 @@ class Table(BaseModel, OrderedModel):
     number_rows = models.BigIntegerField(blank=True, null=True)
     number_columns = models.BigIntegerField(blank=True, null=True)
     is_closed = models.BooleanField(default=False, help_text="Table is for BD Pro subscribers only")
-    order_with_respect_to = ("dataset",)
+    page_views = models.BigIntegerField(
+        default=0,
+        help_text="Number of page views by Google Analytics",
+    )
 
+    order_with_respect_to = ("dataset",)
     graphql_nested_filter_fields_whitelist = ["id", "dataset"]
 
     def __str__(self):
diff --git a/bd_api/apps/api/v1/tasks.py b/bd_api/apps/api/v1/tasks.py
index 61144805..91a03adf 100644
--- a/bd_api/apps/api/v1/tasks.py
+++ b/bd_api/apps/api/v1/tasks.py
@@ -1,4 +1,6 @@
 # -*- coding: utf-8 -*-
+from datetime import datetime, timedelta
+
 from django.core.management import call_command
 from google.api_core.exceptions import BadRequest, NotFound
 from google.cloud.bigquery import Table as GBQTable
@@ -7,7 +9,7 @@
 from loguru import logger
 from pandas import read_gbq
 
-from bd_api.apps.api.v1.models import Table
+from bd_api.apps.api.v1.models import Dataset, Table
 from bd_api.custom.client import get_gbq_client, get_gcs_client, send_discord_message
 from bd_api.utils import production_task
 
@@ -121,3 +123,54 @@ def format_msg(msg: list[str]) -> str:
 
     if msg := format_msg(msg):
         send_discord_message(msg)
+
+
+@periodic_task(crontab(hour="6", minute="0"))
+@production_task
+def update_page_views_task(backfill: bool = False):
+    if backfill:
+        event_table = "events_*"
+    else:
+        yesterday = datetime.now() - timedelta(1)
+        yesterday = yesterday.strftime("%Y%m%d")
+        event_table = f"events_{yesterday}"
+
+    query = f"""
+        select
+            count(1) page_views
+            , regexp_extract(param.value.string_value, r'table=([a-z0-9-]{36})') table_id
+            , regexp_extract(param.value.string_value, r'dataset\/([a-z0-9-]{36})') dataset_id
+        from `basedosdados.analytics_295884852.{event_table}` event
+            join unnest(event_params) param
+        where
+            true
+            and event_name = 'page_view'
+            and param.key = 'page_location'
+            and param.value.string_value like '%/dataset/%'
+        group by
+            table_id,
+            dataset_id
+        having
+            true
+            and table_id is not null
+            and dataset_id is not null
+    """  # noqa: W605
+    metadata = read_gbq(query)
+
+    if backfill:
+        for table_id in metadata["table_id"].unique():
+            if table := Table.objects.filter(id=table_id).first():
+                table.page_views = 0
+                table.save()
+        for dataset_id in metadata["dataset_id"].unique():
+            if dataset := Dataset.objects.filter(id=dataset_id).first():
+                dataset.page_views = 0
+                dataset.save()
+
+    for _, (page_views, table_id, dataset_id) in metadata.iterrows():
+        if table := Table.objects.filter(id=table_id).first():
+            table.page_views += page_views
+            table.save()
+        if dataset := Dataset.objects.filter(id=dataset_id).first():
+            dataset.page_views += page_views
+            dataset.save()