Initial attempt to add UI metrics to influx

galaxy/api/access.py

@@ -419,6 +419,22 @@ def can_delete(self, data):
         return False
 
 
+class InfluxSessionAccess(BaseAccess):
+    model = models.InfluxSessionIdentifier
+
+    def can_read(self, obj):
+        return True
+
+    def can_add(self, data):
+        return True
+
+    def can_change(self, obj, data):
+        return False
+
+    def can_delete(self, obj):
+        return False
+
+
 register_access(EmailConfirmation, EmailConfirmationAccess)
 register_access(User, UserAccess)
 register_access(EmailAddress, EmailAddressAccess)
@@ -440,3 +456,4 @@ def can_delete(self, data):
 register_access(models.ContentType, ContentTypeAccess)
 register_access(models.CloudPlatform, CloudPlatformsAccess)
 register_access(models.CommunitySurvey, CommunitySurveyAccess)
+register_access(models.InfluxSessionIdentifier, InfluxSessionAccess)

galaxy/api/serializers/__init__.py

@@ -31,3 +31,4 @@
 from .serializers import *         # noqa
 from .token import *               # noqa
 from .survey import *              # noqa
+from .influx import *              # noqa

galaxy/api/serializers/influx.py

@@ -0,0 +1,188 @@
+# (c) 2012-2018, Ansible by Red Hat
+#
+# This file is part of Ansible Galaxy
+#
+# Ansible Galaxy is free software: you can redistribute it and/or modify
+# it under the terms of the Apache License as published by
+# the Apache Software Foundation, either version 2 of the License, or
+# (at your option) any later version.
+#
+# Ansible Galaxy is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# Apache License for more details.
+#
+# You should have received a copy of the Apache License
+# along with Galaxy.  If not, see <http://www.apache.org/licenses/>.
+
+from . import serializers
+from rest_framework import serializers as drf_serializers
+from django.conf import settings
+
+import influxdb
+
+from galaxy.main import models
+
+__all__ = (
+    'InfluxSessionSerializer',
+    'InfluxTypes',
+)
+
+
+class InfluxSessionSerializer(serializers.BaseSerializer):
+    class Meta:
+        model = models.InfluxSessionIdentifier
+        fields = (
+            'session_id',
+        )
+
+    def get_url(self, obj):
+        return ''
+
+    def get_summary_fields(self, instance):
+        return {}
+
+
+####################################################
+# Influx "Schema"
+####################################################
+# Influx doesn't have set schema's like most relational databases have, so it's
+# up to the programmer to enforce some level of strcture to the data at the
+# application level. Galaxy is currently using Django Rest Framework
+# to verify that the shape of the data is correct before we insert it into
+# influx. This means that the following serializers are roughly equivalent to
+# models, and should be changed as little as possible to ensure data
+# consistency.
+
+
+# Influx data is organized into 'measurements' which are roughly equivalent to
+# database tables. Each measurement contains at least one field and zero to
+# many tags. Fields and tags are similar to columns with a few caveats.
+class BaseMeasurement(drf_serializers.Serializer):
+    measurement = drf_serializers.CharField()
+
+    def save(self):
+        client = influxdb.InfluxDBClient(
+            host=settings.INFLUX_DB_HOST,
+            port=settings.INFLUX_DB_PORT,
+            username=settings.INFLUX_DB_USERNAME,
+            password=settings.INFLUX_DB_PASSWORD
+        )
+
+        client.switch_database(settings.INFLUX_DB_UI_EVENTS_DB_NAME)
+
+        try:
+            client.write_points([self.data])
+        except influxdb.client.InfluxDBClientError:
+            client.create_database(settings.INFLUX_DB_UI_EVENTS_DB_NAME)
+            client.switch_database(settings.INFLUX_DB_UI_EVENTS_DB_NAME)
+            client.write_points([self.data])
+
+
+# Tags are indexed and only support string types. They should be used grouping
+# data that will be queried frequently, but should not be used for variables
+# that can have a large number of values (search keywords, repository names,
+# sessions etc). Data can be grouped by tags, but the number of distinct tags
+# cannot be easily counted.
+class BaseTags(drf_serializers.Serializer):
+    current_component = drf_serializers.CharField(allow_blank=True)
+
+
+# Fields are not indexed, but can be numeric or string types. Database
+# functions such as DISTINCT, COUNT, SUM, AVERAGE etc can be used on fields.
+class BaseFields(drf_serializers.Serializer):
+    session_id = drf_serializers.UUIDField()
+    current_page = drf_serializers.CharField()
+
+
+# The influx library expects data to be shaped like so:
+#
+# {
+#     "measurement": "measurment_name",
+#     "tags": {
+#         "tag1": "value",
+#     },
+#     "fields": {
+#         "field1": 1
+#     }
+# }
+
+# Page Load
+class PageLoadTags(BaseTags):
+    from_component = drf_serializers.CharField(allow_blank=True)
+
+
+class PageLoadFields(BaseFields):
+    load_time = drf_serializers.FloatField()
+    from_page = drf_serializers.CharField()
+
+
+class PageLoadMeasurementSerializer(BaseMeasurement):
+    tags = PageLoadTags()
+    fields = PageLoadFields()
+
+
+# Search
+class SearchQueryTags(BaseTags):
+    cloud_platforms = drf_serializers.CharField(
+        required=False, allow_blank=True
+    )
+    vendor = drf_serializers.BooleanField(required=False)
+    deprecated = drf_serializers.BooleanField(required=False)
+    content_type = drf_serializers.CharField(required=False, allow_blank=True)
+    platforms = drf_serializers.CharField(required=False, allow_blank=True)
+    tags = drf_serializers.CharField(required=False, allow_blank=True)
+    order_by = drf_serializers.CharField(required=False, allow_blank=True)
+
+
+class SearchQueryFields(BaseFields):
+    keywords = drf_serializers.CharField(required=False, allow_blank=True)
+    namespaces = drf_serializers.CharField(required=False, allow_blank=True)
+    number_of_results = drf_serializers.IntegerField()
+
+
+class SearchQueryMeasurementSerializer(BaseMeasurement):
+    tags = SearchQueryTags()
+    fields = SearchQueryFields()
+
+
+class SearchLinkFields(BaseFields):
+    content_clicked = drf_serializers.CharField()
+    position_in_results = drf_serializers.IntegerField()
+    download_rank = drf_serializers.FloatField()
+    search_rank = drf_serializers.FloatField()
+    relevance = drf_serializers.FloatField()
+    keywords = drf_serializers.CharField(required=False, allow_blank=True)
+
+
+class SearchLinkMeasurementSerializer(BaseMeasurement):
+    tags = SearchQueryTags()
+    fields = SearchLinkFields()
+
+
+# UI Interactions
+class ClickFields(BaseFields):
+    name = drf_serializers.CharField()
+
+
+class LinkClickFields(ClickFields):
+    href = drf_serializers.CharField(allow_blank=True)
+
+
+class ButtonClickMeasurementSerializer(BaseMeasurement):
+    tags = BaseTags()
+    fields = ClickFields()
+
+
+class LinkClickMeasurementSerializer(BaseMeasurement):
+    tags = BaseTags()
+    fields = LinkClickFields()
+
+
+InfluxTypes = {
+    'page_load': PageLoadMeasurementSerializer,
+    'search_query': SearchQueryMeasurementSerializer,
+    'search_click': SearchLinkMeasurementSerializer,
+    'button_click': ButtonClickMeasurementSerializer,
+    'link_click': LinkClickMeasurementSerializer
+}

galaxy/api/urls.py

@@ -19,6 +19,13 @@
 
 from galaxy.api import views
 
+event_tracking_urls = [
+    url(r'^influx_session/$',
+        views.InfluxSession.as_view(), name='influx_session'),
+    url(r'^$',
+        views.InfluxMetrics.as_view(), name='influx_submit'),
+]
+
 email_urls = [
     url(r'^$', views.EmailList.as_view(), name='email_list'),
     url(r'^(?P<pk>[0-9]+)/$',
@@ -271,7 +278,12 @@
     url(r'^community_surveys/', include(community_survey_urls)),
 ]
 
+internal_urls = [
+    url(r'^events/', include(event_tracking_urls)),
+]
+
 urlpatterns = [
     url(r'^$', views.ApiRootView.as_view(), name='api_root_view'),
     url(r'^v1/', include(v1_urls)),
+    url(r'^internal/', include(internal_urls)),
 ]

galaxy/api/views/__init__.py

@@ -32,3 +32,4 @@
 from .users import *                # noqa
 from .views import *                # noqa
 from .survey import *               # noqa
+from .influx import *               # noqa

galaxy/api/views/influx.py

@@ -0,0 +1,99 @@
+# (c) 2012-2018, Ansible by Red Hat
+#
+# This file is part of Ansible Galaxy
+#
+# Ansible Galaxy is free software: you can redistribute it and/or modify
+# it under the terms of the Apache License as published by
+# the Apache Software Foundation, either version 2 of the License, or
+# (at your option) any later version.
+#
+# Ansible Galaxy is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# Apache License for more details.
+#
+# You should have received a copy of the Apache License
+# along with Galaxy.  If not, see <http://www.apache.org/licenses/>.
+
+import logging
+import datetime
+
+from galaxy.main import models
+from galaxy.api import serializers
+
+from . import base_views
+
+from rest_framework.response import Response
+from rest_framework import views
+
+logger = logging.getLogger(__name__)
+
+__all__ = [
+    'InfluxSession',
+    'InfluxMetrics'
+]
+
+
+def set_cookie(response, session):
+    expiration = datetime.datetime.utcnow() + datetime.timedelta(hours=1)
+
+    response.set_cookie(
+        'influx_session',
+        session,
+        expires=expiration
+    )
+
+    return response
+
+
+class InfluxSession(base_views.ListCreateAPIView):
+    model = models.InfluxSessionIdentifier
+    serializer_class = serializers.InfluxSessionSerializer
+
+    def post(self, request, *args, **kwargs):
+        influx_session = ''
+        if not request.COOKIES.get('influx_session'):
+            influx_session = self.model.objects.create()
+        else:
+            influx_session = models.InfluxSessionIdentifier.objects.get(
+                session_id=request.COOKIES.get('influx_session')
+            )
+
+        serializer = self.get_serializer(instance=influx_session)
+        headers = self.get_success_headers(serializer.data)
+
+        response = set_cookie(
+            Response(serializer.data, headers=headers),
+            influx_session.session_id
+        )
+
+        return response
+
+    def get(self, request, *args, **kwargs):
+        return ''
+
+
+class InfluxMetrics(views.APIView):
+    def get(self, request):
+        # TODO: Return real response
+        return Response('HOWDY')
+
+    def post(self, request):
+        serializer = self.load_serializer(request)
+        if serializer:
+            if serializer.is_valid(raise_exception=True):
+                serializer.save()
+            return Response(serializer.data)
+
+        # TODO: Return real response
+        return Response('NO GO')
+
+    # Can't name this get_serializer() for some reason
+    def load_serializer(self, request):
+        if request.data['measurement'] in serializers.InfluxTypes:
+            serializer_type = serializers.InfluxTypes[
+                request.data['measurement']
+            ]
+            return serializer_type(data=request.data)
+        else:
+            return None

galaxy/main/migrations/0118_influxsessionidentifier.py

@@ -0,0 +1,34 @@
+# -*- coding: utf-8 -*-
+# Generated by Django 1.11.15 on 2018-09-24 18:24
+from __future__ import unicode_literals
+
+from django.db import migrations, models
+import galaxy.main.mixins
+import uuid
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('main', '0117_namespace_search_index'),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name='InfluxSessionIdentifier',
+            fields=[
+                ('created', models.DateTimeField(auto_now_add=True)),
+                ('modified', models.DateTimeField(auto_now=True)),
+                ('session_id', models.UUIDField(
+                    default=uuid.uuid4,
+                    editable=False,
+                    primary_key=True,
+                    serialize=False
+                )),
+            ],
+            options={
+                'abstract': False,
+            },
+            bases=(models.Model, galaxy.main.mixins.DirtyMixin),
+        ),
+    ]