Python Integration
Click here for the Posthog Python Library. This is an optional library you can install if you're working with Python.
This page of the Docs refers specifically to the Official PostHog Python Library to capture and send events to any PostHog instance (including posthog.com).
This library uses an internal queue to make calls fast and non-blocking. It also batches requests and flushes asynchronously, making it perfect to use in any part of your web app or other server side application that needs performance.
Installation
pip install posthog
In your app, import the posthog
library and set your api key and host before making any calls.
import posthog# Substitutes posthog.api_key which still exists but has been deprecatedposthog.project_api_key = '<ph_project_api_key>'# Only necessary if you want to use feature flagsposthog.personal_api_key = '<ph_personal_api_key>'# You can remove this line if you're using app.posthog.composthog.host = '<ph_instance_address>'
You can read more about the differences between the project and personal API keys in the dedicated API Authentication section of the Docs.
Note: As a general rule of thumb, we do not recommend having API keys in plaintext. Setting it as an environment variable would be best.
You can find your keys in the 'Project Settings' page in PostHog.
To debug, you can toggle debug mode on:
posthog.debug = True
And to make sure no calls happen during your tests, you can disable them, like so:
if settings.TEST:posthog.disabled = True
Making Calls
Capture
Capture allows you to capture anything a user does within your system, which you can later use in PostHog to find patterns in usage, work out which features to improve or where people are giving up.
A capture
call requires:
distinct id
which uniquely identifies your userevent name
to specify the event
- We recommend naming events with "[noun][verb]", such as
movie played
ormovie updated
, in order to easily identify what your events mean later on (we know this from experience).
Optionally you can submit:
properties
, which is a dictionary with any information you'd like to addtimestamp
, a datetime object for when the event happened. If this isn't submitted, it'll be set to the current time
For example:
posthog.capture('distinct id', 'movie played', {'movie_id': '123', 'category': 'romcom'})
or
posthog.capture('distinct id', event='movie played', properties={'movie_id': '123', 'category': 'romcom'}, timestamp=datetime.utcnow().replace(tzinfo=tzutc())
Identify
Identify lets you add metadata to your users so you can easily identify who they are in PostHog, as well as do things like segment users by these properties.
An identify
call requires:
distinct id
which uniquely identifies your userproperties
with a dict with any key:value pairs
For example:
posthog.identify('distinct id', {'email': 'dwayne@gmail.com','name': 'Dwayne Johnson'})
The most obvious place to make this call is whenever a user signs up, or when they update their information.
Alias
To connect whatever a user does before they sign up or log in with what they do after you need to make an alias call. This will allow you to answer questions like "Which marketing channels leads to users churning after a month?" or "What do users do on our website before signing up?"
In a purely back-end implementation, this means whenever an anonymous user does something, you'll want to send a session ID (Django, Flask) with the capture call. Then, when that users signs up, you want to do an alias call with the session ID and the newly created user ID.
The same concept applies for when a user logs in.
If you're using PostHog in the front-end and back-end, doing the identify
call in the frontend will be enough.
An alias
call requires:
previous distinct id
the unique ID of the user beforedistinct id
the current unique id
For example:
posthog.alias('anonymous session id', 'distinct id')
Feature Flags
Note: To use feature flags you must also set a
personal_api_key
when configuring the integration, as described in the Installation section.
PostHog's Feature Flags allow you to safely deploy and roll back new features.
When using them with one of libraries, you should check if a feature flag is enabled and use the result to toggle functionality on and off in you application.
How to check if a flag is enabled
posthog.feature_enabled('beta-feature', 'distinct id')
Example Use Case
Here's how you might send different users a different version of your homepage, for example:
def homepage(request):template = "new.html" if posthog.feature_enabled('new_ui', 'distinct id') else "old.html"return render_template(template, request=request)
Note: Feature flags are persistent for users across sessions. Read more about feature flag persistence on our dedicated page.
Sending Page Views
If you're aiming for a full back-end implementation of PostHog, you can send pageviews
from your backend
posthog.capture('distinct id', '$pageview', {'$current_url': 'https://example.com'})
Django
For Django, you can do the initialisation of the key in the AppConfig, so that it's available everywhere.
in yourapp/apps.py
from django.apps import AppConfigimport posthogclass YourAppConfig(AppConfig):def ready(self):posthog.api_key = '<ph_project_api_key>'posthog.host = '<ph_instance_address>' # You can remove this line if you're using app.posthog.com
Then, anywhere else in your app you can do:
import posthogdef purchase(request):# example captureposthog.capture(request.user.pk, 'purchase', ...)
Integrations
Sentry
When using Sentry in Python, you can connect to PostHog in order to link Sentry errors to PostHog user profiles.
Example Implementation
from posthog.sentry import update_posthogsentry_sdk.init(dsn="https://examplePublicKey@o0.ingest.sentry.io/0",before_send=update_posthog)# Also set `posthog_distinct_id` tagfrom sentry_sdk import configure_scopewith configure_scope() as scope:scope.set_tag('posthog_distinct_id', 'some distinct id')
Example Implementation with Django
This can be made automatic in Dejango, by adding the following middleware and settings to settings.py
:
MIDDLEWARE = ["posthog.sentry.django.PosthogDistinctIdMiddleware"]POSTHOG_DJANGO = {"distinct_id": lambda request: request.user and request.user.distinct_id}
Development
Naming Confusion
As our open source project PostHog shares the same module name, we created a special posthoganalytics
package, mostly for internal use to avoid module collision. It is the exact same.
Thank You
This library is largely based on the analytics-python
package.
How to release
Instructions on how to release a new version of this package have moved to our Handbook.