A simple python wrapper for the Firebase API.

pip install pyrebase

Pyrebase was written for python 3 and will not work correctly with python 2.

For use with only user based authentication we can create the following configuration:

import pyrebase

config = {
  "apiKey": "apiKey",
  "authDomain": "projectId.firebaseapp.com",
  "databaseURL": "https://databaseName.firebaseio.com",
  "storageBucket": "projectId.appspot.com"
}

firebase = pyrebase.initialize_app(config)

We can optionally add a service account credential to our configuration that will allow our server to authenticate with Firebase as an admin and disregard any security rules.

import pyrebase

config = {
  "apiKey": "apiKey",
  "authDomain": "projectId.firebaseapp.com",
  "databaseURL": "https://databaseName.firebaseio.com",
  "storageBucket": "projectId.appspot.com",
  "serviceAccount": "path/to/serviceAccountCredentials.json"
}

firebase = pyrebase.initialize_app(config)

Adding a service account will authenticate as an admin by default for all database queries, check out the Authentication documentation for how to authenticate users.

A Pyrebase app can use multiple Firebase services.

firebase.auth() - Authentication

firebase.database() - Database

firebase.storage() - Storage

Check out the documentation for each service for further details.

The sign_in_with_email_and_password() method will return user data including a token you can use to adhere to security rules.

Each of the following methods accepts a user token: get(), push(), set(), update(), remove() and stream().

# Get a reference to the auth service
auth = firebase.auth()

# Log the user in
user = auth.sign_in_with_email_and_password(email, password)

# Get a reference to the database service
db = firebase.database()

# data to save
data = {
    "name": "Mortimer 'Morty' Smith"
}

# Pass the user's idToken to the push method
results = db.child("users").push(data, user['idToken'])

A user's idToken expires after 1 hour, so be sure to use the user's refreshToken to avoid stale tokens.

user = auth.sign_in_with_email_and_password(email, password)
# before the 1 hour expiry:
user = auth.refresh(user['refreshToken'])
# now we have a fresh token
user['idToken']

You can also create users using custom tokens, for example:

token = auth.create_custom_token("your_custom_id")

You can also pass in additional claims.

token_with_additional_claims = auth.create_custom_token("your_custom_id", {"premium_account": True})

You can then send these tokens to the client to sign in, or sign in as the user on the server.

user = auth.sign_in_with_custom_token(token)

auth.create_user_with_email_and_password(email, password)

Note: Make sure you have the Email/password provider enabled in your Firebase dashboard under Auth -> Sign In Method.

auth.send_email_verification(user['idToken'])

auth.send_password_reset_email("email")

auth.get_account_info(user['idToken'])

user = auth.refresh(user['refreshToken'])

You can build paths to your data by using the child() method.

db = firebase.database()
db.child("users").child("Morty")

To save data with a unique, auto-generated, timestamp-based key, use the push() method.

data = {"name": "Mortimer 'Morty' Smith"}
db.child("users").push(data)

To create your own keys use the set() method. The key in the example below is "Morty".

data = {"name": "Mortimer 'Morty' Smith"}
db.child("users").child("Morty").set(data)

To update data for an existing entry use the update() method.

db.child("users").child("Morty").update({"name": "Mortiest Morty"})

To delete data for an existing entry use the remove() method.

db.child("users").child("Morty").remove()

You can also perform multi-location updates with the update() method.

data = {
    "users/Morty/": {
        "name": "Mortimer 'Morty' Smith"
    },
    "users/Rick/": {
        "name": "Rick Sanchez"
    }
}

db.update(data)

To perform multi-location writes to new locations we can use the generate_key() method.

data = {
    "users/"+ref.generate_key(): {
        "name": "Mortimer 'Morty' Smith"
    },
    "users/"+ref.generate_key(): {
        "name": "Rick Sanchez"
    }
}

db.update(data)

Queries return a PyreResponse object. Calling val() on these objects returns the query data.

users = db.child("users").get()
print(users.val()) # {"Morty": {"name": "Mortimer 'Morty' Smith"}, "Rick": {"name": "Rick Sanchez"}}

Calling key() returns the key for the query data.

user = db.child("users").get()
print(user.key()) # users

Returns a list of objects on each of which you can call val() and key().

all_users = db.child("users").get()
for user in all_users.each():
    print(user.key()) # Morty
    print(user.val()) # {name": "Mortimer 'Morty' Smith"}

To return data from a path simply call the get() method.

all_users = db.child("users").get()

To return just the keys at a particular path use the shallow() method.

all_user_ids = db.child("users").shallow().get()

Note: shallow() can not be used in conjunction with any complex queries.

You can listen to live changes to your data with the stream() method.

def stream_handler(message):
    print(message["event"]) # put
    print(message["path"]) # /-K7yGTTEp7O549EzTYtI
    print(message["data"]) # {'title': 'Pyrebase', "body": "etc..."}

my_stream = db.child("posts").stream(stream_handler)

You should at least handle put and patch events. Refer to "Streaming from the REST API" for details.

You can also add a stream_id to help you identify a stream if you have multiple running:

my_stream = db.child("posts").stream(stream_handler, stream_id="new_posts")

my_stream.close()

Queries can be built by chaining multiple query parameters together.

users_by_name = db.child("users").order_by_child("name").limit_to_first(3).get()

This query will return the first three users ordered by name.

We begin any complex query with order_by_child().

users_by_name = db.child("users").order_by_child("name").get()

This query will return users ordered by name.

Return data with a specific value.

users_by_score = db.child("users").order_by_child("score").equal_to(10).get()

This query will return users with a score of 10.

Specify a range in your data.

users_by_score = db.child("users").order_by_child("score").start_at(3).end_at(10).get()

This query returns users ordered by score and with a score between 3 and 10.

Limits data returned.

users_by_score = db.child("users").order_by_child("score").limit_to_first(5).get()

This query returns the first five users ordered by score.

When using order_by_key() to sort your data, data is returned in ascending order by key.

users_by_key = db.child("users").order_by_key().get()

When using order_by_value(), children are ordered by their value.

users_by_value = db.child("users").order_by_value().get()

The storage service allows you to upload images to Firebase.

Just like with the Database service, you can build paths to your data with the Storage service.

storage.child("images/example.jpg")

The put method takes the path to the local file and an optional user token.

storage = firebase.storage()
# as admin
storage.child("images/example.jpg").put("example2.jpg")
# as user
storage.child("images/example.jpg").put("example2.jpg", user['idToken'])

The download method takes the path to the saved database file and the name you want the downloaded file to have.

storage.child("images/example.jpg").download("downloaded.jpg")

The get_url method takes the path to the saved database file and returns the storage url.

storage.child("images/example.jpg").get_url()
# https://firebasestorage.googleapis.com/v0/b/storage-url.appspot.com/o/images%2Fexample.jpg?alt=media

db.generate_key() is an implementation of Firebase's key generation algorithm.

See multi-location updates for a potential use case.

Sometimes we might want to sort our data multiple times. For example, we might want to retrieve all articles written between a certain date then sort those articles based on the number of likes.

Currently the REST API only allows us to sort our data once, so the sort() method bridges this gap.

articles = db.child("articles").order_by_child("date").start_at(startDate).end_at(endDate).get()
articles_by_likes = db.sort(articles, "likes")

Indexing is not enabled for the database reference.