The python serverless microframework for AWS allows you to quickly create and deploy applications that use Amazon API Gateway and AWS Lambda. It provides:
- A command line tool for creating, deploying, and managing your app
- A familiar and easy to use API for declaring views in python code
- Automatic IAM policy generation
$ pip install chalice $ chalice new-project helloworld && cd helloworld $ cat app.py from chalice import Chalice app = Chalice(app_name="helloworld") @app.route("/") def index(): return {"hello": "world"} $ chalice deploy ... Your application is available at: https://endpoint/dev $ curl https://endpoint/dev {"hello": "world"}
Up and running in less than 30 seconds.
This project is published as a preview project. Until the 1.0 GA release, we may introduce backwards incompatible changes based on customer feedback we receive. As a best practice, we do not recommend launching production APIs until 1.0 GA because of potential backward incompatible releases. Please see the CHANGELOG.rst file to see the changes for each release.
Give this project a try and share your feedback with us here on Github.
The documentation is available on readthedocs.
In this tutorial, you'll use the chalice
command line utility
to create and deploy a basic REST API.
First, you'll need to install chalice
. Using a virtualenv
is recommended:
$ pip install virtualenv $ virtualenv ~/.virtualenvs/chalice-demo $ source ~/.virtualenvs/chalice-demo/bin/activate
Note: make sure you are using python2.7. The chalice
CLI
as well as the chalice
python package will support the versions
of python supported by AWS Lambda. Currently, AWS Lambda only supports
python2.7, so this is what this project supports. You can ensure
you're creating a virtualenv with python2.7 by running:
# Double check you have python2.7 $ which python2.7 /usr/local/bin/python2.7 $ virtualenv --python $(which python2.7) ~/.virtualenvs/chalice-demo $ source ~/.virtualenvs/chalice-demo/bin/activate
Next, in your virtualenv, install chalice
:
$ pip install chalice
You can verify you have chalice installed by running:
$ chalice --help Usage: chalice [OPTIONS] COMMAND [ARGS]... ...
Before you can deploy an application, be sure you have credentials configured. If you have previously configured your machine to run boto3 (the AWS SDK for Python) or the AWS CLI then you can skip this section.
If this is your first time configuring credentials for AWS you can follow these steps to quickly get started:
$ mkdir ~/.aws $ cat >> ~/.aws/config [default] aws_access_key_id=YOUR_ACCESS_KEY_HERE aws_secret_access_key=YOUR_SECRET_ACCESS_KEY region=YOUR_REGION (such as us-west-2, us-west-1, etc)
If you want more information on all the supported methods for configuring credentials, see the boto3 docs.
The next thing we'll do is use the chalice
command to create a new
project:
$ chalice new-project helloworld
This will create a helloworld
directory. Cd into this
directory. You'll see several files have been created for you:
$ cd helloworld $ ls -la drwxr-xr-x .chalice -rw-r--r-- app.py -rw-r--r-- requirements.txt
You can ignore the .chalice
directory for now, the two main files
we'll focus on is app.py
and requirements.txt
.
Let's take a look at the app.py
file:
from chalice import Chalice
app = Chalice(app_name='helloworld')
@app.route('/')
def index():
return {'hello': 'world'}
The new-project
command created a sample app that defines a
single view, /
, that when called will return the JSON body
{"hello": "world"}
.
Let's deploy this app. Make sure you're in the helloworld
directory and run chalice deploy
:
$ chalice deploy ... Initiating first time deployment... https://qxea58oupc.execute-api.us-west-2.amazonaws.com/dev/
You now have an API up and running using API Gateway and Lambda:
$ curl https://qxea58oupc.execute-api.us-west-2.amazonaws.com/dev/ {"hello": "world"}
Try making a change to the returned dictionary from the index()
function. You can then redeploy your changes by running chalice deploy
.
For the rest of these tutorials, we'll be using httpie
instead of curl
(https://github.com/jkbrzt/httpie) to test our API. You can install httpie
using pip install httpie
, or if you're on Mac, you can run brew install
httpie
. The Github link has more information on installation instructions.
Here's an example of using httpie
to request the root resource of the API
we just created. Note that the command name is http
:
$ http https://qxea58oupc.execute-api.us-west-2.amazonaws.com/dev/ HTTP/1.1 200 OK Connection: keep-alive Content-Length: 18 Content-Type: application/json Date: Mon, 30 May 2016 17:55:50 GMT X-Cache: Miss from cloudfront { "hello": "world" }
Additionally, the API Gateway endpoints will be shortened to
https://endpoint/dev/
for brevity. Be sure to substitute
https://endpoint/dev/
for the actual endpoint that the chalice
CLI displays when you deploy your API (it will look something like
https://abcdefg.execute-api.us-west-2.amazonaws.com/dev/
.
You've now created your first app using chalice
.
The next few sections will build on this quickstart section and introduce you to additional features including: URL parameter capturing, error handling, advanced routing, current request metadata, and automatic policy generation.
Now we're going to make a few changes to our app.py
file that
demonstrate additional capabilities provided by the python serverless
microframework for AWS.
Our application so far has a single view that allows you to make
an HTTP GET request to /
. Now let's suppose we want to capture
parts of the URI:
from chalice import Chalice
app = Chalice(app_name='helloworld')
CITIES_TO_STATE = {
'seattle': 'WA',
'portland': 'OR',
}
@app.route('/')
def index():
return {'hello': 'world'}
@app.route('/cities/{city}')
def state_of_city(city):
return {'state': CITIES_TO_STATE[city]}
In the example above, we've now added a state_of_city
view that allows
a user to specify a city name. The view function takes the city
name and returns name of the state the city is in. Notice that the
@app.route
decorator has a URL pattern of /cities/{city}
. This
means that the value of {city}
is captured and passed to the view
function. You can also see that the state_of_city
takes a single
argument. This argument is the name of the city provided by the user.
For example:
GET /cities/seattle --> state_of_city('seattle') GET /cities/portland --> state_of_city('portland')
Now that we've updated our app.py
file with this new view function,
let's redeploy our application. You can run chalice deploy
from
the helloworld
directory and it will deploy your application:
$ chalice deploy
Let's try it out. Note the examples below use the http
command
from the httpie
package. You can install this using pip install httpie
:
$ http https://endpoint/dev/cities/seattle HTTP/1.1 200 OK { "state": "WA" } $ http https://endpoint/dev/cities/portland HTTP/1.1 200 OK { "state": "OR" }
Notice what happens if we try to request a city that's not in our
CITIES_TO_STATE
map:
$ http https://endpoint/dev/cities/vancouver HTTP/1.1 500 Internal Server Error Content-Type: application/json X-Cache: Error from cloudfront { "Code": "ChaliceViewError", "Message": "ChaliceViewError: An internal server error occurred." }
In the next section, we'll see how to fix this and provide better error messages.
In the example above, you'll notice that when our app raised an uncaught exception, a 500 internal server error was returned.
In this section, we're going to show how you can debug and improve these error messages.
The first thing we're going to look at is how we can debug this issue. By default, debugging is turned off, but you can enable debugging to get more information:
from chalice import Chalice
app = Chalice(app_name='helloworld')
app.debug = True
The app.debug = True
enables debugging for your app.
Save this file and redeploy your changes:
$ chalice deploy ... https://endpoint/dev/
Now, when you request the same URL that returned an internal server error, you'll get back the original stack trace:
$ http https://endpoint/dev/cities/vancouver Traceback (most recent call last): File "/var/task/chalice/app.py", line 304, in _get_view_function_response response = view_function(*function_args) File "/var/task/app.py", line 18, in state_of_city return {'state': CITIES_TO_STATE[city]} KeyError: u'vancouver'
We can see that the error is caused from an uncaught KeyError
resulting
from trying to access the vancouver
key.
Now that we know the error, we can fix our code. What we'd like to do is catch this exception and instead return a more helpful error message to the user. Here's the updated code:
from chalice import BadRequestError
@app.route('/cities/{city}')
def state_of_city(city):
try:
return {'state': CITIES_TO_STATE[city]}
except KeyError:
raise BadRequestError("Unknown city '%s', valid choices are: %s" % (
city, ', '.join(CITIES_TO_STATE.keys())))
Save and deploy these changes:
$ chalice deploy $ http https://endpoint/dev/cities/vancouver HTTP/1.1 400 Bad Request { "Code": "BadRequestError", "Message": "Unknown city 'vancouver', valid choices are: portland, seattle" }
We can see now that we have received a Code
and Message
key, with the message
being the value we passed to BadRequestError
. Whenever you raise
a BadRequestError
from your view function, the framework will return an
HTTP status code of 400 along with a JSON body with a Code
and Message
.
There are a few additional exceptions you can raise from your python code:
* BadRequestError - return a status code of 400 * UnauthorizedError - return a status code of 401 * ForbiddenError - return a status code of 403 * NotFoundError - return a status code of 404 * ConflictError - return a status code of 409 * TooManyRequestsError - return a status code of 429 * ChaliceViewError - return a status code of 500
You can import these directly from the chalice
package:
from chalice import UnauthorizedError
So far, our examples have only allowed GET requests. It's actually possible to support additional HTTP methods. Here's an example of a view function that supports PUT:
@app.route('/resource/{value}', methods=['PUT'])
def put_test(value):
return {"value": value}
We can test this method using the http
command:
$ http PUT https://endpoint/dev/resource/foo HTTP/1.1 200 OK { "value": "foo" }
Note that the methods
kwarg accepts a list of methods. Your view function
will be called when any of the HTTP methods you specify are used for the
specified resource. For example:
@app.route('/myview', methods=['POST', 'PUT'])
def myview():
pass
The above view function will be called when either an HTTP POST or
PUT is sent to /myview
. In the next section we'll go over
how you can introspect the given request in order to differentiate between
various HTTP methods.
In the examples above, you saw how to create a view function that supports an HTTP PUT request as well as a view function that supports both POST and PUT via the same view function. However, there's more information we might need about a given request:
- In a PUT/POST, you frequently send a request body. We need some way of accessing the contents of the request body.
- For view functions that support multiple HTTP methods, we'd like to detect which HTTP method was used so we can have different code paths for PUTs vs. POSTs.
All of this and more is handled by the current request object that the
chalice
library makes available to each view function when it's called.
Let's see an example of this. Suppose we want to create a view function that allowed you to PUT data to an object and retrieve that data via a corresponding GET. We could accomplish that with the following view function:
from chalice import NotFoundError
OBJECTS = {
}
@app.route('/objects/{key}', methods=['GET', 'PUT'])
def myobject(key):
request = app.current_request
if request.method == 'PUT':
OBJECTS[key] = request.json_body
elif request.method == 'GET':
try:
return {key: OBJECTS[key]}
except KeyError:
raise NotFoundError(key)
Save this in your app.py
file and rerun chalice deploy
.
Now, you can make a PUT request to /objects/your-key
with a request
body, and retrieve the value of that body by making a subsequent
GET
request to the same resource. Here's an example of its usage:
# First, trying to retrieve the key will return a 404. $ http GET https://endpoint/dev/objects/mykey HTTP/1.1 404 Not Found { "Code": "NotFoundError", "Message": "mykey" } # Next, we'll create that key by sending a PUT request. $ echo '{"foo": "bar"}' | http PUT https://endpoint/dev/objects/mykey HTTP/1.1 200 OK null # And now we no longer get a 404, we instead get the value we previously # put. $ http GET https://endpoint/dev/objects/mykey HTTP/1.1 200 OK { "mykey": { "foo": "bar" } }
You might see a problem with storing the objects in a module level
OBJECTS
variable. We address this in the next section.
The app.current_request
object also has the following properties.
current_request.query_params
- A dict of the query params for the request.current_request.headers
- A dict of the request headers.current_request.uri_params
- A dict of the captured URI params.current_request.method
- The HTTP method (as a string).current_request.json_body
- The parsed JSON body (json.loads(raw_body)
)current_request.raw_body
- The raw HTTP body as bytes.current_request.context
- A dict of additional context informationcurrent_request.stage_vars
- Configuration for the API Gateway stage
Don't worry about the context
and stage_vars
for now. We haven't
discussed those concepts yet. The current_request
object also
has a to_dict
method, which returns all the information about the
current request as a dictionary. Let's use this method to write a view
function that returns everything it knows about the request:
@app.route('/introspect')
def introspect():
return app.current_request.to_dict()
Save this to your app.py
file and redeploy with chalice deploy
.
Here's an example of hitting the /introspect
URL. Note how we're
sending a query string as well as a custom X-TestHeader
header:
$ http 'https://endpoint/dev/introspect?query1=value1&query2=value2' 'X-TestHeader: Foo' HTTP/1.1 200 OK { "context": { "apiId": "apiId", "httpMethod": "GET", "identity": { "accessKey": null, "accountId": null, "apiKey": null, "caller": null, "cognitoAuthenticationProvider": null, "cognitoAuthenticationType": null, "cognitoIdentityId": null, "cognitoIdentityPoolId": null, "sourceIp": "1.1.1.1", "userAgent": "HTTPie/0.9.3", "userArn": null }, "requestId": "request-id", "resourceId": "resourceId", "resourcePath": "/introspect", "stage": "dev" }, "headers": { "accept": "*/*", ... "x-testheader": "Foo" }, "method": "GET", "query_params": { "query1": "value1", "query2": "value2" }, "raw_body": null, "stage_vars": null, "uri_params": null }
The default behavior of a view function supports
a request body of application/json
. When a request is
made with a Content-Type
of application/json
, the
app.current_request.json_body
attribute is automatically
set for you. This value is the parsed JSON body.
You can also configure a view function to support other
content types. You can do this by specifying the
content_types
paramter value to your app.route
function. This parameter is a list of acceptable content
types. Here's an example of this feature:
import urlparse
from chalice import Chalice
app = Chalice(app_name='helloworld')
@app.route('/', methods=['POST'],
content_types=['application/x-www-form-urlencoded'])
def index():
parsed = urlparse.parse_qs(app.current_request.raw_body)
return {
'states': parsed.get('states', [])
}
There's a few things worth noting in this view function.
First, we've specified that we only accept the
application/x-www-form-urlencoded
content type. If we
try to send a request with application/json
, we'll now
get a 415 Unsupported Media Type
response:
$ http POST https://endpoint/dev/ states=WA states=CA --debug ... >>> requests.request(**{'allow_redirects': False, 'headers': {'Accept': 'application/json', 'Content-Type': 'application/json', ... HTTP/1.1 415 Unsupported Media Type { "message": "Unsupported Media Type" }
If we use the --form
argument, we can see the
expected behavior of this view function because httpie
sets the
Content-Type
header to application/x-www-form-urlencoded
:
$ http --form POST https://endpoint/dev/formtest states=WA states=CA --debug ... >>> requests.request(**{'allow_redirects': False, 'headers': {'Content-Type': 'application/x-www-form-urlencoded; charset=utf-8', ... HTTP/1.1 200 OK { "states": [ "WA", "CA" ] }
The second thing worth noting is that app.current_request.json_body
is only available for the application/json content type.
In our example above, we used app.current_request.raw_body
to access
the raw body bytes:
parsed = urlparse.parse_qs(app.current_request.raw_body)
app.current_request.json_body
is set to None
whenever the
Content-Type
is not application/json
. This means that
you will need to use app.current_request.raw_body
and parse
the request body as needed.
The return value from a chalice view function is serialized as JSON as the response body returned back to the caller. This makes it easy to create rest APIs that return JSON resonse bodies.
Chalice allows you to control this behavior by returning an instance of
a chalice specific Response
class. This behavior allows you to:
- Specify the status code to return
- Specify custom headers to add to the response
- Specify response bodies that are not
application/json
Here's an example of this:
from chalice import Chalice, Response
app = Chalice(app_name='custom-response')
@app.route('/')
def index():
return Response(body='hello world!',
status_code=200,
headers={'Content-Type': 'text/plain'})
This will result in a plain text response body:
$ http https://endpoint/dev/ HTTP/1.1 200 OK Content-Length: 12 Content-Type: text/plain hello world!
You can specify whether a view supports CORS by adding the
cors=True
parameter to your @app.route()
call. By
default this value is false:
@app.route('/supports-cors', methods=['PUT'], cors=True)
def supports_cors():
return {}
Settings cors=True
has similar behavior to enabling CORS
using the AWS Console. This includes:
- Injecting the
Access-Control-Allow-Origin: *
header to your responses, including all error responses you can return. - Automatically adding an
OPTIONS
method so support preflighting requests.
The preflight request will return a response that includes:
Access-Control-Allow-Origin: *
- The
Access-Control-Allow-Methods
header will return a list of all HTTP methods you've called out in your view function. In the example above, this will bePUT,OPTIONS
. Access-Control-Allow-Headers: Content-Type,X-Amz-Date,Authorization, X-Api-Key,X-Amz-Security-Token
.
There's a couple of things to keep in mind when enabling cors for a view:
- An
OPTIONS
method for preflighting is always injected. Ensure that you don't haveOPTIONS
in themethods=[...]
list of your view function. - Every view function must explicitly enable CORS support.
- There's no support for customizing the CORS configuration.
The last two points will change in the future. See this issue for more information.
In the previous section we created a basic rest API that
allowed you to store JSON objects by sending the JSON
in the body of an HTTP PUT request to /objects/{name}
.
You could then retrieve objects by sending a GET request to
/objects/{name}
.
However, there's a problem with the code we wrote:
OBJECTS = {
}
@app.route('/objects/{key}', methods=['GET', 'PUT'])
def myobject(key):
request = app.current_request
if request.method == 'PUT':
OBJECTS[key] = request.json_body
elif request.method == 'GET':
try:
return {key: OBJECTS[key]}
except KeyError:
raise NotFoundError(key)
We're storing the key value pairs in a module level OBJECTS
variable. We can't rely on local storage like this persisting
across requests.
A better solution would be to store this information in Amazon S3. To do this, we're going to use boto3, the AWS SDK for Python. First, install boto3:
$ pip install boto3
Next, add boto3
to your requirements.txt file:
$ echo 'boto3==1.3.1' >> requirements.txt
The requirements.txt file should be in the same directory that contains
your app.py
file. Next, let's update our view code to use boto3:
import json
import boto3
from botocore.exceptions import ClientError
from chalice import NotFoundError
S3 = boto3.client('s3', region_name='us-west-2')
BUCKET = 'your-bucket-name'
@app.route('/objects/{key}', methods=['GET', 'PUT'])
def s3objects(key):
request = app.current_request
if request.method == 'PUT':
S3.put_object(Bucket=BUCKET, Key=key,
Body=json.dumps(request.json_body))
elif request.method == 'GET':
try:
response = S3.get_object(Bucket=BUCKET, Key=key)
return json.loads(response['Body'].read())
except ClientError as e:
raise NotFoundError(key)
Make sure to change BUCKET
with the name of an S3 bucket
you own. Redeploy your changes with chalice deploy
.
Now, whenever we make a PUT
request to /objects/keyname
, the
data send will be stored in S3. Any subsequent GET
requests will
retrieve this data from S3.
IAM permissions can be auto generated, provided manually or can be pre-created and explicitly configured. To use a pre-configured IAM role ARN for chalice, add these two keys to your chalice configuration. Setting manage_iam_role to false tells Chalice to not attempt to generate policies and create IAM role.
"manage_iam_role":false "iam_role_arn":"arn:aws:iam::<account-id>:role/<role-name>"
Whenever your application is deployed using chalice
, the
auto generated policy is written to disk at
<projectdir>/.chalice/policy.json
. When you run the
chalice deploy
command, you can also specify the
--no-autogen-policy
option. Doing so will result in the
chalice
CLI loading the <projectdir>/.chalice/policy.json
file and using that file as the policy for the IAM role.
You can manually edit this file and specify --no-autogen-policy
if you'd like to have full control over what IAM policy to associate
with the IAM role.
You can also run the chalice gen-policy
command from your project
directory to print the auto generated policy to stdout. You can
then use this as a starting point for your policy.
$ chalice gen-policy { "Version": "2012-10-17", "Statement": [ { "Action": [ "s3:ListAllMyBuckets" ], "Resource": [ "*" ], "Effect": "Allow", "Sid": "9155de6ad1d74e4c8b1448255770e60c" } ] }
The automatic policy generation is still in the early stages, it should
be considered experimental. You can always disable policy
generation with --no-autogen-policy
for complete control.
Additionally, you will be prompted for confirmation whenever the auto policy generator detects actions that it would like to add or remove:
$ chalice deploy Updating IAM policy. The following action will be added to the execution policy: s3:ListBucket Would you like to continue? [Y/n]:
AWS API Gateway routes can be authenticated in multiple ways:
- API Key
- Custom Auth Handler
@app.route('/authenticated', methods=['GET'], api_key_required=True)
def authenticated():
return {"secure": True}
Only requests sent with a valid X-Api-Key header will be accepted.
A custom Authorizer is required for this to work, details can be found here; http://docs.aws.amazon.com/apigateway/latest/developerguide/use-custom-authorizer.html
@app.route('/authenticated', methods=['GET'], authorization_type='CUSTOM', authorizer_id='ab12cd')
def authenticated():
return {"secure": True}
Only requests sent with a valid X-Api-Key header will be accepted.
As you develop your application, you may want to experiment locally before
deploying your changes. You can use chalice local
to spin up a local
HTTP server you can use for testing.
For example, if we have the following app.py
file:
from chalice import Chalice
app = Chalice(app_name='helloworld')
@app.route('/')
def index():
return {'hello': 'world'}
We can run chalice local
to test this API locally:
$ chalice local Serving on localhost:8000
We can override the port using:
$ chalice --port=8080 local
We can now test our API using localhost:8000
:
$ http localhost:8000/ HTTP/1.0 200 OK Content-Length: 18 Content-Type: application/json Date: Thu, 27 Oct 2016 20:08:43 GMT Server: BaseHTTP/0.3 Python/2.7.11 { "hello": "world" }
The chalice local
command does not assume the
role associated with your lambda function, so you'll
need to use an AWS_PROFILE
that has sufficient permissions
to your AWS resources used in your app.py
.
These are features that are in the backlog:
- Adding full support for API gateway stages - issue 20
Please share any feedback on the above issues. We'd also love to hear from you. Please create any Github issues for additional features you'd like to see: https://github.com/awslabs/chalice/issues
Q: How does the Python Serverless Microframework for AWS compare to other similar frameworks?
The biggest difference between this framework and others is that the Python Serverless Microframework for AWS is singularly focused on using a familiar, decorator-based API to write python applications that run on Amazon API Gateway and AWS Lambda. You can think of it as Flask/Bottle for serverless APIs. Its goal is to make writing and deploying these types of applications as simple as possible specifically for Python developers.
To achieve this goal, it has to make certain tradeoffs. Python will always remain the only supported language in this framework. Not every feature of API Gateway and Lambda is exposed in the framework. It makes assumptions about how applications will be deployed, and it has restrictions on how an application can be structured. It does not address the creation and lifecycle of other AWS resources your application may need (Amazon S3 buckets, Amazon DynamoDB tables, etc.). The feature set is purposefully small.
Other full-stack frameworks offer a lot more features and configurability than what this framework has and likely will ever have. Those frameworks are excellent choices for applications that need more than what is offered by this microframework. If all you need is to create a simple rest API in Python that runs on Amazon API Gateway and AWS Lambda, consider giving the Python Serverless Microframework for AWS a try.
- serverless - Build applications comprised of microservices that run in response to events, auto-scale for you, and only charge you when they run.
- Zappa - Deploy python WSGI applications on AWS Lambda and API Gateway.
- claudia - Deploy node.js projects to AWS Lambda and API Gateway.