When creating a Python function using the func
CLI, the project directory
looks like a typical Python project. Both HTTP and Event functions have the same
template structure.
❯ func create -l python fn
Project path: /home/developer/src/fn
Function name: fn
Runtime: python
❯ tree
fn
├── func.py
├── func.yaml
├── requirements.txt
└── test_func.py
Aside from the func.yaml
file, this looks like the beginning of just about
any Python project. For now, we will ignore the func.yaml
file, and just
say that it is a configuration file that is used when building your project.
If you're really interested, check out the reference doc.
To learn more about the CLI and the details for each supported command, see
the CLI Commands document.
To run a function, you'll first need to build it. This step creates an OCI container image that can be run locally on your computer, or on a Kubernetes cluster.
❯ func build
After the function has been built, it can be run locally.
❯ func run
Functions can be invoked with a simple HTTP request.
You can test to see if the function is working by using your browser to visit
http://localhost:8080. You can also access liveness and readiness
endpoints at http://localhost:8080/health/liveness and
http://localhost:8080/health/readiness. These two endpoints are used
by Kubernetes to determine the health of your function. If everything
is good, both of these will return OK
.
To deploy your function to a Kubernetes cluster, use the deploy
command.
❯ func deploy
You can get the URL for your deployed function with the info
command.
❯ func info
Python functions can be tested locally on your computer. In the project there is
a test_func.py
file which contains a simple unit test. To run the test locally,
you'll need to install the required dependencies. You do this as you would
with any Python project.
❯ pip install -r requirements.txt
Once you have done this, you can run the provided tests with python3 test_func.py
.
The default test framework for Python functions is unittest
. If you prefer another,
that's no problem. Just install a test framework more to your liking.
Boson Python functions have very few restrictions. You can add any required dependencies
in requirements.txt
, and you may include additional local Python files. The only real
requirements are that your project contain a func.py
file which contains a main()
function.
In this section, we will look in a little more detail at how Boson functions are invoked,
and what APIs are available to you as a developer.
When using the func
CLI to create a function project, you may choose to generate a project
that responds to a CloudEvent
or simple HTTP. CloudEvents
in Knative are transported over
HTTP as a POST
request, so in many ways, the two types of functions are very much the same.
They each will listen and respond to incoming HTTP events.
When an incoming request is received, your function will be invoked with a Context
object as the first parameter. This object is a Python class with two attributes. The
request
attribute will always be present, and contains the Flask request
object.
The second attribute, cloud_event
, will be populated if the incoming request is a
CloudEvent
. Developers may access any CloudEvent
data from the context object.
For example:
def main(context: Context):
"""
The context parameter contains the Flask request object and any
CloudEvent received with the request.
"""
print(f"Method: {context.request.method}")
print(f"Event data {context.cloud_event.data})
# ... business logic here
Functions may return any value supported by Flask, as the invocation framework proxies these values directly to the Flask server. See the Flask documentation for more information.
def main(context: Context):
body = { "message": "Howdy!" }
headers = { "content-type": "application/json" }
return body, 200, headers
Note that functions may set both headers and response codes as secondary and tertiary response values from function invocation.
All event messages in Knative are sent as CloudEvents
over HTTP. As noted
above, function developers may access an event through the context
parameter
when the function is invoked. Additionally, developers may use an @event
decorator to inform the invoker that this function's return value should be
converted to a CloudEvent
before sending the response. For example:
@event("event_source"="/my/function", "event_type"="my.type")
def main(context):
# business logic here
data = do_something()
# more data processing
return data
This will result in a CloudEvent
as the response value, with a type of
"my.type"
, a source of "/my/function"
, and the data property set to data
.
Both the event_source
and event_type
decorator attributes are optional.
If not supplied, the CloudEvent's source attribute will be set to
"/parliament/function"
and the type will be set to "parliament.response"
.
Developers are not restricted to the dependencies provided in the template
requirements.txt
file. Additional dependencies can be added as they would be
in any other project by simply adding them to the requirements.txt
file.
When the project is built for deployment, these dependencies will be included
in the container image.