gestalt is a Python application framework for building distributed systems
Project description
Gestalt
Gestalt is an asyncio based Python package to help build distributed systems.
Gestalt provides a variety of building blocks to assist building long-running network applications that leverage asyncio concurrency primitives.
gestalt:
an organized whole that is perceived as more than the sum of its parts.
This project is in the early stages of development. Expects lots of changes.
Features
The main features of Gestalt are:
-
Inter-process Communications
-
Automatic serialization and compression of message payloads.
-
Socket endpoints.
-
TCP
- Stream
- Netstring
- Message Type Identifier
-
UDP
- Datagram
- Message Type Identifier
-
-
Message queuing (i.e. AMQP) components. The Advanced Message Queuing Protocol (AMQP) is an open standard protocol specification for message passing, queuing, routing, reliability and security. One of the most popular implementations of AMQP is RabbitMQ. The Gestalt package provides high level components that support automatic message serialization and compression of message payloads.
- Topic Publisher and Subscriber
- Request and Reply (RPC)
-
-
A Timer component that simplifies creating periodic calls to a function. Timers can be created as single-shot, repeat a specified number of times or repeat forever.
-
An application runner that simplifies running an asyncio application by performing common setup such as creating an event loop, registering signal handlers, registering a global exception handler and performing graceful shutdown.
Installation
Gestalt is available on PyPI and can be installed using pip.
Install core functionality using the command:
$ pip install gestalt
Gestalt uses optional third party Python packages to obtain support for features such as serialization, compression and message queuing. To install optional extras simply specify one or more of them using the extras specifier notation as shown in the example below which installs all of the available extras:
$ pip install gestalt[amq,protobuf,msgpack,avro,brotli,snappy,yaml]
where the available extras are:
-
protobuf
will install support for serializing Google Protocol Buffers structures. -
msgpack
will install support for serializing Msgpack structures. -
yaml
will install support for serializing YAML structures. -
avro
will install support for serializing Apache Avro structures. -
snappy
will install Snappy compression support. The Python snappy package is simply a binding to a system library. Therefore you must install that first before the Python binding will install successfully. For example, on Debian systems you will wantsudo apt-get install libsnappy-dev
-
brotli
will install Brotli compression support. -
amq
will install the AMQP extras (asyncio bindings for RabbitMQ) -
develop
will install the extras needed to development Gestalt.
Once installed you can begin using Gestalt to develop applications.
Usage
There are many examples under the examples
directory.
The API Reference provides API-level documentation.
Performance Tip
Consider using the high performance uvloop event loop implementation instead of the default asyncio event loop.
Install the package using:
(venv) $ pip install uvloop
Then, simply call uvloop.install()
before creating an asyncio event loop
either via the standard asyncio.get_event_loop()
, asyncio.run()
or the
gestalt.runner.run
function.
Change Log
Notable changes to this project will be documented in this file.
Version History
20.1.0
- Update package to support Python3.8
- Use yarl package for URLs.
- Add linting to improve code sustainment.
- Fix bug in stream protocols that affected msg_len in scenarios where messages were fragmented.
19.9.2
- Fix markdown displayed on PyPI.
19.9.1
- Adopted CalVer for package versioning.
- Initial functionality release.
- Clean up type annotation to remove all Mypy errors.
- Add Mypy type check to CI.
- Improve unit test code coverage.
0.2.0
- Added basic functionality.
0.0.1
- Project created.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.