Skip to main content

No project description provided

Project description

Pipez - lightweight library for fast deploy stream handling

Install

For installing default version of library use

pip install pipez

If you want install specific version pipez - use

pip install pipez[<your choice>]

Now available cv, fastapi and onnxruntime versions. If you want install pypez with all depencies, you can use

pip install pipez[all]

If you want to install a few version - see nex example:

pip install pipez[cv, onnxruntime]

Quick start

Developing custom node

If you want use your node - you can use Registry.add as class decorator from pipez.registry. You should also import base Node class from pipez.node. For example:

from pipez.core.legacy_node import Node
from pipez.core.legacy_registry import Registry

Registry.add


class MyNode(Node):
    ...

Once required method which you should override: work_func(...) which handle Batch from pipez.batch. However, methods post_init(...) and close(...) also available. See next example:

from typing import Optional

from pipez.core.legacy_batch import Batch, BatchStatus
from pipez.core.legacy_node import Node
from pipez.core.legacy_registry import Registry

Registry.add


class MyNode(Node):
    def __init__(
            self,
            a: int = 1,
            **kwargs
    ):
        super().__init__(**kwargs)
        self._a = a

    def post_init(self):
        self._a *= 10

    def close(self):
        self._a = 0

    def work_func(
            self,
            data: Optional[Batch] = None
    ) -> Batch:
        self._a *= 2
        if self._a > 1000:
            return Batch(status=BatchStatus.END)
        return Batch(data=[dict(a=self._a)])

Build pipelines

When you defined all nodes what you need, we build pipeline from them. You can use json describe or class for node. See next examples:

For using json describing you must add Registry.add as class decorator for you node, else you will get error.

{
    "cls": "MyNode",
    "a": 5,
    "type": "Process",
    "output": "some_trash"
}

For using class you must import your node class.

from pipez.core.legacy_node import NodeType

from ... import MyNode

MyNode(
    a=5,
    type=NodeType.PROCESS,
    output='some_trash'
)

As we can see, we used NodeType, which define type of node.

For building pipeline, we must use build_pipeline from pipez.build. For example:

from pipez.core.legacy_build import build_pipeline
from pipez.nodes import DummyNode
from pipez.core.legacy_node import NodeType
from ... import MyNode

watchdog = build_pipeline(
    pipeline=[
        MyNode(
            a=10,
            type=NodeType.THREAD,
            output='q1'
        ),
        DummyNode(
            type=NodeType.PROCESS,
            input='q1',
            output='q2'
        ),
        DummyNode(
            type=NodeType.THREAD,
            input=['q1, q2'],
            output='q3'
        ),
        {
            "cls": "DummyNode",
            "type": "thread",
            "input": "q3"
        }
    ]
)

As we can see, build_pipeline return watchdog. You can read about it in next section.

WatchDog

TODO

РЎontributors

Project details


Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pipez-0.0.84.tar.gz (1.2 MB view details)

Uploaded Source

Built Distribution

pipez-0.0.84-py3-none-any.whl (1.4 MB view details)

Uploaded Python 3

File details

Details for the file pipez-0.0.84.tar.gz.

File metadata

  • Download URL: pipez-0.0.84.tar.gz
  • Upload date:
  • Size: 1.2 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.3

File hashes

Hashes for pipez-0.0.84.tar.gz
Algorithm Hash digest
SHA256 52e1bc5d37222e782ade6286f865b9489c5fd35b6561444a39eeff324f8c62f4
MD5 d1a045d3988fbb2db1354a6d61a03c14
BLAKE2b-256 407dbb53fbdd5f3f5569d41d190f07663e4e9e180f8457be52031fb11bb5247a

See more details on using hashes here.

File details

Details for the file pipez-0.0.84-py3-none-any.whl.

File metadata

  • Download URL: pipez-0.0.84-py3-none-any.whl
  • Upload date:
  • Size: 1.4 MB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.3

File hashes

Hashes for pipez-0.0.84-py3-none-any.whl
Algorithm Hash digest
SHA256 d2d2303ea39e575d5b9f704f9d69de072bb74d20553c76ce176b89620a17f04f
MD5 a4b90f7c69601203b96da3dfbe143491
BLAKE2b-256 61eed87e0479191b93dc6f38725d98944c7e6b378bd7cabfaff52cf5d6b2ac52

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page