Skip to content

plugins.md

Latest commit

 

History

History
104 lines (70 loc) · 3.64 KB

plugins.md

File metadata and controls

104 lines (70 loc) · 3.64 KB

Plugin module

Plugins structure

A plugin is an uploaded Python script triggered by an event. It can be defined as either :

  • a single python file with everything inside
  • a plugin module (a folder) containing multiple python files and a non mandatory requirements.txt file listing all the dependencies you need for you plugin (providing requirements.txt is not available for On-Premise deployments - see details below).

In the case of the module type plugin, at the root of the folder a file named main.py is strictly necessary, as it serves as the entrypoint of the plugin. In this main.py file, you can import what you need from other python files in the folder. The structure of the folder can be the following (the only constraint being the presence of the main.py file):

plugin_folder
|__ main.py
|__ other_file.py
|__ requirements.txt
|
|___helpers
    |__ helper.py

The plugin you are going to upload has to contain a class PluginHandler(PluginCore) (in the case of the module type plugin it has to be inside main.py) that implements two methods for the different types of events:

  • on_submit
  • on_review

These methods have a predefined set of parameters:

  • the label submitted (a dictionary containing the fields of the GraphQL type Label)
  • the asset_id of the asset labeled

You can add custom methods in your class as well.

Moreover, some attributes are directly available in the class:

  • self.kili
  • self.project_id

Therefore, the skeleton of the plugin (of main.py in the case of the module type plugin) should look like this:

from typing import Dict
import numpy as np

from kili.plugins import PluginCore

def custom_function():
    # Do something...

class PluginHandler(PluginCore):
    """Custom plugin"""

    def custom_method(self):
        # Do something...

    def on_review(self, label: Dict, asset_id: str) -> None:
        """Dedicated handler for Review action"""
        # Do something...

    def on_submit(self, label: Dict, asset_id: str) -> None:
        """Dedicated handler for Submit action"""
        # Do something...

!!! note The plugins run has some limitations, it can use a maximum of 512 MB of ram and will timeout after 60 sec of run.

On-Premise deployment details

The plugins for the on-premise deployments work exactly the same as the plugins for the SaaS version of Kili, with only a few small exceptions :

  1. It's not possible to add custom python packages to your plugin with the help of the requirements.txt file, but we selected a list of the most useful packages that you can directly use, including :
    • numpy, pandas, scikit-learn, opencv-python-headless, Pillow, requests, uuid and of course kili
  2. In order to save the logs during the execution of your plugin, you should only use the provided logger in the plugin class (the simple print function will not save the log). For an example, see the code below:
from logging import Logger
from typing import Dict
from kili.plugins import PluginCore

def custom_function(label: Dict, logger: Logger):
    logger.info("Custom function called")
    # Do something...

class PluginHandler(PluginCore):
    """Custom plugin"""

    def on_submit(self, label: Dict, asset_id: str) -> None:
        """Dedicated handler for Submit action"""
        self.logger.info("On Submit called")
        custom_function(label, self.logger)

Model for Plugins

::: kili.services.plugins.model.PluginCore

Queries

::: kili.entrypoints.queries.plugins.init.QueriesPlugins

Mutations

::: kili.entrypoints.mutations.plugins.init.MutationsPlugins