Developing new plugins

This document describes how to develop a new analysis plugin. For an example of conversion plugins, see Conversion.

A more step-by-step tutorial with slides is available here

What is a plugin?

A plugin is a program that, given a text, will add annotations to it. In practice, a plugin consists of at least two files:

  • Definition file: a .senpy file that describes the plugin (e.g. what input parameters it accepts, what emotion model it uses).
  • Python module: the actual code that will add annotations to each input.

This separation allows us to deploy plugins that use the same code but employ different parameters. For instance, one could use the same classifier and processing in several plugins, but train with different datasets. This scenario is particularly useful for evaluation purposes.

The only limitation is that the name of each plugin needs to be unique.

Plugin Definition files

The definition file contains all the attributes of the plugin, and can be written in YAML or JSON. When the server is launched, it will recursively search for definition files in the plugin folder (the current folder, by default). The most important attributes are:

  • name: unique name that senpy will use internally to identify the plugin.

  • module: indicates the module that contains the plugin code, which will be automatically loaded by senpy.

  • version

  • extra_params: to add parameters to the senpy API when this plugin is requested. Those parameters may be required, and have aliased names. For instance:

    extra_params:
    hello_param:
        aliases: # required
            - hello_param
            - hello
        required: true
        default: Hi you
        values:
            - Hi you
            - Hello y'all
            - Howdy

    Parameter validation will fail if a required parameter without a default has not been provided, or if the definition includes a set of values and the provided one does not match one of them.

A complete example:

name: <Name of the plugin>
module: <Python file>
version: 0.1

And the json equivalent:

{
  "name": "<Name of the plugin>",
  "module": "<Python file>",
  "version": "0.1"
}

Plugins Code

The basic methods in a plugin are:

  • __init__
  • activate: used to load memory-hungry resources
  • deactivate: used to free up resources
  • analyse_entry: called in every user requests. It takes two parameters: Entry, the entry object, and params, the parameters supplied by the user. It should yield one or more Entry objects.

Plugins are loaded asynchronously, so don’t worry if the activate method takes too long. The plugin will be marked as activated once it is finished executing the method.

Entries

Entries are objects that can be annotated. By default, entries are NIF contexts represented in JSON-LD format. Annotations are added to the object like this:

entry = Entry()
entry.vocabulary__annotationName = 'myvalue'
entry['vocabulary:annotationName'] = 'myvalue'
entry['annotationNameURI'] = 'myvalue'

Where vocabulary is one of the prefixes defined in the default senpy context, and annotationURI is a full URI. The value may be any valid JSON-LD dictionary. For simplicity, senpy includes a series of models by default in the senpy.models module.

Example plugin

In this section, we will implement a basic sentiment analysis plugin. To determine the polarity of each entry, the plugin will compare the length of the string to a threshold. This threshold will be included in the definition file.

The definition file would look like this:

name: helloworld
module: helloworld
version: 0.0
threshold: 10
description: Hello World

Now, in a file named helloworld.py:

#!/bin/env python
#helloworld.py

from senpy.plugins import AnalysisPlugin
from senpy.models import Sentiment


class HelloWorld(AnalysisPlugin):

    def analyse_entry(entry, params):
        '''Basically do nothing with each entry'''

        sentiment = Sentiment()
        if len(entry.text) < self.threshold:
            sentiment['marl:hasPolarity'] = 'marl:Positive'
        else:
            sentiment['marl:hasPolarity'] = 'marl:Negative'
        entry.sentiments.append(sentiment)
        yield entry

The complete code of the example plugin is available here.

Loading data and files

Most plugins will need access to files (dictionaries, lexicons, etc.). It is good practice to specify the paths of these files in the plugin configuration, so the same code can be reused with different resources.

name: dictworld
module: dictworld
dictionary_path: <PATH OF THE FILE>

The path can be either absolute, or relative.

From absolute paths

Absolute paths (such as /data/dictionary.csv are straightfoward:

with open(os.path.join(self.dictionary_path) as f:
   ...

From relative paths

Since plugins are loading dynamically, relative paths will refer to the current working directory. Instead, what you usually want is to load files relative to the plugin source folder, like so:

.
..
plugin.senpy
plugin.py
dictionary.csv

For this, we need to first get the path of your source folder first, like so:

import os
root = os.path.realpath(__file__)
with open(os.path.join(root, self.dictionary_path) as f:
    ...

Docker image

Add the following dockerfile to your project to generate a docker image with your plugin:

FROM gsiupm/senpy:0.8.8

This will copy your source folder to the image, and install all dependencies. Now, to build an image:

docker build . -t gsiupm/exampleplugin

And you can run it with:

docker run -p 5000:5000 gsiupm/exampleplugin

If the plugin non-source files (Loading data and files), the recommended way is to use absolute paths. Data can then be mounted in the container or added to the image. The former is recommended for open source plugins with licensed resources, whereas the latter is the most convenient and can be used for private images.

Mounting data:

docker run -v $PWD/data:/data gsiupm/exampleplugin

Adding data to the image:

FROM gsiupm/senpy:0.8.8
COPY data /

F.A.Q.

What annotations can I use?

You can add almost any annotation to an entry. The most common use cases are covered in the API and Examples.

Why does the analyse function yield instead of return?

This is so that plugins may add new entries to the response or filter some of them. For instance, a context detection plugin may add a new entry for each context in the original entry. On the other hand, a conversion plugin may leave out those entries that do not contain relevant information.

If I’m using a classifier, where should I train it?

Training a classifier can be time time consuming. To avoid running the training unnecessarily, you can use ShelfMixin to store the classifier. For instance:

from senpy.plugins import ShelfMixin, AnalysisPlugin

class MyPlugin(ShelfMixin, AnalysisPlugin):
    def train(self):
        ''' Code to train the classifier
        '''
        # Here goes the code
        # ...
        return classifier

    def activate(self):
        if 'classifier' not in self.sh:
            classifier = self.train()
            self.sh['classifier'] = classifier
        self.classifier = self.sh['classifier']

    def deactivate(self):
        self.close()

You can specify a ‘shelf_file’ in your .senpy file. By default the ShelfMixin creates a file based on the plugin name and stores it in that plugin’s folder.

Shelves may get corrupted if the plugin exists unexpectedly. A corrupt shelf prevents the plugin from loading. If you do not care about the pickle, you can force your plugin to remove the corrupted file and load anyway, set the ‘force_shelf’ to True in your .senpy file.

How can I turn an external service into a plugin?

This example ilustrate how to implement a plugin that accesses the Sentiment140 service.

class Sentiment140Plugin(SentimentPlugin):
    def analyse_entry(self, entry, params):
        text = entry.text
        lang = params.get("language", "auto")
        res = requests.post("http://www.sentiment140.com/api/bulkClassifyJson",
                            json.dumps({"language": lang,
                                        "data": [{"text": text}]
                                        }
                                       )
                            )

        p = params.get("prefix", None)
        polarity_value = self.maxPolarityValue*int(res.json()["data"][0]
                                                   ["polarity"]) * 0.25
        polarity = "marl:Neutral"
        neutral_value = self.maxPolarityValue / 2.0
        if polarity_value > neutral_value:
            polarity = "marl:Positive"
        elif polarity_value < neutral_value:
            polarity = "marl:Negative"

        sentiment = Sentiment(id="Sentiment0",
                            prefix=p,
                            marl__hasPolarity=polarity,
                            marl__polarityValue=polarity_value)
        sentiment.prov__wasGeneratedBy = self.id
        entry.sentiments.append(sentiment)
        yield entry

Can my plugin require additional parameters from the user?

You can add extra parameters in the definition file under the attribute extra_params. It takes a dictionary, where the keys are the name of the argument/parameter, and the value has the following fields:

  • aliases: the different names which can be used in the request to use the parameter.
  • required: if set to true, users need to provide this parameter unless a default is set.
  • options: the different acceptable values of the parameter (i.e. an enum). If set, the value provided must match one of the options.
  • default: the default value of the parameter, if none is provided in the request.
extra_params
   language:
      aliases:
          - language
          - lang
          - l
      required: true,
      options:
          - es
          - en
      default: es

This example shows how to introduce a parameter associated with language. The extraction of this paremeter is used in the analyse method of the Plugin interface.

lang = params.get("language")

Where can I set up variables for using them in my plugin?

You can add these variables in the definition file with the structure of attribute-value pairs.

Every field added to the definition file is available to the plugin instance.

Can I activate a DEBUG mode for my plugin?

You can activate the DEBUG mode by the command-line tool using the option -d.

senpy -d

Additionally, with the --pdb option you will be dropped into a pdb post mortem shell if an exception is raised.

senpy --pdb