Antispam

Antispam is a bayesian anti-spam classifier written in Python.

Installation

pip install antispam

Usage

Use the built-in model provided & trained by author:

import antispam

antispam.score("Cheap shoes for sale at DSW shoe store!")
# => 0.9657724517163143

antispam.is_spam("Cheap shoes for sale at DSW shoe store!")
# => True

antispam.score("Hi mark could you please send me a copy of your machine learning homework? thanks")
# => 0.0008064840568731558

antispam.is_spam("Hi mark could you please send me a copy of your machine learning homework? thanks")
# => False

Documentation

API

antispam.score(msg)

Score the message based on the built-in model.

Parameters:msg – Message to be scored in string format.

antispam.is_spam(msg)

Decide whether the message is a spam or not based on the built-in model.

Parameters:msg – Message to be classified in string format.

antispam.Model

class antispam.Model(file_path=None, create_new=False)

Save & Load the model in/from the file system using Python’s json module.

Constructs a Model object by the indicated file_path, if the file does not exist, create a new file and contruct a empty model.

Parameters:

• file_path – (optional) Path for the model file indicated, if path is not indicated, use the built-in model file provided by the author, which is located in the antispam package folder.
• create_new – (option) Boolean. If True, create an empty model. file_path will be used when saving the model. If there is an existing model file on the path, the existing model file will be overwritten.

load(file_path=None)

Load the serialized file from the specified file_path, and return spam_count_total, ham_count_total and token_table.

Parameters:file_path – (optional) Path for the model file. If the path does not exist, create a new one.

save()

Serialize the model using Python’s json module, and save the serialized modle as a file which is indicated by self.file_path.

antispam.Detector

class antispam.Detector(path=None, create_new=False)

A baysian spam filter

Parameters:path – (optional) Path for the model file, will be passes to Model and construct a Model object based on path.

train(msg, is_spam)

Train the model.

Parameters:

• msg – Message in string format.
• is_spam – Boolean. If True, train the message as a spam, if False, train the message as a ham.

score(msg)

Calculate and return the spam score of a msg. The higher the score, the stronger the liklihood that the msg is a spam is.

Parameters:msg – Message in string format.

save()

Save self.model based on self.model.file_path.

LICENSE

MIT License