Metadata-Version: 2.1
Name: lemmy_mod_bot
Version: 0.0.5
Summary: lemmy_mod_bot is a Lemmy Mod Bot written in Python
Project-URL: Homepage, https://github.com/noenfugler/LemmyModBot
Project-URL: Bug Tracker, https://github.com/noenfugler/LemmyModBot/issues
Author-email: noenfugler <noenfugler@greenwolf.com.au>
License-File: LICENCE.txt
Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.7
Description-Content-Type: text/markdown

# LemmyModBot
A Lemmy bot written in Python using [pylemmy](https://github.com/dcferreira/pylemmy) to access comment and create reports.  

This bot is a work in progress and is expected to expand and grow.

## Toxic Content Detection
LemmyModBot can detect and report toxic written content on Lemmy communities using a 512-64-2 bag of words deep text classification Neural Network using Pytorch and a local dataset for training.

The neural network with the supplied training dataset is currently achieving accuracies of around 75-80% in detecting toxic content.  Ongoing addition of ambiguous or misclassified data to the training set is hoped to improve this.

## Regexp
LemmyModBot can currently detect posts using regular expressions that do not contain a question mark.  This functionality is planned to be expended to run arbitrary regular expressions.

## User Watchlist
Selected users placed on a watachlist can be reported for moderator review.

## Matrix integration
Messages can be sent to a matrix room.

## Assessed post and comment history
History is stored in an SQLite database

## Example Usage

from lemmybot import LemmyBot

lemmy_world = LemmyBot(rebuild_model=True)
  
lemmy_world.run()
 
