SmokeDetector

Headless chatbot that detects spam and posts it to chatrooms. Uses ChatExchange, takes questions from the Stack Exchange realtime tab, and accesses answers via the Stack Exchange API.

Example chat post:

Documentation

User documentation is in the wiki.

Detailed documentation for setting up and running SmokeDetector is in the wiki.

Basic setup

To set up SmokeDetector, please use

git clone https://github.com/Charcoal-SE/SmokeDetector.git
cd SmokeDetector
git checkout deploy
sudo pip3 install -r requirements.txt --upgrade
pip3 install --user -r user_requirements.txt --upgrade

Next, copy config.sample to a new file called config, and edit the values required.

To run, use python3 nocrash.py (preferably in a daemon-able mode, like a screen session.) You can also use python3 ws.py, but then SmokeDetector will be shut down after 6 hours; when running from nocrash.py, it will be restarted. (This is to be sure that closed websockets, if any, are reopened.)

Virtual environment setup

Running in a virtual environment is a good way to isolate dependency packages from your local system. To set up SmokeDetector in a virtual environment, you can use

git clone https://github.com/Charcoal-SE/SmokeDetector.git
cd SmokeDetector
git config user.email "smokey@erwaysoftware.com"
git config user.name "SmokeDetector"
git checkout deploy

python3 -m venv env
env/bin/pip3 install -r requirements.txt --upgrade
env/bin/pip3 install --user -r user_requirements.txt --upgrade

Next, copy the config file and edit as said above. To run SmokeDetector in this virtual environment, use env/bin/python3 nocrash.py.

[Note: On some systems (e.g. Mac's and Linux), some circumstances may require the --user option to be removed from the last pip3 command line in the above instructions. However, the --user option is known to be necessary in other circumstances. Further testing is necessary to resolve the discrepancy.]

Docker setup

Running in a Docker container is an even better way to isolate dependency packages from your local system. To set up SmokeDetector in a Docker container, follow the steps below.

1. Grab the Dockerfile and build an image of SmokeDetector:

   DATE=$(date +%F)
   mkdir temp
   cd temp
   wget https://raw.githubusercontent.com/Charcoal-SE/SmokeDetector/master/Dockerfile
   docker build -t smokey:$DATE .

2. Create a container from the image you just built

   docker create --name=mysmokedetector smokey:$DATE

3. Start the container. Don't worry, SmokeDetector won't run until it's ready, so you have the chance to edit the configuration file before SmokeDetector runs.

   Copy config.sample to a new file named config and edit the values required, then copy the file into the container with this command:

   docker cp config mysmokedetector:/home/smokey/SmokeDetector/config

4. If you would like to set up additional stuff (SSH, Git etc.), you can do so with a Bash shell in the container:

   docker exec -it mysmokedetector bash

   After you're ready, put a file named ready under /home/smokey:

   touch ~smokey/ready

Automate Docker deployment with Docker Compose

I'll assume you have the basic ideas of Docker and Docker Compose.

The first thing you need is a properly filled config file. You can start with the sample.

Create a directory (name it whatever you like), place the config file and docker-compose.yml file. Run docker-compose up -d and your SmokeDetector instance is up.

If you want additional control like memory and CPU constraint, you can edit docker-compose.yml and add the following keys to smokey. The example values are recommended values.

restart: always  # when your host reboots Smokey can autostart
mem_limit: 512M
cpus: 0.5  # Recommend 2.0 or more for spam waves

Requirements

SmokeDetector only supports Stack Exchange logins.

SmokeDetector supports the Python versions which are in the supported phase of the Python life cycle (as defined as between "First release" and "End of life"). We run CI testing on that span of versions. SmokeDetector may work on older versions of Python, but we don't support them and may, at any time, write code that prevents use in older, unsupported versions. We know SmokeDetector is broken on Python 3.6 and lower. While we don't support versions that haven't reached "First release", we're not adverse to hearing about changes in new Python versions which will require us to make changes to SmokeDetector's code, so we can make the transition to supporting new versions of Python smoother.

To allow committing blacklist and watchlist modifications back to GitHub, your system also needs Git 1.8 or higher, although we recommend Git 2.11+.

License

Licensed under either of

• Apache License, Version 2.0, (LICENSE-APACHE or https://www.apache.org/licenses/LICENSE-2.0)
• MIT license (LICENSE-MIT or https://opensource.org/licenses/MIT)

at your option.

Contribution Licensing

By submitting your contribution for inclusion in the work as defined in the Apache-2.0 license, you agree that it be dual licensed as above, without any additional terms or conditions.