What is Weechat-Matrix?

Weechat is an extensible chat client.

Matrix is an open network for secure, decentralized communication.

weechat-matrix is a Python script for Weechat that lets Weechat communicate over the Matrix protocol.

Project Status

weechat-matrix is stable and quite usable as a daily driver. It already supports large parts of the Matrix protocol, including end-to-end encryption (though some features like cross-signing and session unwedging are unimplemented).

However, due to some inherent limitations of Weechat scripts, development has moved to weechat-matrix-rs, a Weechat plugin written in Rust. As such, weechat-matrix is in maintenance mode and will likely not be receiving substantial new features. PRs are still accepted and welcome.

Installation

1. Install libolm 3.1+

   • Debian 11+ (testing/sid) or Ubuntu 19.10+ install libolm-dev

   • Archlinux based distribution can install the libolm package from the Community repository

   • FreeBSD pkg install olm

   • macOS brew install libolm

   • Failing any of the above see https://gitlab.matrix.org/matrix-org/olm for instructions about building it from sources

2. Clone the repo and install dependencies

   git clone https://github.com/poljar/weechat-matrix.git
   cd weechat-matrix
   pip install -r requirements.txt
   

3. As your regular user, just run: make install in this repository directory.

   This installs the main python file (main.py) into ~/.weechat/python/ (renamed to matrix.py) along with the other python files it needs (from the matrix subdir).

   Note that weechat only supports Python2 OR Python3, and that setting is determined at the time that Weechat is compiled. Weechat-Matrix can work with either Python2 or Python3, but when you install dependencies you will have to take into account which version of Python your Weechat was built to use.

   The minimal supported python2 version is 2.7.10.

   The minimal supported python3 version is 3.5.4 or 3.6.1.

   To check the python version that weechat is using, run:

   /python version
   

Using virtualenv

If you want to install dependencies inside a virtualenv, rather than globally for your system or user, you can use a virtualenv. Weechat-Matrix will automatically use any virtualenv it finds in a directory called venv next to its main Python file (after resolving symlinks). Typically, this means ~/.weechat/python/venv.

To create such a virtualenv, you can use something like below. This only needs to happen once:

virtualenv ~/.weechat/python/venv

Then, activate the virtualenv:

. ~/.weechat/python/venv/bin/activate

This needs to be done whenever you want to install packages inside the virtualenv (so before running the pip install command documented above.

Once the virtualenv is prepared in the right location, Weechat-Matrix will automatically activate it when the script is loaded. This should not affect other script, which seem to have a separate Python environment.

Note that this only supports virtualenv tools that support the activate_this.py way of activation. This includes the virtualenv command, but excludes pyvenv and the Python3 venv module. In particular, this works if (for a typical installation of matrix.py) the file ~/.weechat/python/venv/bin/activate_this.py exists.

Run from git directly

Rather than copying files into ~/.weechat (step 3 above), it is also possible to run from a git checkout directly using symlinks.

For this, you need two symlinks:

ln -s /path/to/weechat-matrix/main.py ~/.weechat/python/matrix.py
ln -s /path/to/weechat-matrix/matrix ~/.weechat/python/matrix

This first link is the main python file, that can be loaded using /script load matrix.py. The second link is to the directory with extra python files used by the main script. This directory must be linked as ~/.weechat/python/matrix so it ends up in the python library path and its files can be imported using e.g. import matrix from the main python file.

Note that these symlinks are essentially the same as the files that would have been copied using make install.

Uploading files

Uploads are done using a helper script, which is found under contrib/matrix_upload. We recommend you install this under your PATH as matrix_upload (without the .py suffix). Uploads can be done from Weechat with: /upload <file>.

Downloading encrypted files

Encrypted files are displayed as an emxc:// URI which cannot be directly opened. They can be opened in two different ways:

• In the CLI by running the contrib/matrix_decrypt helper script.

• In the browser by using matrix-decryptapp. This is a static website which cannot see your data, all the decryption happens on the client side. You can either host it yourself or directly use the instance hosted on seirl.github.io. This weechat trigger will convert all your emxc:// URLs into clickable https links:

  /trigger addreplace emxc_decrypt modifier weechat_print "" ";($|[^\w/#:\[])(emxc://([^ ]+));${re:1}https://seirl.github.io/matrix-decryptapp/#${re:2};"
  

Configuration

Configuration is completed primarily through the Weechat interface. First start Weechat, and then issue the following commands:

1. Start by loading the Weechat-Matrix script:

   /script load matrix.py
   

2. Now set your username and password:

   /set matrix.server.matrix_org.username johndoe
   /set matrix.server.matrix_org.password jd_is_awesome
   

3. Now try to connect:

   /matrix connect matrix_org
   

4. Automatically load the script

   $ ln -s ../matrix.py ~/.weechat/python/autoload
   

5. Automatically connect to the server

   /set matrix.server.matrix_org.autoconnect on
   

6. If everything works, save the configuration

   /save
   

For using a custom (not matrix.org) matrix server:

1. Add your custom server to the script:

   /matrix server add myserver myserver.org
   

2. Add the appropriate credentials

   /set matrix.server.myserver.username johndoe
   /set matrix.server.myserver.password jd_is_awesome
   

3. If everything works, save the configuration

   /save
   

Single sign-on:

Single sign-on is supported using a helper script, the script found under contrib/matrix_sso_helper should be installed under your PATH as matrix_sso_helper (without the .py suffix).

For single sign-on to be the preferred leave the servers username and password empty.

After connecting a URL will be presented which needs to be used to perform the sign on. Please note that the helper script spawns a HTTP server which waits for the sign-on token to be passed back. This makes it necessary to do the sign on on the same host as Weechat.

A hsignal is sent out when the SSO helper spawns as well, the name of the hsignal is matrix_sso_login and it will contain the name of the server in the server variable and the full URL that can be used to log in in the url variable.

To open the login URL automatically in a browser a trigger can be added:

    /trigger add sso_browser hsignal matrix_sso_login "" "" "/exec -bg firefox ${url}"

If signing on on the same host as Weechat is undesirable the listening port of the SSO helper should be set to a static value using the sso_helper_listening_port setting:

   /set matrix.server.myserver.sso_helper_listening_port 8443

After setting the listening port the same port on the local machine can be forwarded using ssh to the remote host:

    ssh -L 8443:localhost:8443 example.org

This forwards the local port 8443 to the localhost:8443 address on example.org. Note that it is necessary to forward the port to the localhost address on the remote host because the helper only listens on localhost.

Bar items

There are two bar items provided by this script:

1. matrix_typing_notice - shows the currently typing users

2. matrix_modes - shows room and server info (encryption status of the room, server connection status)

They can be added to the weechat status bar as usual: /set weechat.bar.status.items

The matrix_modes bar item is replicated in the already used buffer_modes bar item.

Typing notifications and read receipts

The sending of typing notifications and read receipts can be temporarily disabled for a given room via the /room command. They can also be permanently configured using standard weechat conditions settings with the following settings:

1. matrix.network.read_markers_conditions
2. matrix.network.typing_notice_conditions

Cursor bindings

While you can reply on a matrix message using the /reply-matrix command (see its help in weechat), weechat-matrix also adds a binding in /cursor mode to easily reply to a particular message. This mode can be triggered either by running /cursor, or by middle-clicking somewhere on the screen. See weechat's help for /cursor.

The default binding is:

/key bindctxt cursor @chat(python.matrix.*):r hsignal:matrix_cursor_reply

This means that you can reply to a message in a Matrix buffer using the middle mouse button, then r.

This binding is automatically set when the script is loaded and there is no such binding yet. If you want to use a different key than r, you can execute the above command with a different key in place of r. To use modifier keys like control and alt, use alt-k, then your wanted binding key combo, to enter weechat's representation of that key combo in the input bar.

Navigating room buffers using go.py

If you try to use the go.py script to navigate buffers created by weechat-matrix, go.py will by default use the full buffer name which does not contain a human-readable room display name but only the Matrix room ID. This is necessary so that the logger file is able to produce unique, permanent filenames for a room.

However, buffers also have human-readable short names. To make go.py use the short names for navigation, you can run the following command:

/set plugins.var.python.go.short_name "on"

As an alternative, you can also force weechat-matrix to use human-readable names as the full buffer names by running

/set matrix.look.human_buffer_names on

Beware that you will then also need to adjust your logger setup to prevent room name conflicts from causing logger file conflicts.

Helpful Commands

/help matrix will print information about the /matrix command.

/help olm will print information about the /olm command that is used for device verification.

/matrix help [command] will print information for subcommands, such as /matrix help server