reel 0.0.5

An async subprocess manager

A package for asynchronous subprocess pipelines:

import reel

tracks = ['track01', 'track02', 'http://w.w.w/track03']
plst = reel.Reel(
    [reel.cmd.ffmpeg.read(_) for _ in tracks],
    announce_to=print
)
dest = reel.cmd.sox.speakers()
norm = reel.cmd.ffmpeg.norm()
vol = reel.cmd.ffmpeg.volume(1.0)

async with plst | norm | vol | dest as transport:
    await transport.play()

Motivation

This project is a simplified version of Python subprocess control with pipes and asynchronous support. It is being developed to support a music streaming package which uses ffmpeg and other shell commands to get music from various sources to various destinations.

Mythology

The end goal is something like this:

import reel

reel.Spool

Logging

reel will log useful messages to a file called reel.log if you configure a log level (e.g. bash):

$ export REEL_LOGGING_LEVEL='INFO'

That’s all you need to set.

Available log levels, ranked by verbosity, with DEBUG the most verbose, are:

• DEBUG - mostly useless information

• INFO - mostly useful information

• WARNING - might be a problem: suitable default for production

• ERROR - something bad happened

• CRITICAL - rare, show-stopping malfunction

• NOTSET - the default: no logging

By default, reel places the log file in $XDG_DATA_HOME. If $XDG_DATA_HOME is not set, reel chooses a suitable default directory. To view the choice, ask reel to print the current configuration (e.g. bash):

$ reel --config | grep LOGGING

For direct control, explicitly set the logging directory with:

$ export REEL_LOGGING_DIR='~/.local/share/reel'

In addition to sending useful information to reel.log, you can reel also logs output produced by subprocesses. A subprocess can generate log files in two ways:

1. The process might write it’s own log file (e.g. a web server).

   In this case, reel might be able to control where the log file is written if the command is configured in reel.cmd. For example, reel.cmd.icecast will automatically write it’s server log file to $REEL_LOGGING_DIR / icecast.log.

2. You might decide to log stderr and/or stdout from a subprocess.

   You can decide what to do with any subprocess output, including logging it all to a file…

In general, reel attempts to keep all log files in one directory and will sparingly create subdirectories if needed.