Automatically read chat text from Minecraft

There’s in fact an even better way to read the chat from Minecraft, and it doesn’t require either screen scraping or packet decoding.

Minecraft automatically writes chat messages (and numerous other things) to log files, both in singleplayer and in multiplayer. On Windows, they are located in %appdata%/.minecraft/logs. Previous logs are compressed using gzip, but the latest session’s log is written to the text file latest.log in realtime. Chat messages contain the text [Client thread/INFO]: [CHAT]. You can either open it as you would with a normal file using:

import os
with open(os.getenv("APPDATA")+"/.minecraft/logs/latest.log", "r") as logfile:
    for line in logfile:
        if "[Client thread/INFO]: [CHAT]" in line:
            print line,

Or if you want to read chat in realtime, you can use the code below, slightly modified from the code from this answer:

import time, os

def follow(thefile):
    thefile.seek(0,2)
    while True:
        line = thefile.readline()
        if not line:
            time.sleep(0.1)
            continue
        yield line

if __name__ == "__main__":
    logfile = open(os.getenv("APPDATA")+"/.minecraft/logs/latest.log", "r")
    loglines = follow(logfile)
    for line in loglines:
        if "[Client thread/INFO]: [CHAT]" in line:
            print line,

First, as kuyan suggested, see http://wiki.vg/Main_Page, which has links to various programs that may be useful, either directly or for source to look at.

For example, under Utilities, the first thing that comes up is a logging proxy.

And a bit down, there's mc3p, the program suggested by Joran Beasley—a Python proxy, with plugin support. It only works up to 1.2.5, but sadimusi/mc3p claims to be a 1.4.2-compatible fork. As J.F. Sebastian says, mc3p has an interface for log plugins, so you can just write one that logs to postgres.

If you want to read the packets yourself, that's not hard. You can write a generic TCP proxy in a few dozen lines of Python—or write one in 2 lines of shellscript around netcat that tees the data to your Python script.

The hard part isn't intercepting the data; it's parsing the protocol. Minecraft probably isn't sending "Nightbane: 1 tnt for 100.000 Dollars each", but something like "offer:Nightbane:1:tnt:100" or "\x13\x09Nightbane\x00\x01\x72\x00\x64". From what the wiki says, the protocol is documented, but poorly, and sometimes inaccurately, and the wiki is sometimes incorrect too, and the official code is very ugly and hard to read. Which means that the best way to figure out the protocol is probably by reading sadimusi/mc3p or one of the other projects like McPacketSniffer or ProtoProxy, at which point you have to ask whether it would be easier to just use that project instead of reimplementing it.

At any rate, scraping the screen should be your last resort.