checking irssi log parser rewrite progress.... got a TON done on it

This commit is contained in:
brent s 2018-08-02 01:48:26 -04:00
parent aac603b4ee
commit 9f1515b96c
1 changed files with 279 additions and 161 deletions

View File

@ -1,179 +1,275 @@
#!/usr/bin/env python3

## REFERENCE ##
# https://github.com/myano/jenni/wiki/IRC-String-Formatting
# https://www.mirc.com/colors.html
# https://en.wikipedia.org/wiki/ANSI_escape_code
# https://github.com/shabble/irssi-docs/wiki/Formats#Colourising-Text-in-IRC-Messages


import argparse
import curses
import os
import pprint
import re
import sys
try:
import magic
has_magic = True
except ImportError:
print(('Warning: you do not have the magic module installed ' +
'(you can install it via "pip3 install --user python-magic"). ' +
'Automatic log decompression will not work.'))
print('Warning: you do not have the magic module installed (you can '
'install it via "pip3 install --user file-magic"). Automatic log '
'decompression will not work.')
has_magic = False

class logParser(object):
# This is a map to determine which module to use to decompress,
# if we should.
cmprsn_map = {'text/plain': None, # Plain ol' text
# Sometimes the formatting with color gives this
'application/octet-stream': None,
'application/x-bzip2': 'bz2', # Bzip2
'application/x-gzip': 'gzip', # Gzip
'application/x-xz': 'lzma'} # XZ

# irssi/mIRC to ANSI
# Split into 3 maps (truecolor will be populated later, currently uses 8-bit):
# - 8 (3/4 bit color values, 8 colors)
# - 256 (8-bit, 256 colors)
# - 'truecolor' (24-bit, ISO-8613-3, 16777216 colors)
# Keys are the mIRC color value
# Values are:
# - 8: tuple for ANSI fg and bg values
# - 256: single value (same number is used for fg and bg)
# - 'truecolor': tuple of (R#, G#, B#) (same number is used for fg and bg)
# In addition, all three have the following:
# - ansi_wrap: the string formatter.
# fg: foreground color
# bg: background color (if present)
# They are concatted together in that order.
## https://en.wikipedia.org/wiki/ANSI_escape_code#3/4_bit
colormap = {8: {'0': ('97', '107'),
'1': ('30', '40'),
'2': ('34', '44'),
'3': ('32', '42'),
'4': ('91', '101'),
'5': ('31', '41'),
'6': ('35', '45'),
'7': ('33', '43'),
'8': ('93', '103'),
'9': ('92', '102'),
'10': ('36', '46'),
'11': ('96', '106'),
'12': ('94', '104'),
'13': ('95', '105'),
'14': ('90', '100'),
'15': ('37', '47'),
'ansi_wrap': {'fg': '\x1b[{0}',
'bg': ';{0}m'}},
## https://en.wikipedia.org/wiki/ANSI_escape_code#8-bit
256: {'0': '15',
'1': '0',
'2': '19',
'3': '34',
'4': '196',
'5': '52',
'6': '90',
'7': '208',
'8': '226',
'9': '82',
'10': '37',
'11': '51',
'12': '21',
'13': '199',
'14': '241',
'15': '252',
'ansi_wrap': {'fg': '\x1b[38;5;{0}m',
'bg': '\x1b[48;5;{0}m'}},
## https://en.wikipedia.org/wiki/ANSI_escape_code#24-bit
# (can just use mIRC's R,G,B)
'truecolor': {'0': ('255', '255', '255'),
'1': ('0', '0', '0'),
'2': ('0', '0', '127'),
'3': ('0', '147', '0'),
'4': ('255', '0', '0'),
'5': ('127', '0', '0'),
'6': ('156', '0', '156'),
'7': ('252', '127', '0'),
'8': ('255', '255', '0'),
'9': ('0', '252', '0'),
'10': ('0', '147', '147'),
'11': ('0', '255', '255'),
'12': ('0', '0', '252'),
'13': ('255', '0', '255'),
'14': ('127', '127', '127'),
'15': ('210', '210', '210'),
'ansi_wrap': {'fg': '\x1b[38;2;{0[0]},{0[1]},{0[2]}',
'bg': '\x1b[48;2;{0[0]},'
'{0[1]},{0[2]}'}}}

def get_palette():
# Return 8, 256, or 'truecolor'
colorterm = os.getenv('COLORTERM', None)
if colorterm == 'truecolor':
# TODO: 24-bit support (16777216 colors) instead of 8-bit.
# See note above.
#return('truecolor')
return(256)
else:
curses.initscr()
curses.start_color()
c = curses.COLORS
curses.endwin()
return(c)

def color_converter(data_in, palette_map):
# Only used if logParser().args['color'] = True
# Convert mIRC/Irssi color coding to ANSI color codes.
color_ptrn = re.compile('\x03[0-9]{1,2}(,[0-9]{1,2})?')
_colors = colormap[palette_map]
# the value to reset formatting to the terminal default
reset_char = '\x1b[0m'
# the value to reset the foreground text
def_fg = '\x1b[39m'
# the value to reset the background text
def_bg = '\x1b[49m'
regex = {'nick': re.compile('\x04[89]/'),
'bold': re.compile('\x02'),
## Doublecheck this. what is the significance of \x04(.*) chars?
'reset': re.compile('\x04(g|c|[389;]/?|e|>)?/?'),
'color_clear': re.compile('\x0f(g|c|[389;]/?|e)?/?')}
# A sub-function that generates the replacement characters.
def _repl(ch_in):
_ch = ch_in.group().lstrip('\x03')
_chars = [i.strip() for i in _ch.split(',', 1)]
if len(_chars) == 1:
ch_out = _colors['ansi_wrap']['fg'].format(_chars[0])
elif len(_chars) == 2:
ch_out = (_colors['ansi_wrap']['fg'].format(_chars[0]) +
_colors['ansi_wrap']['bg'].format(_chars[1]))
else:
raise RuntimeError('Parsing error! "{0}"'.format(ch_in))
return(ch_out)
data = data_in.splitlines()
for idx, line in enumerate(data[:]):
# Get some preliminary replacements out of the way.
line = regex['nick'].sub(' ', line, 1)
line = regex['reset'].sub(reset_char, line)
line = regex['color_clear'].sub(def_fg + def_bg, line)
# TODO: use ptrn.sub(_repl, line) instead since that works and
# this does not
# First set is text color
# Second set, if present, is bg color
line = color_ptrn.sub(_repl, line)
data[idx] = line
return('\n'.join(data))

def plain_stripper(data_in):
# Strip to plaintext only.
data = data_in.splitlines()
ptrns = [re.compile('\x04(g|c|[389;]/?|e|>)/?'),
re.compile('((\x03)\d\d?,\d\d?|(\x03)\d\d?|[\x01-\x1F])')]
for idx, line in enumerate(data[:]):
# This cleans the nick field
l = re.sub('\x04[89]/', ' ', line, 1)
# And these clean the actual chat messages
for p in ptrns:
l = p.sub('', l)
data[idx] = l
return('\n'.join(data))

class irssiLogParser(object):
def __init__(self, args, data = None):
# We'll need these accessible across the entire class.
self.args = args
# If specified, self.data takes precedence over self.args['logfile']
# (if it was specified).
self.data = data
self.raw = data
self.has_html = False
# This is a map to determine which module to use to decompress,
# if we should.
self.cmprsn_map = {'text/plain': None, # Plain ol' text
'application/octet-stream': None, # Sometimes the formatting with color gives this
'application/x-bzip2': 'bz2', # Bzip2
'application/x-gzip': 'gzip', # Gzip
'application/x-xz': 'lzma'} # XZ
# I though y'all liked GUIs.
# ANSI, which is interpreted by the shell.
# Only used if args['color'] = True
self.ansi_prefix = '\033['
# irssi to ANSI
self.colormap = {'00': '1;37m', # White
'01': '0;30m', # Black
'02': '0;34m', # Blue
'03': '0;32m', # Green
'04': '1;31m', # Light Red
'05': '0;31m', # Red
'06': '0;35m', # Magenta (translated as Purple)
'07': '0;33m', # Orange (translated as Brown)
'08': '1;33m', # Yellow
'09': '1:32m', # Light Green
'10': '0;36m', # Cyan
'11': '1;36m', # Light Cyan
'12': '1;34m', # Light Blue
'13': '1;35m', # Light Magenta (translated as Light Purple)
'14': '0;37m', # Gray
'15': '1;37'} # Light Gray (translated as White)
# The full, interpreted path.
if 'logfile' in self.args.keys():
self.args['logfile'] = os.path.abspath(os.path.expanduser(self.args['logfile']))
if not self.data:
self.getLog()
else:
self.data = self.data.decode('utf-8').splitlines()
self.decompress = None
if has_magic:
# Determine what decompressor to use, if we need to.
_mime = magic.detect_from_content(self.data).mime_type
self.decompress = self.cmprsn_map[_mime]
if self.args['html']:
try:
import ansi2html
self.has_html = True
except ImportError:
print(('Warning: you have selected HTML output but do not ' +
'have the ansi2html module installed. Rendering HTML ' +
'output is not possible.'))
self.has_html = False
if 'color' in self.args and self.args['color']:
if not self.args['html']:
# Ensure that we support color output.
curses.initscr()
self.args['color'] = curses.can_change_color()
curses.endwin()
if not self.args['color'] and not self.args['raw']:
raise RuntimeError('You have specified ANSI colorized '
'output but your terminal does not '
'support it. Use -fc/--force-color '
'to force.')
elif not self.args['color'] and self.args['raw']:
self.args['color'] = True # Force the output anyways.
if self.args['color']:
if not self.args['raw']:
self.colors = get_palette()
else:
self.colors = 8 # Best play it safe for maximum compatibility.
# The full, interpreted path.
if ('logfile' in self.args.keys() and
self.args['logfile'] is not None):
self.args['logfile'] = os.path.abspath(
os.path.expanduser(
self.args['logfile']))
if not self.data:
self.getlog()
else:
self.has_html = False
# Conform everything to bytes.
if not isinstance(self.data, bytes):
self.data = self.data.encode('utf-8')
self.decompressor()
self.parser()

def getLog(self):
if not os.path.isfile(self.args['logfile']):
raise FileNotFoundError('{0} does not exist.'.formatself.args['logfile'])
with open(self.args['logfile'], 'rb') as f:
self.data = f.read()
def getlog(self):
# A filepath was specified
if self.args['logfile']:
if not os.path.isfile(self.args['logfile']):
raise FileNotFoundError('{0} does not exist'.format(
self.args['logfile']))
with open(self.args['logfile'], 'rb') as f:
self.data = f.read()
# Try to get it from stdin
else:
if not sys.stdin.isatty():
self.data = sys.stdin.buffer.read()
else:
raise ValueError('Either a path to a logfile must be '
'specified or you must pipe a log in from '
'stdin.')
self.raw = self.data
return()

def parseLog(self):
if self.decompress:
import importlib
self.decmp = importlib.import_module(self.decompress)
self.data = self.decmp.decompress(self.data)
if self.args['color']:
_idx = 0
_datalst = self.data.split(b'\n')
for line in _datalst[:]: # not really "lines", per se, but...
# First we strip out some basic formatting at the beginning
# of lines. Status lines are \x049/, chat lines are \x048/.
# \x04g seem to be formatting resets of sort.
line = re.sub('\x04[89]/'.encode('utf-8'),
''.encode('utf-8'),
line)
line = re.sub('\x04g'.encode('utf-8'),
''.encode('utf-8'),
line)
# Formatting resets
line = re.sub('\x04e'.encode('utf-8'),
'\033[0m'.encode('utf-8'),
line)
# Then we substitute bolds in. This is trickier, because
# bolds (\x04c) *alternate*. So does the other? bold, \x02.
for b in ('\x04c'.encode('utf-8'), '\x02'.encode('utf-8')):
_linelst = line.split(b)
_bold = False
_cnt = 0
for i in _linelst[:]:
if _bold:
_linelst[_cnt] = re.sub('^'.encode('utf-8'),
(self.ansi_prefix + '1m').encode('utf-8'),
i)
else:
_linelst[_cnt] = re.sub('^'.encode('utf-8'),
(self.ansi_prefix + '0m').encode('utf-8'),
i)
_cnt += 1
_bold = not _bold
line = b''.join(_linelst)
# Then we handle colors.
_cnt = 0
_linelst = line.split(b'\x03')
for i in _linelst[:]:
_color_idx = re.sub('^([0-9]{2}).*$'.encode('utf-8'),
'\g<1>',
i,
re.MULTILINE).decode('utf-8')
if _color_idx in self.colormap.keys():
_linelst[_cnt] = re.sub('^[0-9]{2}'.encode('utf-8'),
(self.ansi_prefix + self.colormap[_color_idx]).encode('utf-8'),
i)
_cnt += 1
line = b''.join(_linelst)
# Lastly, we fix join/part and other messages.
_cnt = 0
_linelst = line.split(b'\x04;/')
for i in _linelst[:]:
_templine = re.sub('^'.encode('utf-8'),
''.encode('utf-8'),
i,
re.MULTILINE)
_templine = re.sub('-!-'.encode('utf-8'),
'\033[2m-!-'.encode('utf-8'),
_templine)
_linelst[_cnt] = re.sub('\x043/'.encode('utf-8'),
''.encode('utf-8'),
_templine)
_cnt += 1
line = re.sub(b'^\x1b\[0;32m\x1b\[0m\x1b\[0m', b'\033[0m', b''.join(_linelst))
# Lastly we strip out \x04>/
line = re.sub(b'\x04>/', b'', line)
###
_datalst[_idx] = line
_idx += 1
###
self.data = b'\n'.join(_datalst)
if self.args['html']:
try:
import ansi2html
_has_html = True
except ImportError:
print(('Warning: you have selected HTML output but do not ' +
'have the ansi2html module installed. Rendering HTML ' +
'output is not possible.'))
_has_html = False
else:
_has_html = False
if _has_html:
# This... basically sucks. It currently doesn't properly interpret the ANSI.
_html = ansi2html.Ansi2HTMLConverter()
self.data = _html.convert(self.data.decode('utf-8'))
else: # We want plaintext, so strip ALL formatting.
_stripbytes = ['\x04>/', '\x02', '\x043/', '\x048/', '\x049/', '\x04g', '\x04e', '\x04c', '\x04;/']
for b in _stripbytes:
self.data = re.sub(b.encode('utf-8'), ''.encode('utf-8'), self.data)
self.data = re.sub('\\x03[0-9]{2}'.encode('utf-8'), ''.encode('utf-8'), self.data)
def decompressor(self):
# TODO: use mime module as fallback?
# https://docs.python.org/3/library/mimetypes.html
# VERY less-than-ideal since it won't work without self.args['logfile']
# (and has iffy detection at best, since it relies on file extensions).
# Determine what decompressor to use, if we need to.
if has_magic:
_mime = magic.detect_from_content(self.data).mime_type
self.decompress = cmprsn_map[_mime]
if self.decompress:
import importlib
decmp = importlib.import_module(self.decompress)
self.raw = decmp.decompress(self.data)
else:
# Assume that it's text and that it isn't compressed.
# We'll get a UnicodeDecodeError exception if it isn't.
pass
try:
self.raw = self.data.decode('utf-8')
except UnicodeDecodeError:
pass
self.data = self.raw
return()

def parser(self):
if 'color' not in self.args or not self.args['color']:
self.data = plain_stripper(self.data)
else:
self.data = color_converter(self.data, self.colors)
# Just in case...
self.data += '\x1b[0m'
return()

def parseArgs():
@ -181,19 +277,41 @@ def parseArgs():
args.add_argument('-c', '--color',
dest = 'color',
action = 'store_true',
help = ('Print the log with converted colors.'))
help = ('Print the log with converted colors (ANSI)'))
args.add_argument('-r', '--raw',
dest = 'raw',
action = 'store_true',
help = ('Use this switch if your terminal is detected '
'as not supporting color output but wish to '
'force it anyways. A string representation of '
'the ANSI output will be produced instead ('
'suitable for pasting elsewhere). Only used if '
'-c/--color is enabled (ignored with '
'-H/--html)'))
args.add_argument('-H', '--html',
dest = 'html',
action = 'store_true',
help = ('Render HTML output.'))
help = ('Render HTML output'))
args.add_argument(dest = 'logfile',
default = None,
nargs = '?',
metavar = 'path/to/logfile',
help = ('The path to the log file. It can be uncompressed ' +
'or compressed with XZ/LZMA, Gzip, or Bzip2.'))
'or compressed with XZ/LZMA, Gzip, or Bzip2. '
'If not specified, read from stdin'))
return(args)

if __name__ == '__main__':
args = vars(parseArgs().parse_args())
l = logParser(args)
l.parseLog()
print(l.data.decode('utf-8'))
l = irssiLogParser(args)
import shutil
cols = shutil.get_terminal_size().columns
#pprint.pprint(l.args, width = cols)
pprint.pprint(l.raw, width = cols)
with open('/tmp/freenode.formatted', 'r') as f:
print(f.read())
#pprint.pprint(l.data, width = cols)
#pprint.pprint(repr(l.data).split('\\n'))
print(l.data)
# l.parseLog()
# print(l.data.decode('utf-8'))