209 lines
6.8 KiB
Python
209 lines
6.8 KiB
Python
import os
|
|
import datetime
|
|
import json
|
|
|
|
from irclogtools import logfile_pattern
|
|
|
|
|
|
class CombinedLogfile(object):
|
|
HEADER = "#$$$COMBINEDLOG"
|
|
PORTIONHEADER = "#$$$BEGINPORTION"
|
|
ENDPORTIONHEADER = "#$$$ENDPORTION"
|
|
|
|
def __init__(self, fpath):
|
|
self.path = fpath
|
|
self.data = []
|
|
if os.path.exists(self.path):
|
|
self._parse()
|
|
# TODO maybe an interface to limit the date range of added stuff?
|
|
|
|
def _parse(self):
|
|
"""
|
|
Open the logfile and load each section into memory
|
|
"""
|
|
with open(self.path, "rb") as f:
|
|
# Read the magic header
|
|
header = f.readline().decode("UTF-8")
|
|
assert self.HEADER and header[0:len(self.HEADER)] == self.HEADER, "Invalid header!"
|
|
|
|
channel = None
|
|
network = None
|
|
|
|
meta = None
|
|
portion = None
|
|
|
|
while True:
|
|
line = f.readline()
|
|
if not line:
|
|
break
|
|
|
|
if line.startswith(self.PORTIONHEADER.encode("UTF-8")):
|
|
assert portion is None and meta is None, "Started portion while already in portion?"
|
|
meta = json.loads(line.decode("UTF-8").split(" ", 1)[1].strip())
|
|
portion = b''
|
|
if not channel:
|
|
channel = meta["channel"]
|
|
if not network:
|
|
network = meta["network"]
|
|
assert channel == meta["channel"], "Portion does not match first portion's channel"
|
|
# assert network == meta["network"], "Portion does not match first portion's network"
|
|
|
|
elif line.startswith(self.ENDPORTIONHEADER.encode("UTF-8")):
|
|
assert portion is not None and meta is not None, "Ended portion while not in portion?"
|
|
self.data.append(VirtualLogFile(meta["name"], portion))
|
|
portion = None
|
|
meta = None
|
|
|
|
else:
|
|
portion += line
|
|
|
|
assert portion is None and meta is None, "Unexpected EOF during open portion"
|
|
|
|
def write(self, target_path=None, raw=False):
|
|
"""
|
|
Write the in-memory contents to disk. A log archive is a UTF-8 text file and is described below.
|
|
|
|
The files start with a header, containing only the channel name: #TODO number of portions + check on parsing
|
|
|
|
#$$$COMBINEDLOG '#chan'
|
|
|
|
Then sorted repeating units of:
|
|
|
|
#$$$BEGINPORTION {"channel": "#chan", "date": "20140119", "name": "#chan_20140119.log", "network": null}
|
|
newline-separated UTF-8 log messages
|
|
#$$$ENDPORTION #hcsmp_20140119.log
|
|
|
|
the metadata is json and must be sorted by key. network may be null but no other fields may be. date must be
|
|
formatted as above and name, the original file name, must match by irclogtools.logfile_pattern.
|
|
"""
|
|
if not target_path:
|
|
target_path = self.path
|
|
|
|
channel = self.data[0].channel
|
|
print("{}: writing {}{} portions".format(target_path, len(self.data), " raw" if raw else ''))
|
|
|
|
with open(target_path, "wb") as f:
|
|
# Write the magic header
|
|
if not raw:
|
|
f.write("{} '{}'\n".format(self.HEADER, channel).encode("UTF-8"))
|
|
|
|
# Put portions in order
|
|
self.sort()
|
|
|
|
# Write each portion
|
|
for section in self.data:
|
|
if not raw:
|
|
meta = {"name": section.name,
|
|
"network": section.network,
|
|
"channel": section.channel,
|
|
"date": section.date.strftime("%Y%m%d"),
|
|
"lines": section.lines(),
|
|
"size": section.bytes()}
|
|
f.write("{} {}\n".format(self.PORTIONHEADER, json.dumps(meta, sort_keys=True)).encode("UTF-8"))
|
|
contents = section.contents()
|
|
f.write(contents)
|
|
if not raw:
|
|
if not contents.endswith(b"\n"):
|
|
f.write(b"\n")
|
|
f.write("{} {}\n".format(self.ENDPORTIONHEADER, section.name).encode("UTF-8"))
|
|
|
|
def sort(self):
|
|
self.data.sort(key=lambda x: x.date)
|
|
|
|
def add_section(self, section):
|
|
"""
|
|
Add a portion (as a LogFile object) to the log file. If a portion with matching dates exists, it will
|
|
be replaced
|
|
"""
|
|
for s in self.data:
|
|
assert section.channel == s.channel
|
|
if s.date == section.date:
|
|
return
|
|
self.data.append(section)
|
|
|
|
def get_range(self):
|
|
"""
|
|
Return (start, end) datetime tuple of sections
|
|
"""
|
|
start = self.data[0].date
|
|
end = self.data[0].date
|
|
|
|
for item in self.data:
|
|
if item.date > end:
|
|
end = item.date
|
|
if item.date < start:
|
|
start = item.date
|
|
|
|
return (start, end, )
|
|
|
|
def limit(self, end=None, start=None):
|
|
"""
|
|
Drop all portions newer than end or older than start
|
|
"""
|
|
assert end or start, "Need an start, end, or both"
|
|
for item in self.data[:]:
|
|
if (end and item.date > end) or (start and item.date < start):
|
|
self.data.remove(item)
|
|
|
|
|
|
class LogFile(object):
|
|
|
|
def __init__(self, fname, root=None):
|
|
self.dir = root
|
|
self.name = fname
|
|
self.network = None
|
|
self.channel = None
|
|
self.date = None # datetime object for this channel
|
|
self._parse()
|
|
|
|
def _parse(self):
|
|
matches = logfile_pattern.match(self.name).groupdict()
|
|
self.network = matches["network"]
|
|
self.channel = matches["channel"]
|
|
date = matches["date"]
|
|
self.date = datetime.datetime.strptime(date, '%Y%m%d')
|
|
|
|
def contents(self):
|
|
"""
|
|
Return log contents
|
|
"""
|
|
with open(os.path.join(self.dir, self.name), "rb") as f:
|
|
return f.read()
|
|
|
|
def lines(self):
|
|
"""
|
|
Return line count
|
|
"""
|
|
lines = 0
|
|
with open(os.path.join(self.dir, self.name), "rb") as f:
|
|
for _ in f.readlines():
|
|
lines += 1
|
|
return lines
|
|
|
|
def bytes(self):
|
|
return os.path.getsize(os.path.join(self.dir, self.name))
|
|
|
|
@staticmethod
|
|
def create(fname):
|
|
return LogFile(os.path.basename(fname), root=os.path.dirname(fname))
|
|
|
|
def __str__(self):
|
|
return "<__main__.LogFile '{}'>".format(self.name)
|
|
|
|
__repr__ = __str__
|
|
|
|
|
|
class VirtualLogFile(LogFile):
|
|
def __init__(self, fname, contents):
|
|
super().__init__(fname)
|
|
self.data = contents
|
|
|
|
def contents(self):
|
|
return self.data
|
|
|
|
def lines(self):
|
|
return len(self.data.split(b'\n'))
|
|
|
|
def bytes(self):
|
|
return len(self.data)
|