lbry-sdk/torba/server/util.py

360 lines
10 KiB
Python
Raw Normal View History

# Copyright (c) 2016-2017, Neil Booth
#
# All rights reserved.
#
# The MIT License (MIT)
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
# LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
# OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
# WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
# and warranty status of this software.
"""Miscellaneous utility classes and functions."""
import array
import inspect
from ipaddress import ip_address
import logging
import re
import sys
from collections import Container, Mapping
from struct import pack, Struct
# Logging utilities
class ConnectionLogger(logging.LoggerAdapter):
"""Prepends a connection identifier to a logging message."""
def process(self, msg, kwargs):
conn_id = self.extra.get('conn_id', 'unknown')
return f'[{conn_id}] {msg}', kwargs
class CompactFormatter(logging.Formatter):
"""Strips the module from the logger name to leave the class only."""
def format(self, record):
record.name = record.name.rpartition('.')[-1]
return super().format(record)
def make_logger(name, *, handler, level):
"""Return the root ElectrumX logger."""
logger = logging.getLogger(name)
logger.addHandler(handler)
logger.setLevel(logging.INFO)
logger.propagate = False
return logger
def class_logger(path, classname):
"""Return a hierarchical logger for a class."""
return logging.getLogger(path).getChild(classname)
# Method decorator. To be used for calculations that will always
# deliver the same result. The method cannot take any arguments
# and should be accessed as an attribute.
2018-11-26 02:51:41 +01:00
class cachedproperty:
def __init__(self, f):
self.f = f
def __get__(self, obj, type):
obj = obj or type
value = self.f(obj)
setattr(obj, self.f.__name__, value)
return value
def formatted_time(t, sep=' '):
"""Return a number of seconds as a string in days, hours, mins and
maybe secs."""
t = int(t)
fmts = (('{:d}d', 86400), ('{:02d}h', 3600), ('{:02d}m', 60))
parts = []
for fmt, n in fmts:
val = t // n
if parts or val:
parts.append(fmt.format(val))
t %= n
if len(parts) < 3:
parts.append('{:02d}s'.format(t))
return sep.join(parts)
def deep_getsizeof(obj):
"""Find the memory footprint of a Python object.
Based on code from code.tutsplus.com: http://goo.gl/fZ0DXK
This is a recursive function that drills down a Python object graph
like a dictionary holding nested dictionaries with lists of lists
and tuples and sets.
The sys.getsizeof function does a shallow size of only. It counts each
object inside a container as pointer only regardless of how big it
really is.
"""
ids = set()
def size(o):
if id(o) in ids:
return 0
r = sys.getsizeof(o)
ids.add(id(o))
if isinstance(o, (str, bytes, bytearray, array.array)):
return r
if isinstance(o, Mapping):
return r + sum(size(k) + size(v) for k, v in o.items())
if isinstance(o, Container):
return r + sum(size(x) for x in o)
return r
return size(obj)
def subclasses(base_class, strict=True):
"""Return a list of subclasses of base_class in its module."""
def select(obj):
return (inspect.isclass(obj) and issubclass(obj, base_class) and
(not strict or obj != base_class))
pairs = inspect.getmembers(sys.modules[base_class.__module__], select)
return [pair[1] for pair in pairs]
def chunks(items, size):
"""Break up items, an iterable, into chunks of length size."""
for i in range(0, len(items), size):
yield items[i: i + size]
def resolve_limit(limit):
if limit is None:
return -1
assert isinstance(limit, int) and limit >= 0
return limit
def bytes_to_int(be_bytes):
"""Interprets a big-endian sequence of bytes as an integer"""
return int.from_bytes(be_bytes, 'big')
def int_to_bytes(value):
"""Converts an integer to a big-endian sequence of bytes"""
return value.to_bytes((value.bit_length() + 7) // 8, 'big')
def increment_byte_string(bs):
"""Return the lexicographically next byte string of the same length.
Return None if there is none (when the input is all 0xff bytes)."""
for n in range(1, len(bs) + 1):
if bs[-n] != 0xff:
return bs[:-n] + bytes([bs[-n] + 1]) + bytes(n - 1)
return None
2018-11-26 02:51:41 +01:00
class LogicalFile:
"""A logical binary file split across several separate files on disk."""
def __init__(self, prefix, digits, file_size):
digit_fmt = '{' + ':0{:d}d'.format(digits) + '}'
self.filename_fmt = prefix + digit_fmt
self.file_size = file_size
def read(self, start, size=-1):
"""Read up to size bytes from the virtual file, starting at offset
start, and return them.
If size is -1 all bytes are read."""
parts = []
while size != 0:
try:
with self.open_file(start, False) as f:
part = f.read(size)
if not part:
break
except FileNotFoundError:
break
parts.append(part)
start += len(part)
if size > 0:
size -= len(part)
return b''.join(parts)
def write(self, start, b):
"""Write the bytes-like object, b, to the underlying virtual file."""
while b:
size = min(len(b), self.file_size - (start % self.file_size))
with self.open_file(start, True) as f:
f.write(b if size == len(b) else b[:size])
b = b[size:]
start += size
def open_file(self, start, create):
"""Open the virtual file and seek to start. Return a file handle.
Raise FileNotFoundError if the file does not exist and create
is False.
"""
file_num, offset = divmod(start, self.file_size)
filename = self.filename_fmt.format(file_num)
f = open_file(filename, create)
f.seek(offset)
return f
def open_file(filename, create=False):
"""Open the file name. Return its handle."""
try:
return open(filename, 'rb+')
except FileNotFoundError:
if create:
return open(filename, 'wb+')
raise
def open_truncate(filename):
"""Open the file name. Return its handle."""
return open(filename, 'wb+')
def address_string(address):
"""Return an address as a correctly formatted string."""
fmt = '{}:{:d}'
host, port = address
try:
host = ip_address(host)
except ValueError:
pass
else:
if host.version == 6:
fmt = '[{}]:{:d}'
return fmt.format(host, port)
# See http://stackoverflow.com/questions/2532053/validate-a-hostname-string
# Note underscores are valid in domain names, but strictly invalid in host
# names. We ignore that distinction.
SEGMENT_REGEX = re.compile("(?!-)[A-Z_\\d-]{1,63}(?<!-)$", re.IGNORECASE)
def is_valid_hostname(hostname):
if len(hostname) > 255:
return False
# strip exactly one dot from the right, if present
if hostname and hostname[-1] == ".":
hostname = hostname[:-1]
return all(SEGMENT_REGEX.match(x) for x in hostname.split("."))
def protocol_tuple(s):
"""Converts a protocol version number, such as "1.0" to a tuple (1, 0).
If the version number is bad, (0, ) indicating version 0 is returned."""
try:
return tuple(int(part) for part in s.split('.'))
except Exception:
return (0, )
def version_string(ptuple):
"""Convert a version tuple such as (1, 2) to "1.2".
There is always at least one dot, so (1, ) becomes "1.0"."""
while len(ptuple) < 2:
ptuple += (0, )
return '.'.join(str(p) for p in ptuple)
def protocol_version(client_req, min_tuple, max_tuple):
"""Given a client's protocol version string, return a pair of
protocol tuples:
(negotiated version, client min request)
If the request is unsupported, the negotiated protocol tuple is
None.
"""
if client_req is None:
client_min = client_max = min_tuple
else:
if isinstance(client_req, list) and len(client_req) == 2:
client_min, client_max = client_req
else:
client_min = client_max = client_req
client_min = protocol_tuple(client_min)
client_max = protocol_tuple(client_max)
result = min(client_max, max_tuple)
if result < max(client_min, min_tuple) or result == (0, ):
result = None
return result, client_min
struct_le_i = Struct('<i')
struct_le_q = Struct('<q')
struct_le_H = Struct('<H')
struct_le_I = Struct('<I')
struct_le_Q = Struct('<Q')
struct_be_H = Struct('>H')
struct_be_I = Struct('>I')
structB = Struct('B')
unpack_le_int32_from = struct_le_i.unpack_from
unpack_le_int64_from = struct_le_q.unpack_from
unpack_le_uint16_from = struct_le_H.unpack_from
unpack_le_uint32_from = struct_le_I.unpack_from
unpack_le_uint64_from = struct_le_Q.unpack_from
unpack_be_uint16_from = struct_be_H.unpack_from
unpack_be_uint32_from = struct_be_I.unpack_from
pack_le_int32 = struct_le_i.pack
pack_le_int64 = struct_le_q.pack
pack_le_uint16 = struct_le_H.pack
pack_le_uint32 = struct_le_I.pack
pack_le_uint64 = struct_le_Q.pack
pack_be_uint16 = struct_be_H.pack
pack_be_uint32 = struct_be_I.pack
pack_byte = structB.pack
hex_to_bytes = bytes.fromhex
def pack_varint(n):
if n < 253:
return pack_byte(n)
if n < 65536:
return pack_byte(253) + pack_le_uint16(n)
if n < 4294967296:
return pack_byte(254) + pack_le_uint32(n)
return pack_byte(255) + pack_le_uint64(n)
def pack_varbytes(data):
return pack_varint(len(data)) + data