Abstract away treating a file as a stream

Makes upload_file cleaner (context manager instead of try-finally)
and helps keep the logic "we may own this stream and need to close
it or not" separated.

It can be overengineered to allow nesting the same instance in
context managers and also provide implementations for synchronous
context managers but it adds quite a bit of unnecessary complexity
for the time being. YAGNI.
This commit is contained in:
Lonami Exo 2020-09-14 16:20:44 +02:00
parent 9ec5707c32
commit 219b4ecb77
3 changed files with 147 additions and 80 deletions

View File

@ -540,47 +540,11 @@ class UploadMethods:
if isinstance(file, (types.InputFile, types.InputFileBig)):
return file # Already uploaded
if not file_name and getattr(file, 'name', None):
file_name = file.name
pos = 0
async with helpers._FileStream(file, file_size=file_size) as stream:
# Opening the stream will determine the correct file size
file_size = stream.file_size
if file_size is not None:
pass # do nothing as it's already kwown
elif isinstance(file, str):
file_size = os.path.getsize(file)
stream = open(file, 'rb')
close_stream = True
elif isinstance(file, bytes):
file_size = len(file)
stream = io.BytesIO(file)
close_stream = True
else:
if not callable(getattr(file, 'read', None)):
raise TypeError('file description should have a `read` method')
if callable(getattr(file, 'seekable', None)):
seekable = await helpers._maybe_await(file.seekable())
else:
seekable = False
if seekable:
pos = await helpers._maybe_await(file.tell())
await helpers._maybe_await(file.seek(0, os.SEEK_END))
file_size = await helpers._maybe_await(file.tell())
await helpers._maybe_await(file.seek(pos, os.SEEK_SET))
stream = file
close_stream = False
else:
self._log[__name__].warning(
'Could not determine file size beforehand so the entire '
'file will be read in-memory')
data = await helpers._maybe_await(file.read())
stream = io.BytesIO(data)
close_stream = True
file_size = len(data)
# File will now either be a string or bytes
if not part_size_kb:
part_size_kb = utils.get_appropriated_part_size(file_size)
@ -595,16 +559,13 @@ class UploadMethods:
# Set a default file name if None was specified
file_id = helpers.generate_random_long()
if not file_name:
if isinstance(file, str):
file_name = os.path.basename(file)
else:
file_name = str(file_id)
file_name = stream.name or str(file_id)
# If the file name lacks extension, add it if possible.
# Else Telegram complains with `PHOTO_EXT_INVALID_ERROR`
# even if the uploaded image is indeed a photo.
if not os.path.splitext(file_name)[-1]:
file_name += utils._get_extension(file)
file_name += utils._get_extension(stream)
# Determine whether the file is too big (over 10MB) or not
# Telegram does make a distinction between smaller or larger files
@ -616,7 +577,6 @@ class UploadMethods:
file_size, part_count, part_size)
pos = 0
try:
for part_index in range(part_count):
# Read the file by in chunks of size part_size
part = await helpers._maybe_await(stream.read(part_size))
@ -663,9 +623,6 @@ class UploadMethods:
else:
raise RuntimeError(
'Failed to upload file part {}.'.format(part_index))
finally:
if close_stream:
await helpers._maybe_await(stream.close())
if is_big:
return types.InputFileBig(file_id, part_count, file_name)

View File

@ -1,9 +1,13 @@
"""Various helpers not related to the Telegram API itself"""
import asyncio
import io
import enum
import os
import struct
import inspect
import logging
import functools
from pathlib import Path
from hashlib import sha1
@ -13,6 +17,9 @@ class _EntityType(enum.Enum):
CHANNEL = 2
_log = logging.getLogger(__name__)
# region Multiple utilities
@ -280,4 +287,105 @@ class TotalList(list):
', '.join(repr(x) for x in self), self.total)
class _FileStream(io.IOBase):
"""
Proxy around things that represent a file and need to be used as streams
which may or not need to be closed.
This will handle `pathlib.Path`, `str` paths, in-memory `bytes`, and
anything IO-like (including `aiofiles`).
It also provides access to the name and file size (also necessary).
"""
def __init__(self, file, *, file_size=None):
if isinstance(file, Path):
file = str(file.absolute())
self._file = file
self._name = None
self._size = file_size
self._stream = None
self._close_stream = None
async def __aenter__(self):
if isinstance(self._file, str):
self._name = os.path.basename(self._file)
self._size = os.path.getsize(self._file)
self._stream = open(self._file, 'rb')
self._close_stream = True
elif isinstance(self._file, bytes):
self._size = len(self._file)
self._stream = io.BytesIO(self._file)
self._close_stream = True
elif not callable(getattr(self._file, 'read', None)):
raise TypeError('file description should have a `read` method')
elif self._size is not None:
self._name = getattr(self._file, 'name', None)
self._stream = self._file
self._close_stream = False
else:
if callable(getattr(self._file, 'seekable', None)):
seekable = await _maybe_await(self._file.seekable())
else:
seekable = False
if seekable:
pos = await _maybe_await(self._file.tell())
await _maybe_await(self._file.seek(0, os.SEEK_END))
self._size = await _maybe_await(self._file.tell())
await _maybe_await(self._file.seek(pos, os.SEEK_SET))
self._stream = self._file
self._close_stream = False
else:
_log.warning(
'Could not determine file size beforehand so the entire '
'file will be read in-memory')
data = await _maybe_await(self._file.read())
self._size = len(data)
self._stream = io.BytesIO(data)
self._close_stream = True
return self
async def __aexit__(self, exc_type, exc_val, exc_tb):
if self._close_stream and self._stream:
await _maybe_await(self._stream.close())
@property
def file_size(self):
return self._size
@property
def name(self):
return self._name
# Proxy all the methods. Doesn't need to be readable (makes multiline edits easier)
def read(self, *args, **kwargs): return self._stream.read(*args, **kwargs)
def readinto(self, *args, **kwargs): return self._stream.readinto(*args, **kwargs)
def write(self, *args, **kwargs): return self._stream.write(*args, **kwargs)
def fileno(self, *args, **kwargs): return self._stream.fileno(*args, **kwargs)
def flush(self, *args, **kwargs): return self._stream.flush(*args, **kwargs)
def isatty(self, *args, **kwargs): return self._stream.isatty(*args, **kwargs)
def readable(self, *args, **kwargs): return self._stream.readable(*args, **kwargs)
def readline(self, *args, **kwargs): return self._stream.readline(*args, **kwargs)
def readlines(self, *args, **kwargs): return self._stream.readlines(*args, **kwargs)
def seek(self, *args, **kwargs): return self._stream.seek(*args, **kwargs)
def seekable(self, *args, **kwargs): return self._stream.seekable(*args, **kwargs)
def tell(self, *args, **kwargs): return self._stream.tell(*args, **kwargs)
def truncate(self, *args, **kwargs): return self._stream.truncate(*args, **kwargs)
def writable(self, *args, **kwargs): return self._stream.writable(*args, **kwargs)
def writelines(self, *args, **kwargs): return self._stream.writelines(*args, **kwargs)
# close is special because it will be called by __del__ but we do NOT
# want to close the file unless we have to (we're just a wrapper).
# Instead, we do nothing (we should be used through the decorator which
# has its own mechanism to close the file correctly).
def close(self, *args, **kwargs):
pass
# endregion

View File

@ -611,7 +611,9 @@ def _get_metadata(file):
# The parser may fail and we don't want to crash if
# the extraction process fails.
try:
# Note: aiofiles are intentionally left out for simplicity
# Note: aiofiles are intentionally left out for simplicity.
# `helpers._FileStream` is async only for simplicity too, so can't
# reuse it here.
if isinstance(file, str):
stream = open(file, 'rb')
elif isinstance(file, bytes):