Do not singletonize the retriever

This commit is contained in:
Blallo 2021-09-15 17:53:39 +02:00
parent d967440a6d
commit acc966b488
No known key found for this signature in database
GPG key ID: 0CBE577C9B72DC3F
2 changed files with 27 additions and 50 deletions

View file

@ -8,7 +8,7 @@ from time import sleep
from typing import Callable, Optional
from .config_manager import get_config
from .http_retriever import RETRIEVER
from .http_retriever import download
logger = logging.getLogger("forge")
Validator = Callable[[datetime, datetime, str], bool]
@ -23,8 +23,10 @@ async def get_timefile_exact(time) -> str:
remote_path = os.path.join(
remote_repo, time.strftime(get_config()["AUDIO_INPUT_FORMAT"])
)
if remote.startswith("http"):
local = await RETRIEVER.get(remote_path)
if remote_path.startswith("http://") or remote_path.startswith("https://"):
logger.info(f"downloading {remote_path}")
print(f"DOWNLOADING -> {remote_path}")
local = await download(remote_path)
return local
return local_path

View file

@ -1,6 +1,8 @@
# -*- encoding: utf-8 -*-
import asyncio
import os
from typing import Optional
from tempfile import mkdtemp
import aiohttp # type: ignore
@ -9,52 +11,25 @@ from .config_manager import get_config
CHUNK_SIZE = 2 ** 12
class HTTPRetriever(object):
async def download(remote: str, staging: Optional[str] = None) -> str:
"""
This class offers the `get` method to retrieve the file from the local staging path
or, if missing, from the given remote.
This will download to AUDIO_STAGING the remote file and return the local
path of the downloaded file
"""
_instance = None
def __new__(cls):
if self._instance is None:
self._instance = super().__new__(cls)
return self._instance
def __init__(self):
self.repo_path = get_config()["AUDIO_STAGING"]
self.repo = dict(
path=os.path.join(self.repo_path, path) for i in os.listdir(self.repo_path)
)
async def get(remote: str) -> str:
"""
This will look in the local staging path (ideally on a tmpfs or something
similar), and return the file from there if present. Otherwise, it will download
it from the remote.
"""
if remote in self.repo:
return self.repo[remote]
file = await self._download_from_remote(remote)
self.repo[] = file
return file
async def _download(remote: str) -> str:
"""
This will download to AUDIO_STAGING the remote file and return the local path
of the downloaded file
"""
_, filename = os.path.split(remote)
local = os.path.join(get_config()["AUDIO_STAGING"], filename)
async with aiohttp.ClientSession() as session:
async with session.get(remote) as resp:
with open(local) as f:
while True:
chunk = await resp.content.read(CHUNK_SIZE)
if not chunk:
break
f.write(chunk)
return local
RETRIEVER = HTTPRetriever()
_, filename = os.path.split(remote)
if staging:
base = staging
else:
# if no staging is specified, and you want to clean the storage
# used by techrec: rm -rf /tmp/techrec*
base = mkdtemp(prefix="techrec-", dir="/tmp")
local = os.path.join(base, filename)
async with aiohttp.ClientSession() as session:
async with session.get(remote) as resp:
with open(local, "wb") as f:
while True:
chunk = await resp.content.read(CHUNK_SIZE)
if not chunk:
break
f.write(chunk)
return local