feat: full support for opendal and sync configurations between .env and docker-compose (#11754)
Signed-off-by: -LAN- <laipz8200@outlook.com>
This commit is contained in:
@@ -1,46 +1,57 @@
|
||||
import logging
|
||||
import os
|
||||
from collections.abc import Generator
|
||||
from pathlib import Path
|
||||
from urllib.parse import urlparse
|
||||
|
||||
import opendal
|
||||
from dotenv import dotenv_values
|
||||
|
||||
from configs.middleware.storage.opendal_storage_config import OpenDALScheme
|
||||
from extensions.storage.base_storage import BaseStorage
|
||||
|
||||
S3_R2_HOSTNAME = "r2.cloudflarestorage.com"
|
||||
S3_R2_COMPATIBLE_KWARGS = {
|
||||
"delete_max_size": "700",
|
||||
"disable_stat_with_override": "true",
|
||||
"region": "auto",
|
||||
}
|
||||
S3_SSE_WITH_AWS_MANAGED_IAM_KWARGS = {
|
||||
"server_side_encryption": "aws:kms",
|
||||
}
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
|
||||
def is_r2_endpoint(endpoint: str) -> bool:
|
||||
if not endpoint:
|
||||
return False
|
||||
def _get_opendal_kwargs(*, scheme: str, env_file_path: str = ".env", prefix: str = "OPENDAL_"):
|
||||
kwargs = {}
|
||||
config_prefix = prefix + scheme.upper() + "_"
|
||||
for key, value in os.environ.items():
|
||||
if key.startswith(config_prefix):
|
||||
kwargs[key[len(config_prefix) :].lower()] = value
|
||||
|
||||
parsed_url = urlparse(endpoint)
|
||||
return bool(parsed_url.hostname and parsed_url.hostname.endswith(S3_R2_HOSTNAME))
|
||||
file_env_vars = dotenv_values(env_file_path)
|
||||
for key, value in file_env_vars.items():
|
||||
if key.startswith(config_prefix) and key[len(config_prefix) :].lower() not in kwargs and value:
|
||||
kwargs[key[len(config_prefix) :].lower()] = value
|
||||
|
||||
return kwargs
|
||||
|
||||
|
||||
class OpenDALStorage(BaseStorage):
|
||||
def __init__(self, scheme: OpenDALScheme, **kwargs):
|
||||
if scheme == OpenDALScheme.FS:
|
||||
Path(kwargs["root"]).mkdir(parents=True, exist_ok=True)
|
||||
def __init__(self, scheme: str, **kwargs):
|
||||
kwargs = kwargs or _get_opendal_kwargs(scheme=scheme)
|
||||
|
||||
if scheme == "fs":
|
||||
root = kwargs.get("root", "storage")
|
||||
Path(root).mkdir(parents=True, exist_ok=True)
|
||||
|
||||
# self.op = opendal.Operator(scheme=scheme, **kwargs)
|
||||
self.op = opendal.Operator(scheme=scheme, **kwargs)
|
||||
logger.debug(f"opendal operator created with scheme {scheme}")
|
||||
retry_layer = opendal.layers.RetryLayer(max_times=3, factor=2.0, jitter=True)
|
||||
self.op = self.op.layer(retry_layer)
|
||||
logger.debug("added retry layer to opendal operator")
|
||||
|
||||
def save(self, filename: str, data: bytes) -> None:
|
||||
self.op.write(path=filename, bs=data)
|
||||
logger.debug(f"file {filename} saved")
|
||||
|
||||
def load_once(self, filename: str) -> bytes:
|
||||
if not self.exists(filename):
|
||||
raise FileNotFoundError("File not found")
|
||||
|
||||
return self.op.read(path=filename)
|
||||
content = self.op.read(path=filename)
|
||||
logger.debug(f"file {filename} loaded")
|
||||
return content
|
||||
|
||||
def load_stream(self, filename: str) -> Generator:
|
||||
if not self.exists(filename):
|
||||
@@ -50,6 +61,7 @@ class OpenDALStorage(BaseStorage):
|
||||
file = self.op.open(path=filename, mode="rb")
|
||||
while chunk := file.read(batch_size):
|
||||
yield chunk
|
||||
logger.debug(f"file {filename} loaded as stream")
|
||||
|
||||
def download(self, filename: str, target_filepath: str):
|
||||
if not self.exists(filename):
|
||||
@@ -57,16 +69,22 @@ class OpenDALStorage(BaseStorage):
|
||||
|
||||
with Path(target_filepath).open("wb") as f:
|
||||
f.write(self.op.read(path=filename))
|
||||
logger.debug(f"file {filename} downloaded to {target_filepath}")
|
||||
|
||||
def exists(self, filename: str) -> bool:
|
||||
# FIXME this is a workaround for opendal python-binding do not have a exists method and no better
|
||||
# error handler here when opendal python-binding has a exists method, we should use it
|
||||
# more https://github.com/apache/opendal/blob/main/bindings/python/src/operator.rs
|
||||
try:
|
||||
return self.op.stat(path=filename).mode.is_file()
|
||||
except Exception as e:
|
||||
res = self.op.stat(path=filename).mode.is_file()
|
||||
logger.debug(f"file {filename} checked")
|
||||
return res
|
||||
except Exception:
|
||||
return False
|
||||
|
||||
def delete(self, filename: str):
|
||||
if self.exists(filename):
|
||||
self.op.delete(path=filename)
|
||||
logger.debug(f"file {filename} deleted")
|
||||
return
|
||||
logger.debug(f"file {filename} not found, skip delete")
|
||||
|
Reference in New Issue
Block a user