bellingcat/auto-archiver
1
0
Fork 0
mirror of https://github.com/bellingcat/auto-archiver.git synced 2026-06-11 12:48:28 +03:00
Code Issues Packages Projects Releases Wiki Activity

Refactoring of storage code:

Browse Source 
1. Fix some bugs in local_storage
2. Refactor unit tests to not set Media.key explicitly (unless it's well-known beforehand, which it isn't)
3. Limit length of URL for 'url' type path_generator
4. Throw an error if 'save_to' of local storage is too long
5. A few other tidyups
This commit is contained in:
Patrick Robertson
2025-03-10 16:39:31 +00:00
parent e89a8da3b4
commit 770f4c8a3d
15 changed files with 74 additions and 96 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
src/auto_archiver/core/config.py
View File

@@ -14,7 +14,6 @@ from loguru import logger
from copy import deepcopy
from auto_archiver.core.consts import MODULE_TYPES
from typing import Any, List, Type, Tuple
_yaml: YAML = YAML()

2
src/auto_archiver/core/consts.py
View File

@@ -1,3 +1,5 @@
class SetupError(ValueError):
pass
MODULE_TYPES = [
'feeder',

3
src/auto_archiver/core/extractor.py
View File

@@ -81,7 +81,8 @@ class Extractor(BaseModule):
if len(to_filename) > 64:
to_filename = to_filename[-64:]
to_filename = os.path.join(self.tmp_dir, to_filename)
if verbose: logger.debug(f"downloading {url[0:50]=} {to_filename=}")
if verbose:
logger.debug(f"downloading {url[0:50]=} {to_filename=}")
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/81.0.4044.138 Safari/537.36'
}

9
src/auto_archiver/core/media.py
View File

@@ -21,14 +21,13 @@ class Media:
Represents a media file with associated properties and storage details.
Attributes:
- filename: The file path of the media.
- key: An optional identifier for the media.
- filename: The file path of the media as saved locally (temporarily, before uploading to the storage).
- urls: A list of URLs where the media is stored or accessible.
- properties: Additional metadata or transformations for the media.
- _mimetype: The media's mimetype (e.g., image/jpeg, video/mp4).
"""
filename: str
key: str = None
_key: str = None
urls: List[str] = field(default_factory=list)
properties: dict = field(default_factory=dict)
_mimetype: str = None # eg: image/jpeg
@@ -67,6 +66,10 @@ class Media:
# checks if the media is already stored in the given storage
return len(self.urls) > 0 and len(self.urls) == len(in_storage.config["steps"]["storages"])
@property
def key(self) -> str:
return self._key
def set(self, key: str, value: Any) -> Media:
self.properties[key] = value
return self

4
src/auto_archiver/core/orchestrator.py
View File

@@ -23,15 +23,13 @@ from .config import read_yaml, store_yaml, to_dot_notation, merge_dicts, is_vali
DefaultValidatingParser, UniqueAppendAction, AuthenticationJsonParseAction, DEFAULT_CONFIG_FILE
from .module import ModuleFactory, LazyBaseModule
from . import validators, Feeder, Extractor, Database, Storage, Formatter, Enricher
from .consts import MODULE_TYPES
from .consts import MODULE_TYPES, SetupError
from auto_archiver.utils.url import check_url_or_raise
if TYPE_CHECKING:
from .base_module import BaseModule
from .module import LazyBaseModule
class SetupError(ValueError):
pass
class ArchivingOrchestrator:
# instance variables

42
src/auto_archiver/core/storage.py
View File

@@ -6,6 +6,7 @@ from __future__ import annotations
from abc import abstractmethod
from typing import IO
import os
import platform
from loguru import logger
from slugify import slugify
@@ -43,17 +44,30 @@ class Storage(BaseModule):
def uploadf(self, file: IO[bytes], key: str, **kwargs: dict) -> bool:
"""
Uploads (or saves) a file to the storage service/location.
This method should not be called directly, but instead through the 'store' method,
which sets up the media for storage.
"""
pass
def upload(self, media: Media, **kwargs) -> bool:
"""
Uploads a media object to the storage service.
This method should not be called directly, but instead be called through the 'store' method,
which sets up the media for storage.
"""
logger.debug(f'[{self.__class__.__name__}] storing file {media.filename} with key {media.key}')
with open(media.filename, 'rb') as f:
return self.uploadf(f, media, **kwargs)
def set_key(self, media: Media, url: str, metadata: Metadata) -> None:
"""takes the media and optionally item info and generates a key"""
if media.key is not None and len(media.key) > 0: return
if media.key is not None and len(media.key) > 0:
# media key is already set
return
folder = metadata.get_context('folder', '')
filename, ext = os.path.splitext(media.filename)
@@ -61,10 +75,8 @@ class Storage(BaseModule):
path_generator = self.path_generator
if path_generator == "flat":
path = ""
# TODO: this is never used
filename = slugify(filename) # Ensure filename is slugified
elif path_generator == "url":
path = slugify(url)
path = slugify(url)[:70]
elif path_generator == "random":
path = random_str(24)
else:
@@ -81,25 +93,7 @@ class Storage(BaseModule):
filename = hd[:24]
else:
raise ValueError(f"Invalid filename_generator: {filename_generator}")
key = os.path.join(folder, path, f"{filename}{ext}")
if len(key) > self.max_file_length():
# truncate the path
max_path_length = self.max_file_length() - len(filename) - len(ext) - len(folder) - 1
path = path[:max_path_length]
logger.warning(f'Filename too long ({len(key)} characters), truncating to {self.max_file_length()} characters')
key = os.path.join(folder, path, f"{filename}{ext}")
media.key = key
def max_file_length(self) -> int:
"""
Returns the maximum length of a file name that can be stored in the storage service.
Files are truncated if they exceed this length.
Override this method in subclasses if the storage service has a different maximum file length.
"""
return 255 # safe max file length for most filesystems (macOS, Windows, Linux)
media._key = key

1
src/auto_archiver/modules/cli_feeder/cli_feeder.py
View File

@@ -15,7 +15,6 @@ class CLIFeeder(Feeder):
for url in urls:
logger.debug(f"Processing {url}")
m = Metadata().set_url(url)
m.set_context("folder", "cli")
yield m
logger.success(f"Processed {len(urls)} URL(s)")

43
src/auto_archiver/modules/local_storage/local_storage.py
View File

@@ -6,37 +6,34 @@ from loguru import logger
from auto_archiver.core import Media
from auto_archiver.core import Storage
from auto_archiver.core.consts import SetupError
class LocalStorage(Storage):
MAX_FILE_LENGTH = 255
def setup(self) -> None:
if len(self.save_to) > 200:
raise SetupError(f"Your save_to path is long, this will cause issues saving files on your computer. Please use a shorter path")
def get_cdn_url(self, media: Media) -> str:
# TODO: is this viable with Storage.configs on path/filename?
dest = os.path.join(self.save_to, media.key)
dest = media.key
if self.save_absolute:
dest = os.path.abspath(dest)
return dest
def set_key(self, media, url, metadata):
# clarify we want to save the file to the save_to folder
old_folder = metadata.get('folder', '')
metadata.set_context('folder', os.path.join(self.save_to, metadata.get('folder', '')))
super().set_key(media, url, metadata)
# don't impact other storages that might want a different 'folder' set
metadata.set_context('folder', old_folder)
def upload(self, media: Media, **kwargs) -> bool:
# override parent so that we can use shutil.copy2 and keep metadata
dest = os.path.join(self.save_to, media.key)
if len(dest) > self.max_file_length():
old_dest_length = len(dest)
filename, ext = os.path.splitext(media.key)
dir, filename = os.path.split(filename)
# see whether we should truncate filename or dir
if len(dir) > len(filename):
dir = dir[:self.MAX_FILE_LENGTH - len(self.save_to) - len(ext) - len(filename) - 1]
else:
filename = filename[:self.MAX_FILE_LENGTH - len(self.save_to) - len(ext) - len(filename) - 1]
# override media.key
media.key = os.path.join(dir, f"{filename}{ext}")
dest = os.path.join(self.save_to, dir, f"{filename}{ext}")
logger.warning(f'Filename too long ({old_dest_length} characters), truncating to {len(dest)} characters')
dest = media.key
os.makedirs(os.path.dirname(dest), exist_ok=True)
logger.debug(f'[{self.__class__.__name__}] storing file {media.filename} with key {media.key} to {dest}')
@@ -46,7 +43,5 @@ class LocalStorage(Storage):
return True
# must be implemented even if unused
def uploadf(self, file: IO[bytes], key: str, **kwargs: dict) -> bool: pass
def max_file_length(self):
return self.MAX_FILE_LENGTH
def uploadf(self, file: IO[bytes], key: str, **kwargs: dict) -> bool:
pass

13
src/auto_archiver/modules/s3_storage/s3_storage.py
View File

@@ -42,7 +42,7 @@ class S3Storage(Storage):
logger.warning(f"Unable to get mimetype for {media.key=}, error: {e}")
self.s3.upload_fileobj(file, Bucket=self.bucket, Key=media.key, ExtraArgs=extra_args)
return True
def is_upload_needed(self, media: Media) -> bool:
if self.random_no_duplicate:
# checks if a folder with the hash already exists, if so it skips the upload
@@ -50,13 +50,13 @@ class S3Storage(Storage):
path = os.path.join(NO_DUPLICATES_FOLDER, hd[:24])
if existing_key:=self.file_in_folder(path):
media.key = existing_key
media._key = existing_key
media.set("previously archived", True)
logger.debug(f"skipping upload of {media.filename} because it already exists in {media.key}")
return False
_, ext = os.path.splitext(media.key)
media.key = os.path.join(path, f"{random_str(24)}{ext}")
media._key = os.path.join(path, f"{random_str(24)}{ext}")
return True
def file_in_folder(self, path:str) -> str:
@@ -66,9 +66,4 @@ class S3Storage(Storage):
resp = self.s3.list_objects(Bucket=self.bucket, Prefix=path, Delimiter='/', MaxKeys=1)
if 'Contents' in resp:
return resp['Contents'][0]['Key']
return False
def max_file_length(self):
# Amazon AWS max file length is 1024, but we will use 1000 to be safe
return 1000
return False