mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-24 19:54:25 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Add URL-specific method allow/deny lists

Browse source 
Allows enabling only allow-listed extractors or disabling specific
deny-listed extractors for a regular expression matched against an added
site's URL.
This commit is contained in:
Ross Williams 2023-07-31 11:34:03 -04:00
parent 46e80dd509
commit b44f7e68b1
3 changed files with 85 additions and 16 deletions
Download patch file Download diff file Expand all files Collapse all files

47
archivebox/extractors/__init__.py
View file

@ -4,12 +4,16 @@ import os
import sys
from pathlib import Path
from typing import Optional, List, Iterable, Union
from typing import Callable, Optional, List, Iterable, TypeAlias, Union
from datetime import datetime, timezone
from django.db.models import QuerySet
from ..config import (
SAVE_ALLOWLIST_PTN,
SAVE_DENYLIST_PTN,
)
from ..core.settings import ERROR_LOG
from ..index.schema import Link
from ..index.schema import ArchiveResult, Link
from ..index.sql import write_link_to_sql_index
from ..index import (
load_link_details,
@ -42,7 +46,11 @@ from .archive_org import should_save_archive_dot_org, save_archive_dot_org
from .headers import should_save_headers, save_headers
def get_default_archive_methods():
ShouldSaveFunction: TypeAlias = Callable[[Link, Optional[Path], Optional[bool]], bool]
SaveFunction: TypeAlias = Callable[[Link, Optional[Path], int], ArchiveResult]
ArchiveMethodEntry: TypeAlias = tuple[str, ShouldSaveFunction, SaveFunction]
def get_default_archive_methods() -> List[ArchiveMethodEntry]:
return [
('favicon', should_save_favicon, save_favicon),
('headers', should_save_headers, save_headers),
@ -59,14 +67,31 @@ def get_default_archive_methods():
('archive_org', should_save_archive_dot_org, save_archive_dot_org),
]
@enforce_types
def get_archive_methods_for_link(link: Link) -> Iterable[ArchiveMethodEntry]:
DEFAULT_METHODS = get_default_archive_methods()
allowed_methods = {
m for pat, methods in
SAVE_ALLOWLIST_PTN.items()
if pat.search(link.url)
for m in methods
} or { m[0] for m in DEFAULT_METHODS }
denied_methods = {
m for pat, methods in
SAVE_DENYLIST_PTN.items()
if pat.search(link.url)
for m in methods
}
allowed_methods -= denied_methods
return (m for m in DEFAULT_METHODS if m[0] in allowed_methods)
ARCHIVE_METHODS_INDEXING_PRECEDENCE = [('readability', 1), ('singlefile', 2), ('dom', 3), ('wget', 4)]
@enforce_types
def ignore_methods(to_ignore: List[str]):
def ignore_methods(to_ignore: List[str]) -> Iterable[str]:
ARCHIVE_METHODS = get_default_archive_methods()
methods = filter(lambda x: x[0] not in to_ignore, ARCHIVE_METHODS)
methods = map(lambda x: x[0], methods)
return list(methods)
return [x[0] for x in ARCHIVE_METHODS if x[0] not in to_ignore]
@enforce_types
def archive_link(link: Link, overwrite: bool=False, methods: Optional[Iterable[str]]=None, out_dir: Optional[Path]=None) -> Link:
@ -79,11 +104,11 @@ def archive_link(link: Link, overwrite: bool=False, methods: Optional[Iterable[s
except Snapshot.DoesNotExist:
snapshot = write_link_to_sql_index(link)
ARCHIVE_METHODS = get_default_archive_methods()
active_methods = get_archive_methods_for_link(link)
if methods:
ARCHIVE_METHODS = [
method for method in ARCHIVE_METHODS
active_methods = [
method for method in active_methods
if method[0] in methods
]
@ -100,7 +125,7 @@ def archive_link(link: Link, overwrite: bool=False, methods: Optional[Iterable[s
stats = {'skipped': 0, 'succeeded': 0, 'failed': 0}
start_ts = datetime.now(timezone.utc)
for method_name, should_run, method_function in ARCHIVE_METHODS:
for method_name, should_run, method_function in active_methods:
try:
if method_name not in link.history:
link.history[method_name] = []