alex/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2026-04-06 07:47:53 +10:00
Code Issues Packages Projects Releases Wiki Activity

fix lint

Browse Source
This commit is contained in:
Nick Sweeting
2026-03-15 18:45:29 -07:00
parent f97725d16f
commit 934e02695b
111 changed files with 919 additions and 461 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
archivebox/core/admin_site.py
View File

@@ -1,8 +1,8 @@
__package__ = 'archivebox.core'
from django.contrib import admin
from admin_data_views.admin import get_app_list, admin_data_index_view, get_admin_data_urls, get_urls
import archivebox
class ArchiveBoxAdmin(admin.AdminSite):
site_header = 'ArchiveBox'
@@ -20,7 +20,6 @@ archivebox_admin = ArchiveBoxAdmin()
# patch admin with methods to add data views (implemented by admin_data_views package)
# https://github.com/MrThearMan/django-admin-data-views
# https://mrthearman.github.io/django-admin-data-views/setup/
from admin_data_views.admin import get_app_list, admin_data_index_view, get_admin_data_urls, get_urls
archivebox_admin.get_app_list = get_app_list.__get__(archivebox_admin, ArchiveBoxAdmin)
archivebox_admin.admin_data_index_view = admin_data_index_view.__get__(archivebox_admin, ArchiveBoxAdmin) # type: ignore
archivebox_admin.get_admin_data_urls = get_admin_data_urls.__get__(archivebox_admin, ArchiveBoxAdmin) # type: ignore

6
archivebox/core/admin_snapshots.py
View File

@@ -26,7 +26,7 @@ from archivebox.base_models.admin import BaseModelAdmin, ConfigEditorMixin
from archivebox.workers.tasks import bg_archive_snapshots, bg_add
from archivebox.core.models import Tag, Snapshot, ArchiveResult
from archivebox.core.admin_archiveresults import ArchiveResultInline, render_archiveresults_list
from archivebox.core.admin_archiveresults import render_archiveresults_list
from archivebox.core.widgets import TagEditorWidget, InlineTagEditorWidget
@@ -712,8 +712,6 @@ class SnapshotAdmin(SearchResultsAdminMixin, ConfigEditorMixin, BaseModelAdmin):
description="🔁 Redo Failed"
)
def update_snapshots(self, request, queryset):
count = queryset.count()
queued = bg_archive_snapshots(queryset, kwargs={"overwrite": False, "out_dir": DATA_DIR})
messages.success(
@@ -741,8 +739,6 @@ class SnapshotAdmin(SearchResultsAdminMixin, ConfigEditorMixin, BaseModelAdmin):
description="🔄 Redo"
)
def overwrite_snapshots(self, request, queryset):
count = queryset.count()
queued = bg_archive_snapshots(queryset, kwargs={"overwrite": True, "out_dir": DATA_DIR})
messages.success(

2
archivebox/core/apps.py
View File

@@ -60,7 +60,7 @@ class CoreConfig(AppConfig):
from archivebox.workers.orchestrator import Orchestrator
Process.cleanup_stale_running()
machine = Machine.current()
Machine.current()
if not Orchestrator.is_running():
Orchestrator(exit_on_idle=False).start()

3
archivebox/core/asgi.py
View File

@@ -8,11 +8,10 @@ https://docs.djangoproject.com/en/stable/howto/deployment/asgi/
"""
from archivebox.config.django import setup_django
from django.core.asgi import get_asgi_application
setup_django(in_memory_db=False, check_db=True)
from django.core.asgi import get_asgi_application
# Standard Django ASGI application (no websockets/channels needed)
application = get_asgi_application()

5
archivebox/core/forms.py
View File

@@ -6,6 +6,7 @@ from archivebox.misc.util import URL_REGEX
from taggit.utils import edit_string_for_tags, parse_tags
from archivebox.base_models.admin import KeyValueWidget
from archivebox.crawls.schedule_utils import validate_schedule
from archivebox.hooks import get_plugins
DEPTH_CHOICES = (
('0', 'depth = 0 (archive just these URLs)'),
@@ -15,7 +16,6 @@ DEPTH_CHOICES = (
('4', 'depth = 4 (+ URLs four hops away)'),
)
from archivebox.hooks import get_plugins
def get_plugin_choices():
"""Get available extractor plugins from discovered hooks."""
@@ -210,15 +210,18 @@ class AddLinkForm(forms.Form):
return schedule
class TagWidgetMixin:
def format_value(self, value):
if value is not None and not isinstance(value, str):
value = edit_string_for_tags(value)
return super().format_value(value)
class TagWidget(TagWidgetMixin, forms.TextInput):
pass
class TagField(forms.CharField):
widget = TagWidget

1
archivebox/core/middleware.py
View File

@@ -17,7 +17,6 @@ from archivebox.config import VERSION
from archivebox.config.version import get_COMMIT_HASH
from archivebox.core.host_utils import (
build_admin_url,
build_api_url,
build_web_url,
get_api_host,
get_admin_host,

4
archivebox/core/migrations/0006_auto_20201012_1520.py
View File

@@ -7,10 +7,8 @@ def forwards_func(apps, schema_editor):
SnapshotModel = apps.get_model("core", "Snapshot")
TagModel = apps.get_model("core", "Tag")
db_alias = schema_editor.connection.alias
snapshots = SnapshotModel.objects.all()
for snapshot in snapshots:
tags = snapshot.tags
tag_set = (
set(tag.strip() for tag in (snapshot.tags_old or '').split(','))
)
@@ -23,9 +21,7 @@ def forwards_func(apps, schema_editor):
def reverse_func(apps, schema_editor):
SnapshotModel = apps.get_model("core", "Snapshot")
TagModel = apps.get_model("core", "Tag")
db_alias = schema_editor.connection.alias
snapshots = SnapshotModel.objects.all()
for snapshot in snapshots:
tags = snapshot.tags.values_list("name", flat=True)

2
archivebox/core/migrations/0007_archiveresult.py
View File

@@ -43,7 +43,7 @@ def forwards_func(apps, schema_editor):
try:
with open(out_dir / "index.json", "r") as f:
fs_index = json.load(f)
except Exception as e:
except Exception:
continue
history = fs_index["history"]

1
archivebox/core/migrations/0023_upgrade_to_0_9_0.py
View File

@@ -234,7 +234,6 @@ def upgrade_core_tables(apps, schema_editor):
tag_has_data = cursor.fetchone()[0] > 0
if tag_has_data:
tag_cols = get_table_columns('core_tag')
cursor.execute("PRAGMA table_info(core_tag)")
tag_id_type = None
for row in cursor.fetchall():

1
archivebox/core/migrations/0024_assign_default_crawl.py
View File

@@ -2,7 +2,6 @@
# Creates a default crawl for v0.7.2 migrated snapshots and makes crawl_id NOT NULL
from django.db import migrations, models
import uuid
def create_default_crawl_and_assign_snapshots(apps, schema_editor):

2
archivebox/core/migrations/0027_copy_archiveresult_to_process.py
View File

@@ -347,7 +347,7 @@ def copy_archiveresult_data_to_process(apps, schema_editor):
migrated_count += 1
if i == 0:
print(f'DEBUG 0027: Linked ArchiveResult to Process')
print('DEBUG 0027: Linked ArchiveResult to Process')
except Exception as e:
print(f'✗ Error migrating ArchiveResult {ar_id}: {e}')

75
archivebox/core/models.py
View File

@@ -1,6 +1,6 @@
__package__ = 'archivebox.core'
from typing import Optional, Dict, Iterable, Any, List, TYPE_CHECKING
from typing import Optional, Dict, Iterable, Any, List
from archivebox.uuid_compat import uuid7
from datetime import datetime, timedelta
from django_stubs_ext.db.models import TypedModelMeta
@@ -12,19 +12,18 @@ from pathlib import Path
from statemachine import State, registry
from django.db import models
from django.db.models import QuerySet, Value, Case, When, IntegerField
from django.db.models import QuerySet
from django.utils.functional import cached_property
from django.utils.text import slugify
from django.utils import timezone
from django.core.cache import cache
from django.urls import reverse, reverse_lazy
from django.urls import reverse_lazy
from django.contrib import admin
from django.conf import settings
from archivebox.config import CONSTANTS
from archivebox.misc.system import get_dir_size, atomic_write
from archivebox.misc.util import parse_date, base_url, domain as url_domain, to_json, ts_to_date_str, urlencode, htmlencode, urldecode
from archivebox.misc.hashing import get_dir_info
from archivebox.misc.util import parse_date, domain as url_domain, to_json, ts_to_date_str, urlencode, htmlencode, urldecode
from archivebox.hooks import (
get_plugins, get_plugin_name, get_plugin_icon,
)
@@ -186,7 +185,7 @@ class SnapshotQuerySet(models.QuerySet):
for pattern in patterns:
try:
qsearch |= query_search_index(pattern)
except:
except BaseException:
raise SystemExit(2)
return self.all() & qsearch
@@ -344,8 +343,6 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
@property
def process_set(self):
"""Get all Process objects related to this snapshot's ArchiveResults."""
import json
import json
from archivebox.machine.models import Process
return Process.objects.filter(archiveresult__snapshot_id=self.id)
@@ -458,13 +455,13 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
if not old_dir.exists() or old_dir == new_dir:
# No migration needed
print(f"[DEBUG _fs_migrate] Returning None (early return)")
print("[DEBUG _fs_migrate] Returning None (early return)")
return None
if new_dir.exists():
# New directory already exists (files already copied), but we still need cleanup
# Return cleanup info so old directory can be cleaned up
print(f"[DEBUG _fs_migrate] Returning cleanup info (new_dir exists)")
print("[DEBUG _fs_migrate] Returning cleanup info (new_dir exists)")
return (old_dir, new_dir)
new_dir.mkdir(parents=True, exist_ok=True)
@@ -499,7 +496,6 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
# Schedule cleanup AFTER transaction commits successfully
# This ensures DB changes are committed before we delete old files
from django.db import transaction
transaction.on_commit(lambda: self._cleanup_old_migration_dir(old_dir, new_dir))
# Return cleanup info for manual cleanup if needed (when called directly)
@@ -594,8 +590,8 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
domain = self.extract_domain_from_url(self.url)
return (
CONSTANTS.DATA_DIR / 'users' / username / 'snapshots' /
date_str / domain / str(self.id)
CONSTANTS.DATA_DIR / 'users' / username / 'snapshots'
/ date_str / domain / str(self.id)
)
else:
# Unknown version - use current
@@ -670,7 +666,7 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
print(f"[DEBUG load_from_directory] Found via fuzzy match: {snapshot.timestamp}")
return snapshot
elif candidates.count() > 1:
print(f"[DEBUG load_from_directory] Multiple fuzzy matches, using first")
print("[DEBUG load_from_directory] Multiple fuzzy matches, using first")
return candidates.first()
print(f"[DEBUG load_from_directory] NOT FOUND (fuzzy): {url} @ {timestamp}")
return None
@@ -767,7 +763,7 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
ts_int = int(float(ts))
# 1995-01-01 to 2035-12-31
return 788918400 <= ts_int <= 2082758400
except:
except (TypeError, ValueError, OverflowError):
return False
index_valid = is_valid_timestamp(index_timestamp) if index_timestamp else False
@@ -850,7 +846,7 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
try:
with open(json_path) as f:
index_data = json.load(f)
except:
except (OSError, TypeError, ValueError, json.JSONDecodeError):
pass
# Merge title
@@ -929,7 +925,7 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
if result_data.get('start_ts'):
try:
start_ts = parser.parse(result_data['start_ts'])
except:
except (TypeError, ValueError, OverflowError):
pass
if (plugin, start_ts) in existing:
@@ -940,7 +936,7 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
if result_data.get('end_ts'):
try:
end_ts = parser.parse(result_data['end_ts'])
except:
except (TypeError, ValueError, OverflowError):
pass
# Support both 'output' (legacy) and 'output_str' (new JSONL) field names
@@ -957,7 +953,7 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
start_ts=start_ts,
end_ts=end_ts,
)
except:
except Exception:
pass
def write_index_json(self):
@@ -1176,7 +1172,7 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
try:
shutil.move(str(snapshot_dir), str(dest))
except:
except Exception:
pass
@classmethod
@@ -1208,7 +1204,7 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
try:
cls._merge_snapshots(snapshots)
merged += 1
except:
except Exception:
pass
return merged
@@ -1244,7 +1240,7 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
try:
shutil.rmtree(dup_dir)
except:
except Exception:
pass
# Merge tags
@@ -1615,7 +1611,6 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
"""
import re
from django.utils import timezone
from archivebox.misc.util import parse_date
from archivebox.base_models.models import get_or_create_system_user_pk
from archivebox.config.common import GENERAL_CONFIG
@@ -2125,7 +2120,6 @@ class Snapshot(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWithHea
def to_dict(self, extended: bool = False) -> Dict[str, Any]:
"""Convert Snapshot to a dictionary (replacement for Link._asdict())"""
from archivebox.misc.util import ts_to_date_str
from archivebox.core.host_utils import build_snapshot_url
result = {
@@ -2283,9 +2277,9 @@ class SnapshotMachine(BaseStateMachine):
# Tick Event (polled by workers)
tick = (
queued.to.itself(unless='can_start') |
queued.to(started, cond='can_start') |
started.to(sealed, cond='is_finished')
queued.to.itself(unless='can_start')
| queued.to(started, cond='can_start')
| started.to(sealed, cond='is_finished')
)
# Manual event (can also be triggered by last ArchiveResult finishing)
@@ -2783,7 +2777,7 @@ class ArchiveResult(ModelWithOutputDir, ModelWithConfig, ModelWithNotes, ModelWi
Updates status/output fields, queues discovered URLs, and triggers indexing.
"""
from django.utils import timezone
from archivebox.hooks import BUILTIN_PLUGINS_DIR, USER_PLUGINS_DIR, run_hook, is_background_hook
from archivebox.hooks import BUILTIN_PLUGINS_DIR, USER_PLUGINS_DIR, run_hook
from archivebox.config.configset import get_config
# Get merged config with proper context
@@ -3190,16 +3184,16 @@ class ArchiveResultMachine(BaseStateMachine):
# queued → skipped (if exceeded max attempts)
# started → backoff → started (retry)
tick = (
queued.to(skipped, cond='is_exceeded_max_attempts') | # Check skip first
queued.to.itself(unless='can_start') |
queued.to(started, cond='can_start') |
started.to(succeeded, cond='is_succeeded') |
started.to(failed, cond='is_failed') |
started.to(skipped, cond='is_skipped') |
started.to(backoff, cond='is_backoff') |
backoff.to(skipped, cond='is_exceeded_max_attempts') | # Check skip from backoff too
backoff.to.itself(unless='can_start') |
backoff.to(started, cond='can_start')
queued.to(skipped, cond='is_exceeded_max_attempts') # Check skip first
| queued.to.itself(unless='can_start')
| queued.to(started, cond='can_start')
| started.to(succeeded, cond='is_succeeded')
| started.to(failed, cond='is_failed')
| started.to(skipped, cond='is_skipped')
| started.to(backoff, cond='is_backoff')
| backoff.to(skipped, cond='is_exceeded_max_attempts') # Check skip from backoff too
| backoff.to.itself(unless='can_start')
| backoff.to(started, cond='can_start')
# Removed redundant transitions: backoff.to(succeeded/failed/skipped)
# Reason: backoff should always retry→started, then started→final states
)
@@ -3241,8 +3235,8 @@ class ArchiveResultMachine(BaseStateMachine):
"""Check if we should backoff and retry later."""
# Backoff if status is still started (plugin didn't complete) and output_str is empty
return (
self.archiveresult.status == ArchiveResult.StatusChoices.STARTED and
not self.archiveresult.output_str
self.archiveresult.status == ArchiveResult.StatusChoices.STARTED
and not self.archiveresult.output_str
)
def is_finished(self) -> bool:
@@ -3286,7 +3280,6 @@ class ArchiveResultMachine(BaseStateMachine):
@started.enter
def enter_started(self):
from archivebox.machine.models import NetworkInterface
# Update Process with network interface
if self.archiveresult.process_id:

13
archivebox/core/settings.py
View File

@@ -6,6 +6,7 @@ import inspect
from pathlib import Path
from django.conf.locale.en import formats as en_formats # type: ignore
from django.utils.crypto import get_random_string
import archivebox
@@ -13,6 +14,7 @@ import archivebox
from archivebox.config import DATA_DIR, PACKAGE_DIR, ARCHIVE_DIR, CONSTANTS # noqa
from archivebox.config.common import SHELL_CONFIG, SERVER_CONFIG, STORAGE_CONFIG # noqa
from archivebox.core.host_utils import normalize_base_url, get_admin_base_url, get_api_base_url
from .settings_logging import SETTINGS_LOGGING
IS_MIGRATING = "makemigrations" in sys.argv[:3] or "migrate" in sys.argv[:3]
@@ -54,8 +56,8 @@ INSTALLED_APPS = [
"django.contrib.staticfiles",
"django.contrib.admin",
# 3rd-party apps from PyPI
"signal_webhooks", # handles REST API outbound webhooks https://github.com/MrThearMan/django-signal-webhooks
"django_object_actions", # provides easy Django Admin action buttons on change views https://github.com/crccheck/django-object-actions
"signal_webhooks", # handles REST API outbound webhooks
"django_object_actions", # provides easy Django Admin action buttons on change views
# Our ArchiveBox-provided apps (use fully qualified names)
# NOTE: Order matters! Apps with migrations that depend on other apps must come AFTER their dependencies
# "archivebox.config", # ArchiveBox config settings (no models, not a real Django app)
@@ -117,7 +119,6 @@ try:
try:
# Try to import django-auth-ldap (will fail if not installed)
import django_auth_ldap
from django_auth_ldap.config import LDAPSearch
import ldap
@@ -414,9 +415,6 @@ DATETIME_FORMAT = "Y-m-d h:i:s A"
SHORT_DATETIME_FORMAT = "Y-m-d h:i:s A"
TIME_ZONE = CONSTANTS.TIMEZONE # django convention is TIME_ZONE, archivebox config uses TIMEZONE, they are equivalent
from django.conf.locale.en import formats as en_formats # type: ignore
en_formats.DATETIME_FORMAT = DATETIME_FORMAT # monkey patch en_format default with our preferred format
en_formats.SHORT_DATETIME_FORMAT = SHORT_DATETIME_FORMAT
@@ -425,9 +423,6 @@ en_formats.SHORT_DATETIME_FORMAT = SHORT_DATETIME_FORMAT
### Logging Settings
################################################################################
from .settings_logging import SETTINGS_LOGGING, LOGS_DIR, ERROR_LOG
LOGGING = SETTINGS_LOGGING

2
archivebox/core/settings_logging.py
View File

@@ -5,8 +5,6 @@ import os
import tempfile
import logging
import pydantic
import django.template
from archivebox.config import CONSTANTS

18
archivebox/core/tests.py
View File

@@ -1,5 +1,6 @@
"""Tests for the core views, especially AddView."""
import importlib
import os
import django
from unittest.mock import patch
@@ -8,13 +9,14 @@ from unittest.mock import patch
os.environ.setdefault('DJANGO_SETTINGS_MODULE', 'archivebox.settings')
django.setup()
from django.test import TestCase, Client
from django.contrib.auth.models import User
from django.urls import reverse
from archivebox.crawls.models import Crawl, CrawlSchedule
from archivebox.core.models import Tag
from archivebox.config.common import SERVER_CONFIG
TestCase = importlib.import_module('django.test').TestCase
Client = importlib.import_module('django.test').Client
User = importlib.import_module('django.contrib.auth.models').User
reverse = importlib.import_module('django.urls').reverse
Crawl = importlib.import_module('archivebox.crawls.models').Crawl
CrawlSchedule = importlib.import_module('archivebox.crawls.models').CrawlSchedule
Tag = importlib.import_module('archivebox.core.models').Tag
SERVER_CONFIG = importlib.import_module('archivebox.config.common').SERVER_CONFIG
class AddViewTests(TestCase):
@@ -252,7 +254,7 @@ class AddViewTests(TestCase):
def test_add_staff_admin_custom_config_is_allowed(self):
"""Admin users can override crawl config."""
self.client.logout()
admin_user = User.objects.create_user(
User.objects.create_user(
username='adminuser',
password='adminpass123',
email='admin@example.com',

29
archivebox/core/views.py
View File

@@ -10,7 +10,7 @@ from pathlib import Path
from urllib.parse import urlparse
from django.shortcuts import render, redirect
from django.http import HttpRequest, HttpResponse, Http404, HttpResponseForbidden
from django.http import JsonResponse, HttpRequest, HttpResponse, Http404, HttpResponseForbidden
from django.utils.html import format_html, mark_safe
from django.views import View
from django.views.generic.list import ListView
@@ -24,9 +24,8 @@ from django.utils.decorators import method_decorator
from admin_data_views.typing import TableContext, ItemContext
from admin_data_views.utils import render_with_table_view, render_with_item_view, ItemLink
import archivebox
from archivebox.config import CONSTANTS, CONSTANTS_CONFIG, DATA_DIR, VERSION
from archivebox.config.common import SHELL_CONFIG, SERVER_CONFIG, ARCHIVING_CONFIG
from archivebox.config.common import SHELL_CONFIG, SERVER_CONFIG
from archivebox.config.configset import get_flat_config, get_config, get_all_configs
from archivebox.misc.util import base_url, htmlencode, ts_to_date_str, urldecode
from archivebox.misc.serve_static import serve_static_with_byterange_support
@@ -35,6 +34,9 @@ from archivebox.search import query_search_index
from archivebox.core.models import Snapshot
from archivebox.core.host_utils import build_snapshot_url
from archivebox.core.forms import AddLinkForm
from archivebox.crawls.models import Crawl
from archivebox.hooks import get_enabled_plugins, get_plugin_name
def _files_index_target(snapshot: Snapshot, archivefile: str | None) -> str:
@@ -49,12 +51,6 @@ def _files_index_target(snapshot: Snapshot, archivefile: str | None) -> str:
return target
from archivebox.core.forms import AddLinkForm
from archivebox.crawls.models import Crawl
from archivebox.hooks import get_enabled_plugins, get_plugin_name
class HomepageView(View):
def get(self, request):
if request.user.is_authenticated:
@@ -1066,10 +1062,6 @@ class HealthCheckView(View):
status=200
)
import json
from django.http import JsonResponse
def live_progress_view(request):
"""Simple JSON endpoint for live progress status - used by admin progress monitor."""
try:
@@ -1077,7 +1069,6 @@ def live_progress_view(request):
from archivebox.crawls.models import Crawl
from archivebox.core.models import Snapshot, ArchiveResult
from archivebox.machine.models import Process, Machine
from django.db.models import Case, When, Value, IntegerField
# Get orchestrator status
orchestrator_running = Orchestrator.is_running()
@@ -1133,7 +1124,6 @@ def live_progress_view(request):
})
# Build hierarchical active crawls with nested snapshots and archive results
from django.db.models import Prefetch
running_workers = Process.objects.filter(
machine=machine,
@@ -1387,7 +1377,7 @@ def find_config_default(key: str) -> str:
return default_val
def find_config_type(key: str) -> str:
from typing import get_type_hints, ClassVar
from typing import ClassVar
CONFIGS = get_all_configs()
for config in CONFIGS.values():
@@ -1430,7 +1420,6 @@ def key_is_safe(key: str) -> bool:
def find_config_source(key: str, merged_config: dict) -> str:
"""Determine where a config value comes from."""
import os
from archivebox.machine.models import Machine
# Check if it's from archivebox.machine.config
@@ -1464,12 +1453,11 @@ def live_config_list_view(request: HttpRequest, **kwargs) -> TableContext:
# Get merged config that includes Machine.config overrides
try:
from archivebox.machine.models import Machine
machine = Machine.current()
Machine.current()
merged_config = get_config()
except Exception as e:
except Exception:
# Fallback if Machine model not available
merged_config = get_config()
machine = None
rows = {
"Section": [],
@@ -1525,7 +1513,6 @@ def live_config_list_view(request: HttpRequest, **kwargs) -> TableContext:
@render_with_item_view
def live_config_value_view(request: HttpRequest, key: str, **kwargs) -> ItemContext:
import os
from archivebox.machine.models import Machine
from archivebox.config.configset import BaseConfigSet

3
archivebox/core/widgets.py
View File

@@ -343,20 +343,17 @@ class InlineTagEditorWidget(TagEditorWidget):
snapshot_id = snapshot_id or self.snapshot_id
# Parse value to get list of tag dicts with id and name
tags = []
tag_data = []
if value:
if hasattr(value, 'all'): # QuerySet
for tag in value.all():
tag_data.append({'id': tag.pk, 'name': tag.name})
tag_data.sort(key=lambda x: x['name'].lower())
tags = [t['name'] for t in tag_data]
elif isinstance(value, (list, tuple)):
if value and hasattr(value[0], 'name'):
for tag in value:
tag_data.append({'id': tag.pk, 'name': tag.name})
tag_data.sort(key=lambda x: x['name'].lower())
tags = [t['name'] for t in tag_data]
widget_id_raw = f"inline_tags_{snapshot_id}" if snapshot_id else (attrs.get('id', name) if attrs else name)
widget_id = self._normalize_id(widget_id_raw)

3
archivebox/core/wsgi.py
View File

@@ -9,9 +9,8 @@ https://docs.djangoproject.com/en/2.1/howto/deployment/wsgi/
import archivebox # noqa
from archivebox.config.django import setup_django
from django.core.wsgi import get_wsgi_application
setup_django(in_memory_db=False, check_db=True)
from django.core.wsgi import get_wsgi_application
application = get_wsgi_application()