wip pre-0.7.1 side fixes

.vscode/settings.json

@@ -0,0 +1,3 @@
+{
+    "python.formatting.provider": "black"
+}

archivebox/config.py

@@ -786,6 +786,7 @@ def find_chrome_binary() -> Optional[str]:
     # Precedence: Chromium, Chrome, Beta, Canary, Unstable, Dev
     # make sure data dir finding precedence order always matches binary finding order
     default_executable_paths = (
+        # '~/Library/Caches/ms-playwright/chromium-*/chrome-mac/Chromium.app/Contents/MacOS/Chromium',
         'chromium-browser',
         'chromium',
         '/Applications/Chromium.app/Contents/MacOS/Chromium',

archivebox/core/admin.py

@@ -48,22 +48,23 @@ class TagInline(admin.TabularInline):
 from django.contrib.admin.helpers import ActionForm
 from django.contrib.admin.widgets import AutocompleteSelectMultiple
 
-class AutocompleteTags:
-    model = Tag
-    search_fields = ['name']
+# class AutocompleteTags:
+#     model = Tag
+#     search_fields = ['name']
+#     name = 'tags'
 
-class AutocompleteTagsAdminStub:
-    name = 'admin'
+# class AutocompleteTagsAdminStub:
+#     name = 'admin'
 
 
 class SnapshotActionForm(ActionForm):
     tags = forms.ModelMultipleChoiceField(
         queryset=Tag.objects.all(),
         required=False,
-        widget=AutocompleteSelectMultiple(
-            AutocompleteTags(),
-            AutocompleteTagsAdminStub(),
-        ),
+        # widget=AutocompleteSelectMultiple(
+        #     # AutocompleteTags(),
+        #     # AutocompleteTagsAdminStub(),
+        # ),
     )
 
     # TODO: allow selecting actions for specific extractors? is this useful?

archivebox/core/apps.py

@@ -3,4 +3,4 @@ from django.apps import AppConfig
 
 class CoreConfig(AppConfig):
     name = 'core'
-    default_auto_field = 'django.db.models.UUIDField'
+    # default_auto_field = 'django.db.models.UUIDField'

archivebox/core/settings.py

@@ -268,6 +268,7 @@ AUTH_PASSWORD_VALIDATORS = [
     {'NAME': 'django.contrib.auth.password_validation.NumericPasswordValidator'},
 ]
 
+# DEFAULT_AUTO_FIELD = 'django.db.models.UUIDField'
 
 ################################################################################
 ### Shell Settings

archivebox/logging_util.py

@@ -393,7 +393,11 @@ def log_link_archiving_finished(link: "Link", link_dir: str, is_new: bool, stats
     else:
         _LAST_RUN_STATS.succeeded += 1
 
-    size = get_dir_size(link_dir)
+    try:
+        size = get_dir_size(link_dir)
+    except FileNotFoundError:
+        size = (0, None, '0')
+
     end_ts = datetime.now(timezone.utc)
     duration = str(end_ts - start_ts).split('.')[0]
     print('        {black}{} files ({}) in {}s {reset}'.format(size[2], printable_filesize(size[0]), duration, **ANSI))

archivebox/static

@@ -0,0 +1 @@
+templates/static

archivebox/templates/admin/private_index.html

@@ -88,4 +88,5 @@
       {% endblock %}
     </div>
   </div>
+  <script>alert(1)</script>
 {% endblock %}

docker-compose.yml

@@ -21,7 +21,6 @@ services:
             # - ./etc/crontabs:/var/spool/cron/crontabs  # uncomment this and archivebox_scheduler below to set up automatic recurring archive jobs
             # - ./archivebox:/app/archivebox             # uncomment this to mount the ArchiveBox source code at runtime (for developers working on archivebox)
         # build: .                                       # uncomment this to build the image from source code at buildtime (for developers working on archivebox)
-        
         environment:
             - ALLOWED_HOSTS=*                   # restrict this to only accept incoming traffic via specific domain name
             # - PUBLIC_INDEX=True               # set to False to prevent anonymous users from viewing snapshot list

etc/crontabs/archivebox

@@ -0,0 +1,8 @@
+# DO NOT EDIT THIS FILE - edit the master and reinstall.
+# (/tmp/tmpe3dawo9u installed on Tue Jun 13 23:21:48 2023)
+# (Cron version -- $Id: crontab.c,v 2.13 1994/01/17 03:20:37 vixie Exp $)
+
+@daily cd /data && /usr/local/bin/archivebox add --depth=0 "https://example.com/3" >> /data/logs/schedule.log 2>&1 # archivebox_schedule
+@daily cd /data && /usr/local/bin/archivebox add --depth=0 "https://example.com/2" >> /data/logs/schedule.log 2>&1 # archivebox_schedule
+@daily cd /data && /usr/local/bin/archivebox add --depth=0 "https://example.com" >> /data/logs/schedule.log 2>&1 # archivebox_schedule
+@daily cd /data && /usr/local/bin/archivebox add --depth=0 "update" >> /data/logs/schedule.log 2>&1 # archivebox_schedule