make get_commit_hash more precisce by using exact ref (#1270)

archivebox/config.py

@@ -53,6 +53,7 @@ from .config_stubs import (
 
 ### Pre-Fetch Minimal System Config
 
+TIMEZONE = 'UTC'
 SYSTEM_USER = getpass.getuser() or os.getlogin()
 
 try:
@@ -81,7 +82,6 @@ CONFIG_SCHEMA: Dict[str, ConfigDefaultDict] = {
         'IN_QEMU':                  {'type': bool,  'default': False},
         'PUID':                     {'type': int,   'default': os.getuid()},
         'PGID':                     {'type': int,   'default': os.getgid()},
-        # TODO: 'SHOW_HINTS':       {'type:  bool,  'default': True},   # hints are hidden automatically once collection contains >0 Snapshots, no need to configure
     },
 
     'GENERAL_CONFIG': {
@@ -392,9 +392,19 @@ def get_version(config):
     raise Exception('Failed to detect installed archivebox version!')
 
 def get_commit_hash(config) -> Optional[str]:
+    try:
+        git_dir = config['PACKAGE_DIR'] / '../'
+        ref = (git_dir / 'HEAD').read_text().strip().split(' ')[-1]
+        commit_hash = git_dir.joinpath(ref).read_text().strip()
+        return commit_hash
+    except Exception:
+        pass
+
     try:
         return list((config['PACKAGE_DIR'] / '../.git/refs/heads/').glob('*'))[0].read_text().strip()
     except Exception:
+        pass
+    
     return None
 
 def get_build_time(config) -> str:
@@ -792,6 +802,7 @@ 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',
@@ -1212,7 +1223,7 @@ def check_dependencies(config: ConfigDict=CONFIG, show_help: bool=True) -> None:
 
     if config['USE_YOUTUBEDL'] and config['MEDIA_TIMEOUT'] < 20:
         stderr(f'[!] Warning: MEDIA_TIMEOUT is set too low! (currently set to MEDIA_TIMEOUT={config["MEDIA_TIMEOUT"]} seconds)', color='red')
-        stderr('    Youtube-dl will fail to archive all media if set to less than ~20 seconds.')
+        stderr('    youtube-dl/yt-dlp will fail to archive any media if set to less than ~20 seconds.')
         stderr('    (Setting it somewhere over 60 seconds is recommended)')
         stderr()
         stderr('    If you want to disable media archiving entirely, set SAVE_MEDIA=False instead:')

archivebox/core/admin.py

@@ -48,22 +48,25 @@ 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']
+# WIP: commented out because broken by Django 3.1.2 -> 4.0 migration
+# 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(),
-        ),
+        # WIP: commented out because broken by Django 3.1.2 -> 4.0 migration
+        # 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 +268,8 @@ AUTH_PASSWORD_VALIDATORS = [
     {'NAME': 'django.contrib.auth.password_validation.NumericPasswordValidator'},
 ]
 
+# WIP: commented out because broken by Django 3.1.2 -> 4.0 migration
+# DEFAULT_AUTO_FIELD = 'django.db.models.UUIDField'
 
 ################################################################################
 ### Shell Settings

archivebox/extractors/__init__.py

@@ -184,7 +184,7 @@ def archive_link(link: Link, overwrite: bool=False, methods: Optional[Iterable[s
                         link.url,
                         command,
                         ts
-                    ) + "\n"))
+                    ) + "\n" + str(e) + "\n"))
                     #f.write(f"\n> {command}; ts={ts} version={config['VERSION']} docker={config['IN_DOCKER']} is_tty={config['IS_TTY']}\n")
 
         # print('    ', stats)

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
 
+    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

bin/build_docker.sh

@@ -51,6 +51,7 @@ function create_builder() {
     docker buildx use xbuilder && return 0
     echo "[+] Creating new xbuilder for: $SELECTED_PLATFORMS"
     echo
+    docker pull 'moby/buildkit:buildx-stable-1'
 
     # Switch to buildx builder if already present / previously created
     docker buildx create --name xbuilder --driver docker-container --bootstrap --use --platform "$SELECTED_PLATFORMS" || true
@@ -75,6 +76,7 @@ echo "[+] Generating requirements.txt and pdm.lock from pyproject.toml..."
 pdm lock --group=':all' --strategy="cross_platform" --production
 pdm export --group=':all' --production --without-hashes -o requirements.txt
 
+
 echo "[+] Building archivebox:$VERSION docker image..."
 # docker builder prune
 # docker build . --no-cache -t archivebox-dev \

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