Previously, `archivebox search --json` exported tags as a comma-separated
string (e.g. "tag1,tag2"), which required manual parsing by consumers like
LlamaIndex, LangChain, and other RAG frameworks.
Now `to_dict()` returns tags as a proper JSON array (e.g. ["tag1", "tag2"]),
making the export directly usable as structured metadata in LLM/RAG pipelines
without additional preprocessing.
`from_json()` is updated to accept both list and string formats for backward
compatibility with existing JSON imports.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add prominent view mode switcher with List/Grid toggle buttons
- Improve filter sidebar CSS with modern styling, rounded corners
- Add live progress bar for in-progress snapshots showing hooks status
- Show plugin icons only when output directory has content
- Display archive result output_size sum from new field
- Show hooks succeeded/total count in size column
- Add get_progress_stats() method to Snapshot model
- Add CSS for progress spinner and status badges
- Update grid view template with progress indicator for archiving cards
- Add tests for admin views, search, and progress stats
DELETED:
- workers/pid_utils.py (-192 lines) - replaced by Process model methods
SIMPLIFIED:
- crawls/models.py Crawl.cleanup() (80 lines -> 10 lines)
- hooks.py: deleted process_is_alive() and kill_process() (-45 lines)
UPDATED to use Process model:
- core/models.py: Snapshot.cleanup() and has_running_background_hooks()
- machine/models.py: Binary.cleanup()
- workers/worker.py: Worker.on_startup/shutdown, get_running_workers, start
- workers/orchestrator.py: Orchestrator.on_startup/shutdown, is_running
All subprocess management now uses:
- Process.current() for registering current process
- Process.get_running() / get_running_count() for querying
- Process.cleanup_stale_running() for cleanup
- safe_kill_process() for validated PID killing
Total line reduction: ~250 lines
Phase 1: Model Prerequisites
- Add ArchiveResult.from_json() and from_jsonl() methods
- Fix Snapshot.to_json() to use tags_str (consistent with Crawl)
Phase 2: Shared Utilities
- Create archivebox/cli/cli_utils.py with shared apply_filters()
- Update 7 CLI files to import from cli_utils.py instead of duplicating
Phase 3: Pass-Through Behavior
- Add pass-through to crawl create (non-Crawl records pass unchanged)
- Add pass-through to snapshot create (Crawl records + others pass through)
- Add pass-through to archiveresult create (Snapshot records + others)
- Add create-or-update behavior to run command:
- Records WITHOUT id: Create via Model.from_json()
- Records WITH id: Lookup existing, re-queue
- Outputs JSONL of all processed records for chaining
Phase 4: Test Infrastructure
- Create archivebox/tests/conftest.py with pytest-django fixtures
- Include CLI helpers, output assertions, database assertions
Phase 6: Config Update
- Update supervisord_util.py: orchestrator -> run command
This enables Unix-style piping:
archivebox crawl create URL | archivebox run
archivebox archiveresult list --status=failed | archivebox run
curl API | jq transform | archivebox crawl create | archivebox run
Extended graceful_terminate_background_hooks() to:
- Reap processes with os.waitpid() to get exit codes
- Write returncode to .returncode file for update_from_output()
- Return detailed result dict with status, returncode, and pid
Updated update_from_output() to:
- Read .returncode and .stderr.log files
- Determine status from returncode if no ArchiveResult JSONL record
- Include stderr in output_str for failed hooks
- Handle signal termination (negative returncodes like -9 for SIGKILL)
- Clean up .returncode files along with other hook output files
Changed Snapshot.cleanup() to gracefully terminate background hooks:
1. Send SIGTERM to all background hook processes first
2. Wait up to each hook's plugin-specific timeout
3. Send SIGKILL only to hooks still running after their timeout
Added graceful_terminate_background_hooks() function in hooks.py that:
- Collects all .pid files from output directory
- Validates process identity using mtime
- Sends SIGTERM to all valid processes in phase 1
- Polls each process for up to its plugin-specific timeout
- Sends SIGKILL as last resort if timeout expires
- Returns status for each hook (sigterm/sigkill/already_dead/invalid)
Multiple hooks in the same plugin directory were overwriting each
other's stdout.log, stderr.log, hook.pid, and cmd.sh files. Now
each hook uses filenames prefixed with its hook name:
- on_Snapshot__20_chrome_tab.bg.stdout.log
- on_Snapshot__20_chrome_tab.bg.stderr.log
- on_Snapshot__20_chrome_tab.bg.pid
- on_Snapshot__20_chrome_tab.bg.sh
Updated:
- hooks.py run_hook() to use hook-specific names
- core/models.py cleanup and update_from_output methods
- Plugin scripts to no longer write redundant hook.pid files
- Add JSONL_INDEX_FILENAME to ALLOWED_IN_DATA_DIR for consistency
- Fix fallback logic in legacy.py to try JSON when JSONL parsing fails
- Replace bare except clauses with specific exception types
- Fix stdin double-consumption in archivebox_crawl.py
- Merge CLI --tag option with crawl tags in archivebox_snapshot.py
- Remove tautological mock tests (covered by integration tests)
Co-authored-by: Nick Sweeting <pirate@users.noreply.github.com>
- Add Tag.to_jsonl() method with schema_version
- Add Crawl.to_jsonl() method with schema_version
- Fix Tag.from_jsonl() to not depend on jsonl.py helper
- Update tests to use Snapshot.from_jsonl() instead of non-existent get_or_create_snapshot
Remove model-specific functions from misc/jsonl.py:
- tag_to_jsonl() - use Tag.to_jsonl() instead
- crawl_to_jsonl() - use Crawl.to_jsonl() instead
- get_or_create_tag() - use Tag.from_jsonl() instead
- process_jsonl_records() - use model from_jsonl() methods directly
jsonl.py now only contains generic I/O utilities:
- Type constants (TYPE_SNAPSHOT, etc.)
- parse_line(), read_stdin(), read_file(), read_args_or_stdin()
- write_record(), write_records()
- filter_by_type(), process_records()
Move JSONL serialization from standalone functions to model methods
to mirror the from_jsonl() pattern:
- Add Binary.to_jsonl() method
- Add Process.to_jsonl() method
- Add ArchiveResult.to_jsonl() method
- Add Snapshot.to_jsonl() method
- Update write_index_jsonl() to use model methods
- Update jsonl.py functions to be thin wrappers
Switch from hierarchical index.json to flat index.jsonl format for
snapshot metadata storage. Each line is a self-contained JSON record
with a 'type' field (Snapshot, ArchiveResult, Binary, Process).
Changes:
- Add JSONL_INDEX_FILENAME constant to constants.py
- Add TYPE_PROCESS and TYPE_MACHINE to jsonl.py type constants
- Add binary_to_jsonl(), process_to_jsonl(), machine_to_jsonl() converters
- Add Snapshot.write_index_jsonl() to write new format
- Add Snapshot.read_index_jsonl() to read new format
- Add Snapshot.convert_index_json_to_jsonl() for migration
- Update Snapshot.reconcile_with_index() to handle both formats
- Update fs_migrate to convert during filesystem migration
- Update load_from_directory/create_from_directory for both formats
- Update legacy.py parse_json_links_details for JSONL support
The new format is easier to parse, extend, and mix record types.
