soulsync/core/automation/handlers/personalized_pipeline.py
Broque Thomas 877d0e7d81 Personalized pipeline: auto-refresh stale snapshots after watchlist scan
Snapshots now track when their source data changes. Watchlist scan
emits stale flags on the playlists whose underlying pool just got
refreshed; the next pipeline run sees the flag and regenerates the
snapshot before syncing, so the server playlist never lags the source.

Schema:
- new `is_stale INTEGER NOT NULL DEFAULT 0` column on
  `personalized_playlists`, plus an idempotent ADD COLUMN migration
  in `ensure_personalized_schema` for installs created before this PR.
- `PlaylistRecord.is_stale: bool = False` exposed on the dataclass so
  callers can branch on freshness without re-querying.

Manager:
- new `mark_kinds_stale(kinds, profile_id=None)` flips the flag in
  bulk for a list of kinds (used by upstream data refreshers).
- `_persist_snapshot` clears `is_stale = 0` on successful refresh.
- SELECT statements + `_row_to_record` updated to read the column
  (with tuple-form length guard for safety).

Pipeline:
- `_build_payloads_for_kinds` now branches: refresh_first=True OR
  `existing.is_stale` -> refresh_playlist, else read existing
  snapshot. So the auto-refresh kicks in without needing the user to
  toggle the refresh-each-run option.

Watchlist scanner emits stale flags at three sites:
- after `update_discovery_pool_timestamp` -> marks pool-fed kinds
  stale: hidden_gems, discovery_shuffle, popular_picks, time_machine,
  genre_playlist, daily_mix.
- after release_radar `save_curated_playlist` -> marks `fresh_tape`.
- after discovery_weekly `save_curated_playlist` -> marks `archives`.

All three calls go through a module-level `_mark_personalized_kinds_stale`
helper that builds a PersonalizedPlaylistManager with `deps=None` (only
DB access is needed for the flag update — no generator dispatch). Each
call is wrapped in try/except so a flag failure can never abort the
scan itself.

Tests:
- new `TestStaleFlag` class in `test_personalized_manager.py` (6
  tests): default-false, single-kind flip, multi-kind, profile
  scoping, refresh-clears, empty-list noop.
- two new pipeline tests pin the auto-refresh dispatch:
  `test_stale_snapshot_auto_refreshes_even_without_refresh_first`
  and `test_non_stale_snapshot_skips_refresh`.
- existing stub-manager `SimpleNamespace` returns gained
  `is_stale=False` so the new attribute read doesn't AttributeError.

Full suite: 3391 pass.

User-facing WHATS_NEW entry added under 2.5.2 (above the prior
pipeline auto-sync entry) describing the auto-refresh behavior.
2026-05-15 20:53:03 -07:00

278 lines
10 KiB
Python

"""Personalized Playlist Pipeline automation handler.
Sibling to ``auto_playlist_pipeline`` (mirrored). Where the mirrored
pipeline runs REFRESH external sources → DISCOVER metadata → SYNC →
WISHLIST, the personalized pipeline is simpler:
SNAPSHOT → SYNC → WISHLIST
SNAPSHOT reads the persisted track list from
``PersonalizedPlaylistManager``. When ``refresh_first=True`` (config),
each playlist is refreshed BEFORE syncing — useful when the user
wants the cron to capture a fresh-each-run view (e.g. "give me a new
Hidden Gems set every night"). Default is to sync the existing
snapshot, on the assumption the user / a separate cron has already
refreshed when they wanted to.
Config schema:
{
'kinds': [
{'kind': 'hidden_gems'},
{'kind': 'time_machine', 'variant': '1980s'},
{'kind': 'seasonal_mix', 'variant': 'halloween'},
...
],
'refresh_first': bool, # default false
'skip_wishlist': bool, # default false
}
Each kind dict has at minimum ``kind``; ``variant`` is required for
kinds that need it (time_machine, genre_playlist, daily_mix,
seasonal_mix). Singleton kinds (hidden_gems, discovery_shuffle,
popular_picks, fresh_tape, archives) ignore variant.
Pipeline-running flag (``deps.state.pipeline_running``) is shared
with the mirrored pipeline so the two can't overlap. (One sync
queue, one wishlist worker — overlapping triggers would step on
each other.)"""
from __future__ import annotations
import threading
import time
from typing import Any, Dict, List, Optional
from core.automation.deps import AutomationDeps
from core.automation.handlers._pipeline_shared import run_sync_and_wishlist
# Sync state key prefix so personalized syncs don't collide with
# mirrored ones (`auto_mirror_<id>`).
_SYNC_ID_PREFIX = 'auto_personalized'
def auto_personalized_pipeline(config: Dict[str, Any], deps: AutomationDeps) -> Dict[str, Any]:
"""Run SNAPSHOT → SYNC → WISHLIST for selected personalized playlists."""
deps.state.set_pipeline_running(True)
automation_id = config.get('_automation_id')
pipeline_start = time.time()
try:
kinds_config = config.get('kinds') or []
if not isinstance(kinds_config, list) or not kinds_config:
deps.state.set_pipeline_running(False)
return {
'status': 'error',
'error': 'No personalized playlist kinds selected',
}
refresh_first = bool(config.get('refresh_first', False))
skip_wishlist = bool(config.get('skip_wishlist', False))
manager = deps.build_personalized_manager()
deps.update_progress(
automation_id,
progress=2,
phase=f'Personalized pipeline: {len(kinds_config)} playlist(s)',
log_line=f'Starting pipeline for {len(kinds_config)} playlist(s)',
log_type='info',
)
# ── PHASE 1: SNAPSHOT (optionally refresh) ──────────────────
deps.update_progress(
automation_id,
progress=3,
phase='Phase 1/2: Loading snapshots...' if not refresh_first
else 'Phase 1/2: Refreshing snapshots...',
log_line='Phase 1: Snapshot' + (' (with refresh)' if refresh_first else ''),
log_type='info',
)
profile_id = deps.get_current_profile_id()
playload_payloads = _build_payloads_for_kinds(
deps, manager, kinds_config, profile_id,
automation_id=automation_id,
refresh_first=refresh_first,
)
if not playload_payloads:
deps.state.set_pipeline_running(False)
deps.update_progress(
automation_id,
status='finished', progress=100,
phase='No playlists to sync',
log_line='No personalized playlists had tracks to sync',
log_type='warning',
)
return {
'status': 'completed',
'_manages_own_progress': True,
'playlists_synced': '0',
'tracks_synced': '0',
'duration_seconds': str(int(time.time() - pipeline_start)),
}
deps.update_progress(
automation_id,
progress=50,
phase='Phase 1/2: Snapshot complete',
log_line=f'Phase 1 done: {len(playload_payloads)} playlist(s) ready to sync',
log_type='success',
)
# ── PHASE 2: SYNC + WISHLIST (shared helper) ────────────────
sync_summary = run_sync_and_wishlist(
deps,
automation_id,
playload_payloads,
sync_one_fn=lambda pl: _sync_personalized_playlist(deps, pl),
sync_id_for_fn=lambda pl: pl['sync_id'],
skip_wishlist=skip_wishlist,
progress_start=51,
progress_end=90,
sync_phase_label='Phase 2/2: Syncing to server...',
sync_phase_start_log='Phase 2: Sync',
wishlist_phase_label='Phase 2/2: Processing wishlist...',
wishlist_phase_start_log='Wishlist',
)
# ── COMPLETE ────────────────────────────────────────────────
duration = int(time.time() - pipeline_start)
deps.update_progress(
automation_id,
status='finished', progress=100,
phase='Pipeline complete',
log_line=f'Personalized pipeline finished in {duration // 60}m {duration % 60}s',
log_type='success',
)
deps.state.set_pipeline_running(False)
return {
'status': 'completed',
'_manages_own_progress': True,
'playlists_synced': str(len(playload_payloads)),
'tracks_synced': str(sync_summary['synced']),
'sync_skipped': str(sync_summary['skipped']),
'wishlist_queued': str(sync_summary['wishlist_queued']),
'duration_seconds': str(duration),
}
except Exception as e: # noqa: BLE001 — automation handlers must never raise into engine
deps.state.set_pipeline_running(False)
deps.update_progress(
automation_id,
status='error', progress=100,
phase='Pipeline error',
log_line=f'Personalized pipeline failed: {e}',
log_type='error',
)
return {'status': 'error', 'error': str(e), '_manages_own_progress': True}
def _build_payloads_for_kinds(
deps: AutomationDeps,
manager: Any,
kinds_config: List[Dict[str, Any]],
profile_id: int,
*,
automation_id: Optional[str],
refresh_first: bool,
) -> List[Dict[str, Any]]:
"""Resolve each requested kind+variant into a sync-payload dict.
Each payload has: ``{'name', 'kind', 'variant', 'tracks_json',
'image_url', 'sync_id'}``. Playlists with no tracks (e.g. a
seasonal mix that hasn't been populated yet) are omitted from
the result so the sync loop doesn't waste time on empty pushes.
"""
payloads: List[Dict[str, Any]] = []
for entry in kinds_config:
if not isinstance(entry, dict):
continue
kind = entry.get('kind')
variant = entry.get('variant') or ''
if not kind:
continue
try:
# Determine whether to refresh: explicit user flag, OR the
# snapshot is marked stale because the underlying source
# data (discovery_pool, curated_playlists) changed since
# last generation. Either way → refresh; otherwise just
# read the existing snapshot.
if refresh_first:
record = manager.refresh_playlist(kind, variant, profile_id)
else:
existing = manager.ensure_playlist(kind, variant, profile_id)
if existing.is_stale:
record = manager.refresh_playlist(kind, variant, profile_id)
else:
record = existing
except Exception as exc: # noqa: BLE001 — log + continue with next kind
deps.update_progress(
automation_id,
log_line=f'Skipping {kind}{("/" + variant) if variant else ""}: {exc}',
log_type='warning',
)
continue
tracks = manager.get_playlist_tracks(record.id)
if not tracks:
deps.update_progress(
automation_id,
log_line=f'No tracks in {record.name} — skipping sync',
log_type='skip',
)
continue
tracks_json = [_track_to_sync_shape(t) for t in tracks]
payloads.append({
'name': record.name,
'kind': record.kind,
'variant': record.variant,
'tracks_json': tracks_json,
'image_url': '', # personalized playlists don't have a cover image yet
'sync_id': f'{_SYNC_ID_PREFIX}_{record.kind}_{record.variant or "_"}',
})
return payloads
def _track_to_sync_shape(track: Any) -> Dict[str, Any]:
"""Convert a personalized.types.Track into the dict shape
`_run_sync_task` expects. Mirrors what the mirrored pipeline
builds from extra_data.matched_data — name/artists/album/duration/id."""
primary_id = track.spotify_track_id or track.itunes_track_id or track.deezer_track_id or ''
artists = [{'name': track.artist_name}]
return {
'name': track.track_name,
'artists': artists,
'album': {'name': track.album_name or ''},
'duration_ms': int(track.duration_ms or 0),
'id': primary_id,
}
def _sync_personalized_playlist(deps: AutomationDeps, payload: Dict[str, Any]) -> Dict[str, Any]:
"""Launch a personalized playlist sync via _run_sync_task on a
daemon thread + return immediately with status='started'.
Mirrors the mirrored ``auto_sync_playlist`` return contract so the
shared helper can poll on ``sync_states[sync_id]`` and aggregate
results identically."""
sync_id = payload['sync_id']
name = payload['name']
tracks_json = payload['tracks_json']
profile_id = deps.get_current_profile_id()
threading.Thread(
target=deps.run_sync_task,
args=(sync_id, name, tracks_json, None, profile_id, payload.get('image_url', '')),
daemon=True,
name=f'auto-personalized-{sync_id}',
).start()
return {
'status': 'started',
'playlist_name': name,
'_manages_own_progress': True,
}