Files
feedBack/plugins/achievements/routes.py
d2569cc2a8 feat(achievements): wall sync drain worker + review fixes (epic PR3) (#592)
* feat(achievements): wall sync drain worker (epic PR3, client side)

Background dead-letter worker that POSTs queued Feat unlocks/removals to the
hosted feedback-achievements wall. Idle unless FEEDBACK_ACHIEVEMENTS_WALL_URL
is set; uses requests + the client-token header (mirrors lyrics_transcribe).

Dead-letter, never drop (pure engine.drain_decision):
  network err / 429 / 5xx -> keep pending (retry)
  other 4xx               -> dead_letter (diagnosable, replayable)
  2xx                     -> delete on server ack
remove-me enqueues a wall removal keyed by the reused player_hash.

Verified by an end-to-end staging round-trip (earn a Feat -> drains onto the
wall with name + short hash -> remove-me -> wall empties) with no IP in tables
or access logs. 42 plugin tests pass (test_sync.py adds the decision table +
ack/retry/dead-letter retention + four-field on-the-wire payload).

The hosted service lives in the new feedback-achievements repo.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(achievements): address local review findings (epic)

Bugs caught in the pre-merge review loop:

- secret_witching Feat was DEAD: post_activity wrote witching_nights_run to the
  DB before snapshotting prev_tiers, so diff_unlocks never saw the fresh unlock.
  Fold the run into the activity delta instead (same asymmetry chart_encore
  uses) so the 7th-night unlock is detected. +regression tests.
- chart_encore broke across restarts: per-chart counter keyed on abs(hash(str)),
  which Python salts per-process (PYTHONHASHSEED). Use a stable sha1 digest so
  the same chart accumulates across sessions. +regression test.
- Bounded the per-activity counter read: _read_counters no longer pulls the
  unbounded chart_plays:* rows (they're bumped/read individually).
- screen.js: gate note:hit/miss on an active-song flag so tuner/calibration note
  events can't inflate Feats or flush a phantom chart:null session.
- screen.js: P-III — prefix the plugin localStorage key (achievements:profile-cat).
- screen.js: extract the duplicated local-ISO-date helper.

45 plugin tests pass (3 new). Wall-side review fixes are in the
feedback-achievements repo.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(achievements): default the drain worker to the hosted wall

Point FEEDBACK_ACHIEVEMENTS_WALL_URL's default at the live got-feedback wall
(https://feedback-achievements.onrender.com) so the drain worker targets it out
of the box; still env-overridable for self-hosting/staging. Nothing publishes
unless the user opted in AND has a profile identity, so a default URL alone
sends nothing.

Tests disable the default (autouse fixture) so no test ever POSTs to production;
drain logic is covered via _drain_once() with an injected poster. 45 pass.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-24 17:01:48 +02:00

586 lines
23 KiB
Python

"""Achievements & Feats of Power — local engine (offline).
State lives under ``<config_dir>/achievements/``:
- ``achievements.db`` SQLite — unlocks, activity counters, derived ledger,
and the (PR3) wall sync queue.
Two surfaces, one engine, kept structurally apart (the **integration law**):
* **Feats of Power** — activity/volume. The engine OWNS raw activity counters
(`counters`), evaluates Feat thresholds, and records Feat unlocks. Feats are
the only thing that ever syncs to the public wall.
* **Achievements** — demonstrated competency. The engine RECORDS unlocks the
source reports (`report-unlock`); it never re-derives them from activity.
A baseline catalogue ships here and is driven by the built-in progression
system; richer items are contributed by source plugins at runtime.
Endpoints (all under /api/plugins/achievements/):
POST /activity bump activity counters, eval Feats, return newly-unlocked
POST /report-unlock idempotent upsert of a competency/feat unlock
POST /report-criterion record a (criterion_id, token) pair → distinct count
GET /catalog baseline competency defs + earned state
GET /earned all earned items (id, cls, category, tier, at)
GET /feats earned Feats (for the profile trophy shelf)
POST /remove-me wipe synced state (full wall-removal lands in PR2/PR3)
Pure threshold/criterion math lives in the sibling ``engine.py`` (P-V testable);
this module is the SQLite + HTTP shell.
"""
import hashlib
import json
import logging
import os
import sqlite3
import threading
import time
from pathlib import Path
from pydantic import BaseModel, Field
_lock = threading.Lock()
_state = {
"db_path": None,
"dir": None, # plugin directory (for catalog JSON)
"config_dir": None, # CONFIG_DIR (for reading the opt-in setting)
"meta_db": None, # MetadataDB (for the profile identity: name + hash)
"log": logging.getLogger("feedBack.plugin.achievements"),
"engine": None, # sibling engine.py module (pure helpers)
"feat_defs": [], # parsed feats.json -> list of feat defs
"baseline": {}, # parsed achievements.json
}
# ── SQLite ──────────────────────────────────────────────────────────────────
def _conn():
conn = sqlite3.connect(_state["db_path"], timeout=5)
conn.row_factory = sqlite3.Row
conn.execute("PRAGMA journal_mode=WAL")
return conn
def _init_db():
conn = _conn()
try:
conn.execute(
"""
CREATE TABLE IF NOT EXISTS unlocks (
achievement_id TEXT PRIMARY KEY,
cls TEXT NOT NULL, -- 'competency' | 'feat'
disp_category TEXT, -- global/guitar/bass/...
source_id TEXT,
tier INTEGER NOT NULL DEFAULT 0,
unlocked_at TEXT,
synced INTEGER NOT NULL DEFAULT 0
)
"""
)
conn.execute(
"""
CREATE TABLE IF NOT EXISTS counters (
key TEXT PRIMARY KEY,
value INTEGER NOT NULL DEFAULT 0
)
"""
)
conn.execute(
"""
CREATE TABLE IF NOT EXISTS comp_ledger (
criterion_id TEXT NOT NULL,
token TEXT NOT NULL,
PRIMARY KEY (criterion_id, token)
)
"""
)
conn.execute(
"""
CREATE TABLE IF NOT EXISTS sync_queue (
id INTEGER PRIMARY KEY AUTOINCREMENT,
kind TEXT NOT NULL, -- 'unlock' | 'remove'
payload TEXT NOT NULL,
state TEXT NOT NULL DEFAULT 'pending' -- 'pending' | 'dead_letter'
)
"""
)
conn.commit()
finally:
conn.close()
def _now_iso():
return time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime())
def _opted_in():
"""True only when the user has opted in (core setting ``achievements_enabled``).
Read straight from CONFIG_DIR/config.json — the single source of truth the
/api/settings endpoint persists. Default OFF on any read failure: nothing
leaves the device unless explicitly enabled.
"""
try:
cfg_path = Path(_state["config_dir"]) / "config.json"
cfg = json.loads(cfg_path.read_text(encoding="utf-8"))
return bool(cfg.get("achievements_enabled") is True)
except (OSError, ValueError, TypeError):
return False
def _identity():
"""(display_name, player_hash) from the profile, or (None, None).
Reused as the wall identity (server.py's documented player_hash). Sync is
skipped entirely when either is missing.
"""
db = _state["meta_db"]
if db is None or not hasattr(db, "get_profile"):
return None, None
try:
prof = db.get_profile() or {}
return (prof.get("display_name") or None), (prof.get("player_hash") or None)
except Exception: # noqa: BLE001 — identity is best-effort; never break a request
return None, None
def _enqueue_feat_sync(conn, feat_id, unlocked_at):
"""Enqueue a wall-sync POST for a Feat unlock — opt-in gated, identity gated.
Builds the outbound payload through the SINGLE code-gated serializer
(engine.build_wall_payload, exactly four fields). Competency unlocks never
reach this path (integration law + data-minimization contract). The drain
worker (PR3) POSTs the queued rows; here we only persist intent.
"""
if not _opted_in():
return False
display_name, player_hash = _identity()
if not display_name or not player_hash:
return False
payload = _state["engine"].build_wall_payload(display_name, player_hash, feat_id, unlocked_at)
conn.execute(
"INSERT INTO sync_queue(kind, payload, state) VALUES ('unlock', ?, 'pending')",
(json.dumps(payload),),
)
return True
# ── Wall sync — background drain worker (dead-letter, never drop) ─────────────
# Idle unless a wall URL is configured. POSTs pending rows to the hosted
# feedback-achievements service; the decision state machine
# (engine.drain_decision) is pure + tested. A row leaves the queue only on a
# server ack (or a user opt-out wiping it) — never silently dropped.
# Canonical hosted Feats wall (the got-feedback service). Used by default so the
# drain worker targets it out of the box; override via env for self-hosting or a
# staging wall. Nothing is ever sent unless the user opted in AND has an identity
# (see _enqueue_feat_sync), so a default URL does not publish anything on its own.
_DEFAULT_WALL_URL = "https://feedback-achievements.onrender.com"
_WALL_URL = (os.environ.get("FEEDBACK_ACHIEVEMENTS_WALL_URL")
or os.environ.get("SLOPSMITH_ACHIEVEMENTS_WALL_URL")
or _DEFAULT_WALL_URL).rstrip("/")
_WALL_TOKEN = os.environ.get("FEEDBACK_ACHIEVEMENTS_CLIENT_TOKEN", "fb-wall-v1")
_DRAIN_INTERVAL_S = int(os.environ.get("FEEDBACK_ACHIEVEMENTS_DRAIN_INTERVAL", "30"))
_drain_started = False
def _post_to_wall(kind, payload):
"""POST one queued item; return the HTTP status code, or None on a network
error. Mirrors the lib/lyrics_transcribe outbound pattern (explicit timeout,
no raise on non-2xx — the caller's state machine decides)."""
import requests # local import: only needed when a wall is configured
path = "/api/unlock" if kind == "unlock" else "/api/remove"
try:
resp = requests.post(
_WALL_URL + path, json=payload,
headers={"X-Client-Token": _WALL_TOKEN, "Content-Type": "application/json"},
timeout=10,
)
return resp.status_code
except requests.RequestException:
return None
def _drain_once(post_fn=None):
"""Process all pending queue rows once. ``post_fn(kind, payload) -> status``
is injectable for tests; defaults to the real wall POST."""
post_fn = post_fn or _post_to_wall
engine = _state["engine"]
with _lock:
conn = _conn()
try:
rows = conn.execute(
"SELECT id, kind, payload FROM sync_queue WHERE state='pending'").fetchall()
finally:
conn.close()
for row in rows:
try:
payload = json.loads(row["payload"]) if row["payload"] else {}
except ValueError:
payload = {}
status = post_fn(row["kind"], payload)
action = engine.drain_decision(status)
with _lock:
conn = _conn()
try:
if action == "ack":
conn.execute("DELETE FROM sync_queue WHERE id=?", (row["id"],))
elif action == "dead":
conn.execute("UPDATE sync_queue SET state='dead_letter' WHERE id=?", (row["id"],))
# 'retry' → leave it pending for the next pass
conn.commit()
finally:
conn.close()
def _drain_loop():
while True:
try:
_drain_once()
except Exception as e: # noqa: BLE001 — a worker crash must not kill the thread
_state["log"].warning("achievements wall drain error: %s", e)
time.sleep(_DRAIN_INTERVAL_S)
def _maybe_start_drain():
global _drain_started
if _drain_started or not _WALL_URL:
return
_drain_started = True
threading.Thread(target=_drain_loop, name="ach-wall-drain", daemon=True).start()
_state["log"].info("achievements wall drain worker started → %s", _WALL_URL)
def _chart_key(chart):
"""Stable per-chart counter key — a sha1 digest of the chart id. NOT the
builtin hash(), whose str hashing is salted per process (PYTHONHASHSEED), so
the same chart would land on a different counter after every restart and the
Encore Feat could never accumulate across sessions."""
return "chart_plays:" + hashlib.sha1(str(chart).encode("utf-8")).hexdigest()[:16]
def _read_counters(conn):
# Excludes the per-chart `chart_plays:*` rows: they are bumped + read
# individually via _bump_counter and would otherwise make this aggregate
# round-trip O(distinct charts played) on every activity POST.
return {
row["key"]: int(row["value"])
for row in conn.execute("SELECT key, value FROM counters WHERE key NOT LIKE 'chart_plays:%'")
}
def _write_counters(conn, counters):
for key, value in counters.items():
conn.execute(
"INSERT INTO counters(key, value) VALUES (?, ?) "
"ON CONFLICT(key) DO UPDATE SET value=excluded.value",
(key, int(value)),
)
def _bump_counter(conn, key, delta):
"""Increment a counter and return the new value (used for per-chart plays)."""
conn.execute(
"INSERT INTO counters(key, value) VALUES (?, ?) "
"ON CONFLICT(key) DO UPDATE SET value=value+excluded.value",
(key, int(delta)),
)
row = conn.execute("SELECT value FROM counters WHERE key=?", (key,)).fetchone()
return int(row["value"]) if row else int(delta)
def _earned_feat_tiers(conn):
return {
row["achievement_id"]: int(row["tier"])
for row in conn.execute("SELECT achievement_id, tier FROM unlocks WHERE cls='feat'")
}
def _record_unlock(conn, ach_id, cls, disp_category, source_id, tier, at):
"""Idempotent upsert; only advances the tier upward. Returns True if changed."""
row = conn.execute("SELECT tier FROM unlocks WHERE achievement_id=?", (ach_id,)).fetchone()
if row is not None and int(row["tier"]) >= int(tier):
return False
conn.execute(
"""
INSERT INTO unlocks(achievement_id, cls, disp_category, source_id, tier, unlocked_at, synced)
VALUES (?, ?, ?, ?, ?, ?, 0)
ON CONFLICT(achievement_id) DO UPDATE SET
tier=excluded.tier,
cls=excluded.cls,
disp_category=COALESCE(excluded.disp_category, unlocks.disp_category),
source_id=COALESCE(excluded.source_id, unlocks.source_id)
""",
(ach_id, cls, disp_category, source_id, int(tier), at or _now_iso()),
)
return True
# ── Catalog loading ─────────────────────────────────────────────────────────
def _feat_by_id(fid):
for f in _state["feat_defs"]:
if f.get("id") == fid:
return f
return None
def _load_catalogs():
base = Path(_state["dir"])
try:
feats = json.loads((base / "feats.json").read_text(encoding="utf-8"))
_state["feat_defs"] = feats.get("feats", []) if isinstance(feats, dict) else []
except (OSError, ValueError) as e:
_state["log"].warning("achievements: could not load feats.json: %s", e)
_state["feat_defs"] = []
try:
_state["baseline"] = json.loads((base / "achievements.json").read_text(encoding="utf-8"))
except (OSError, ValueError) as e:
_state["log"].warning("achievements: could not load achievements.json: %s", e)
_state["baseline"] = {}
# ── Request models ──────────────────────────────────────────────────────────
class ActivityIn(BaseModel):
notes: int = Field(ge=0, default=0)
session_notes: int = Field(ge=0, default=0)
in_song_streak: int = Field(ge=0, default=0)
song_done: int = Field(ge=0, default=0)
seconds: int = Field(ge=0, default=0)
chart: str | None = None
night_session: bool = False
night_date: str | None = None # 'YYYY-MM-DD', frontend supplies (no server clock)
class UnlockIn(BaseModel):
id: str
kind: str = "achievement" # 'achievement' | 'feat'
category: str | None = None # display category (global/guitar/...)
sourceId: str | None = None
tier: int = Field(ge=0, default=0)
at: str | None = None
class CriterionIn(BaseModel):
criterion_id: str
token: str
# ── FastAPI wiring ──────────────────────────────────────────────────────────
def setup(app, context):
config_dir = context["config_dir"]
base = Path(config_dir) / "achievements"
base.mkdir(parents=True, exist_ok=True)
_state["db_path"] = str(base / "achievements.db")
_state["dir"] = str(Path(__file__).resolve().parent)
_state["config_dir"] = str(config_dir)
_state["meta_db"] = context.get("meta_db")
_state["log"] = context.get("log") or _state["log"]
# Pure helpers via the per-plugin sibling loader (constitution P-III), with a
# plain-import fallback for pytest / standalone use.
load_sibling = context.get("load_sibling")
try:
_state["engine"] = load_sibling("engine") if load_sibling else __import__("engine")
except Exception: # noqa: BLE001 — last-ditch, keep the plugin alive
import importlib.util
spec = importlib.util.spec_from_file_location(
"achievements_engine", str(Path(__file__).resolve().parent / "engine.py"))
mod = importlib.util.module_from_spec(spec)
spec.loader.exec_module(mod)
_state["engine"] = mod
_init_db()
_load_catalogs()
log = _state["log"]
@app.post("/api/plugins/achievements/activity")
def post_activity(body: ActivityIn):
engine = _state["engine"]
with _lock:
conn = _conn()
try:
# Per-chart play count is the only directly-stateful bit; bump it
# first (stable key) so apply_activity() just takes the new max.
chart_play_count = None
if body.song_done and body.chart:
chart_play_count = _bump_counter(conn, _chart_key(body.chart), 1)
# Night-window ledger → consecutive-night run. Computed here but
# NOT written before the prev snapshot: it is folded into the delta
# below so prev_tiers reflects the OLD run and new_tiers the new one
# (the same asymmetry chart_encore relies on). Pre-writing it would
# make prev already satisfy the Feat, so diff_unlocks would never
# see the freshly-earned witching unlock.
witching_run = None
if body.night_session and body.night_date:
conn.execute(
"INSERT OR IGNORE INTO comp_ledger(criterion_id, token) VALUES ('witching', ?)",
(body.night_date,),
)
nights = [r["token"] for r in conn.execute(
"SELECT token FROM comp_ledger WHERE criterion_id='witching'")]
witching_run = engine.consecutive_run_length(nights)
counters = _read_counters(conn)
prev_tiers = engine.evaluate_feats(_state["feat_defs"], counters)
new_counters = engine.apply_activity(counters, {
"notes": body.notes,
"session_notes": body.session_notes,
"in_song_streak": body.in_song_streak,
"song_done": body.song_done,
"seconds": body.seconds,
"chart_play_count": chart_play_count,
})
if witching_run is not None:
new_counters["witching_nights_run"] = max(
int(new_counters.get("witching_nights_run", 0) or 0), witching_run)
_write_counters(conn, new_counters)
new_tiers = engine.evaluate_feats(_state["feat_defs"], new_counters)
fresh = engine.diff_unlocks(prev_tiers, new_tiers)
unlocked = []
for fid in fresh:
f = _feat_by_id(fid) or {}
tier = new_tiers[fid]
at = _now_iso()
if _record_unlock(conn, fid, "feat", f.get("category"), f.get("sourceId"), tier, at):
_enqueue_feat_sync(conn, fid, at)
unlocked.append(_feat_payload(fid, f, tier))
conn.commit()
return {"ok": True, "unlocked": unlocked, "counters": new_counters}
finally:
conn.close()
@app.post("/api/plugins/achievements/report-unlock")
def post_report_unlock(body: UnlockIn):
cls = "feat" if body.kind == "feat" else "competency"
at = body.at or _now_iso()
with _lock:
conn = _conn()
try:
changed = _record_unlock(
conn, body.id, cls, body.category, body.sourceId, body.tier, at)
# Only Feats sync; competency never enqueues (integration law +
# data-minimization contract).
if changed and cls == "feat":
_enqueue_feat_sync(conn, body.id, at)
conn.commit()
return {"ok": True, "changed": changed, "id": body.id, "tier": body.tier}
finally:
conn.close()
@app.post("/api/plugins/achievements/report-criterion")
def post_report_criterion(body: CriterionIn):
"""Record a distinct (criterion_id, token); return the distinct count.
Lets a baseline subscriber aggregate multi-event criteria (e.g. the set
of distinct days with a real advance → `steady_hands`) without us
re-deriving competency from activity. Bookkeeping over events only.
"""
with _lock:
conn = _conn()
try:
conn.execute(
"INSERT OR IGNORE INTO comp_ledger(criterion_id, token) VALUES (?, ?)",
(body.criterion_id, body.token),
)
row = conn.execute(
"SELECT COUNT(*) AS n FROM comp_ledger WHERE criterion_id=?",
(body.criterion_id,),
).fetchone()
conn.commit()
return {"ok": True, "count": int(row["n"]) if row else 0}
finally:
conn.close()
@app.get("/api/plugins/achievements/catalog")
def get_catalog():
with _lock:
conn = _conn()
try:
earned = _earned_map(conn)
finally:
conn.close()
return {"baseline": _state["baseline"], "earned": earned}
@app.get("/api/plugins/achievements/earned")
def get_earned():
with _lock:
conn = _conn()
try:
return {"earned": list(_earned_map(conn).values())}
finally:
conn.close()
@app.get("/api/plugins/achievements/feats")
def get_feats():
with _lock:
conn = _conn()
try:
rows = conn.execute(
"SELECT achievement_id, tier, unlocked_at FROM unlocks WHERE cls='feat'"
).fetchall()
finally:
conn.close()
out = []
for row in rows:
fid = row["achievement_id"]
f = _feat_by_id(fid) or {}
payload = _feat_payload(fid, f, int(row["tier"]))
payload["unlocked_at"] = row["unlocked_at"]
out.append(payload)
return {"feats": out}
@app.post("/api/plugins/achievements/remove-me")
def post_remove_me():
# Local removal works offline: drop the synced flag so nothing re-syncs,
# and enqueue a wall removal (drained in PR3). The wall identity
# (player_hash) is resolved server-side at drain time, not stored here.
_, player_hash = _identity()
with _lock:
conn = _conn()
try:
conn.execute("UPDATE unlocks SET synced=0 WHERE cls='feat'")
# Enqueue a wall removal only when we have an identity to key it
# by; the drain worker (below) POSTs it. Idempotent server-side.
if player_hash:
conn.execute(
"INSERT INTO sync_queue(kind, payload, state) VALUES ('remove', ?, 'pending')",
(json.dumps({"player_hash": player_hash}),),
)
conn.commit()
return {"ok": True}
finally:
conn.close()
_maybe_start_drain()
log.info("achievements engine ready (%d feats, baseline v%s)",
len(_state["feat_defs"]), str(_state["baseline"].get("version", "?")))
def _feat_payload(fid, feat, tier):
titles = feat.get("tier_titles") or []
title = titles[tier] if 0 <= tier < len(titles) else feat.get("title", fid)
return {
"id": fid,
"cls": "feat",
"tier": tier,
"title": title,
"description": feat.get("description", ""),
"category": feat.get("category", "global"),
"secret": bool(feat.get("secret", False)),
}
def _earned_map(conn):
out = {}
for row in conn.execute(
"SELECT achievement_id, cls, disp_category, tier, unlocked_at FROM unlocks"
):
out[row["achievement_id"]] = {
"id": row["achievement_id"],
"cls": row["cls"],
"category": row["disp_category"],
"tier": int(row["tier"]),
"unlocked_at": row["unlocked_at"],
}
return out