chore: project cleanup — purge stale function-calling refs + sync docs

Followed code-reviewer audit. Findings applied:
- twentyq/README.md, twentyq/index.js header — claimed "function calling"
  + ANSWER_FUNCTION_SCHEMA / submit_answer; rewrite to JSON-in-content
  matching what the code actually does. Added generateRoundStart line.
- wrangler.toml [ai] comment — list both bge-m3 (semantle/doantu) AND
  gemma-4 (twentyq) consumers; drop neuron math that no longer matched.
- scripts/stub-kv.js — drop reference to nonexistent REGISTER_DRYRUN flag.
- twentyq/ai-client.redactSecret — strip dead "if (out.length > 0)" branch
  (String.replace cannot produce empty string from the inputs we pass).
- handlers.test.js — drop noise saveGame() before "no games" stats assert;
  add ai.run call-count guards on two-AI-call flows.
- docs/codebase-summary.md — full rewrite of Active Modules table
  (semantle/doantu/lolschedule/twentyq were missing); fix vitest 2→4 +
  wrangler 3→4 versions; replace stale 200-test count with current ~450.
- docs/architecture.md — file tree includes lolschedule/semantle/doantu/
  twentyq + cron-dispatcher + sql-store* + scripts/migrate.js;
  moduleRegistry snippet matches src/modules/index.js.
- docs/todo.md — entire file obsolete (D1 UUID populated, cron live).
  Deleted.

Tests: 449 pass, lint clean.
This commit is contained in:
2026-04-24 18:30:25 +07:00
parent f6ab94ffb0
commit 3be799d68a
9 changed files with 78 additions and 114 deletions
+29 -15
View File
@@ -17,28 +17,38 @@ For authoring a new plugin module, see [`adding-a-module.md`](./adding-a-module.
```
src/
├── index.js ── fetch router: POST /webhook + GET / health
├── index.js ── fetch + scheduled handlers
├── bot.js ── memoized grammY Bot factory, lazy dispatcher install
├── db/
│ ├── kv-store-interface.js ── JSDoc typedefs only — the contract
│ ├── kv-store-interface.js ── KVStore contract (JSDoc)
│ ├── cf-kv-store.js ── Cloudflare KV adapter
── create-store.js ── per-module prefixing factory
── create-store.js ── KV per-module prefixing factory
│ ├── sql-store-interface.js ── SqlStore contract (JSDoc)
│ ├── cf-sql-store.js ── Cloudflare D1 adapter
│ └── create-sql-store.js ── D1 per-module prefixing factory
├── modules/
│ ├── index.js ── static import map (add new modules here)
│ ├── registry.js ── loader + builder + conflict detection + memoization
│ ├── dispatcher.js ── bot.command() for every visibility
│ ├── validate-command.js ── shared validators
│ ├── util/ ── fully implemented: /info + /help
│ ├── trading/ ── paper trading: VN stocks (dynamic symbol resolution)
│ ├── wordle/ ── 5-letter guessing game (KV storage)
│ ├── loldle/ ── classic-mode LoL champion guesser (KV storage)
── misc/ ── stub that exercises the DB (ping/mstats)
│ ├── cron-dispatcher.js ── routes scheduled events to matching module crons
│ ├── validate-command.js ── command contract validator
│ ├── validate-cron.js ── cron contract validator
│ ├── util/ ── /info + /help
│ ├── misc/ ── stub: /ping + /mstats
── trading/ ── paper trading: VN stocks (D1 + KV, daily cron)
│ ├── wordle/ ── 5-letter guessing game (KV)
│ ├── loldle/ ── classic-mode LoL champion guesser (KV)
│ ├── lolschedule/ ── LoL esports schedule + daily digest subscriptions (KV, cron)
│ ├── semantle/ ── English semantic word guess (KV, word2sim)
│ ├── doantu/ ── Vietnamese semantle (KV, phow2sim)
│ └── twentyq/ ── reverse-Akinator yes/no game (KV + Workers AI)
└── util/
└── escape-html.js
scripts/
├── register.js ── post-deploy: setWebhook + setMyCommands
── stub-kv.js ── no-op KV binding for deploy-time registry build
── migrate.js ── apply D1 migrations
└── stub-kv.js ── no-op KV + AI bindings for deploy-time registry build
```
## 3. Cold-start and the bot factory
@@ -104,11 +114,15 @@ Cloudflare Workers bundle statically via wrangler. A dynamic import from a varia
```js
// src/modules/index.js
export const moduleRegistry = {
util: () => import("./util/index.js"),
wordle: () => import("./wordle/index.js"),
loldle: () => import("./loldle/index.js"),
misc: () => import("./misc/index.js"),
trading: () => import("./trading/index.js"),
util: () => import("./util/index.js"),
wordle: () => import("./wordle/index.js"),
loldle: () => import("./loldle/index.js"),
misc: () => import("./misc/index.js"),
trading: () => import("./trading/index.js"),
lolschedule: () => import("./lolschedule/index.js"),
semantle: () => import("./semantle/index.js"),
doantu: () => import("./doantu/index.js"),
twentyq: () => import("./twentyq/index.js"),
};
```
+18 -24
View File
@@ -2,7 +2,7 @@
## Overview
Telegram bot on Cloudflare Workers with a plug-n-play module system. grammY handles Telegram API; modules register commands with three visibility levels. Data stored in Cloudflare KV behind a prefixed `KVStore` interface.
Telegram bot on Cloudflare Workers with a plug-n-play module system. grammY handles Telegram API; modules register commands with three visibility levels. Data stored in Cloudflare KV (behind a prefixed `KVStore` interface) or D1 (behind `SqlStore` interface).
## Tech Stack
@@ -10,21 +10,25 @@ Telegram bot on Cloudflare Workers with a plug-n-play module system. grammY hand
|-------|-----------|
| Runtime | Cloudflare Workers (V8 isolates) |
| Bot framework | grammY 1.x |
| Storage | Cloudflare KV |
| Storage | Cloudflare KV + D1 |
| AI inference | Workers AI binding (`env.AI`) |
| Linter/Formatter | Biome |
| Tests | Vitest |
| Deploy | Wrangler CLI |
## Active Modules
| Module | Status | Commands | Storage | Crons | Description |
|--------|--------|----------|---------|-------|-------------|
| `util` | Complete | `/info`, `/help`, `/stickerid` (private) | — | — | Bot info, command help renderer, and sticker file_id echo helper |
| `trading` | Complete | `/trade_topup`, `/trade_buy`, `/trade_sell`, `/trade_convert`, `/trade_stats`, `/history` | D1 (trades) + KV (portfolio, symbol cache) | Daily 5PM trim | Paper trading — VN stocks with dynamic symbol resolution. Crypto/gold/forex coming soon. |
| `wordle` | Complete | `/wordle`, `/wordle_new`, `/wordle_giveup`, `/wordle_stats` | KV (game, stats) | — | Classic 5-letter word game. 14,855-word dict sourced from [dracos's gist](https://gist.github.com/dracos/dd0668f281e685bad51479e5acaadb93). |
| `loldle` | Complete | `/loldle`, `/loldle_giveup`, `/loldle_stats` | KV (game, stats) | — | Classic-mode LoL champion guesser (auto-starts a new round after solve/giveup). Champion data synced from `tiennm99/loldle-data`. |
| `twentyq` | Complete | `/twentyq`, `/twentyq_giveup`, `/twentyq_stats` | KV (game, stats) | — | Reverse-Akinator yes/no game. Workers AI (`@cf/google/gemma-4-26b-a4b-it`) judges each question via function calling + generates fresh hints. |
| `misc` | Stub | `/ping`, `/mstats`, `/fortytwo` | KV | — | Health check + DB demo |
| Module | Commands | Storage | Crons | Description |
|--------|----------|---------|-------|-------------|
| `util` | `/info`, `/help`, `/stickerid` (private) | — | — | Bot info, command help renderer, sticker file_id echo helper |
| `misc` | `/ping`, `/mstats`, `/fortytwo` | KV | — | Health check + DB demo stub |
| `trading` | `/trade_topup`, `/trade_buy`, `/trade_sell`, `/trade_convert`, `/trade_stats`, `/history` | D1 (trades) + KV (portfolio, symbol cache) | Daily 5PM trim | Paper trading — VN stocks with dynamic symbol resolution |
| `wordle` | `/wordle`, `/wordle_new`, `/wordle_giveup`, `/wordle_stats` | KV | — | 5-letter word guessing game. 14,855-word dict |
| `loldle` | `/loldle`, `/loldle_giveup`, `/loldle_stats` | KV | — | Classic-mode LoL champion guesser. Data synced from `tiennm99/loldle-data` |
| `lolschedule` | `/lolschedule_today`, `/lolschedule_week`, `/lolschedule_subscribe`, `/lolschedule_unsubscribe` | KV | Daily 01:00 UTC | LoL esports schedule + daily digest subscriptions |
| `semantle` | `/semantle`, `/semantle_giveup`, `/semantle_stats` | KV | — | English semantic word guessing via hosted word2sim service |
| `doantu` | `/doantu`, `/doantu_hint`, `/doantu_giveup`, `/doantu_stats` | KV | — | Vietnamese semantle via hosted phow2sim service |
| `twentyq` | `/twentyq`, `/twentyq_giveup`, `/twentyq_stats` | KV | — | Reverse-Akinator yes/no game. Workers AI (`@cf/google/gemma-4-26b-a4b-it`) generates round-start category+hint and judges each turn via one-line JSON |
## Key Data Flows
@@ -62,23 +66,13 @@ npm run deploy
|-----------|---------|---------|
| `grammy` | Telegram Bot API framework | ^1.30.0 |
| `@biomejs/biome` | Linting + formatting (dev) | ^1.9.0 |
| `vitest` | Test runner (dev) | ^2.1.0 |
| `wrangler` | Cloudflare Workers CLI (dev) | ^3.90.0 |
| `vitest` | Test runner (dev) | ^4.1.4 |
| `wrangler` | Cloudflare Workers CLI (dev) | ^4.84.0 |
## Module Documentation
Each module maintains its own `README.md` with commands, data model, and implementation details. See `src/modules/<name>/README.md`.
## Test Coverage
## Tests
200 tests across 21 test files (run via `npm test` — ~2s):
| Area | Tests | What's Covered |
|------|-------|---------------|
| DB layer (KV) | 19 | KV store, prefixing, JSON helpers, pagination |
| DB layer (D1) | — | Fake D1 in-memory implementation (fake-d1.js) backs trading tests |
| Module framework | 33 | Registry, dispatcher, validators, help renderer, cron validation |
| Utilities | 4 | HTML escaping |
| Trading module | 79 | Symbol resolution, formatters, flat portfolio CRUD, command handlers, history/retention |
| Loldle module | 18 | Classic-mode champion comparison, champion lookup, daily picker |
| Wordle module | 13 | Duplicate-letter two-pass comparison, guess validation |
`npm test` runs the full vitest suite (run in a few seconds — ~450 tests). Structure: one folder per module under `tests/modules/<name>/`, shared fakes under `tests/fakes/` (fake-kv-namespace, fake-d1, fake-bot, fake-modules, fake-ai). No workerd, no Telegram fixtures — pure-logic unit tests with injected fakes.
-47
View File
@@ -1,47 +0,0 @@
# TODO
Manual follow-ups after the D1 + Cron infra rollout (plan: `plans/260415-1010-d1-cron-infra/`).
## Pre-deploy (required before next `npm run deploy`)
- [ ] Create the D1 database:
```bash
npx wrangler d1 create miti99bot-db
```
Copy the returned UUID.
- [ ] Replace `REPLACE_ME_D1_UUID` in `wrangler.toml` (`[[d1_databases]]` → `database_id`) with the real UUID.
- [ ] Commit `wrangler.toml` with the real UUID (the ID is not a secret).
## First deploy verification
- [ ] Run `npm run db:migrate -- --dry-run` — confirm it lists `src/modules/trading/migrations/0001_trades.sql` as pending.
- [ ] Run `npm run deploy` — chain is `wrangler deploy` → `npm run db:migrate` → `npm run register`.
- [ ] Verify in Cloudflare dashboard:
- D1 database `miti99bot-db` shows `trading_trades` + `_migrations` tables
- Worker shows a cron trigger `0 17 * * *`
## Post-deploy smoke tests
- [ ] Send `/buy VNM 10 80000` (or whatever the real buy syntax is) via Telegram, then `/history` — expect 1 row.
- [ ] Manually fire the cron to verify retention:
```bash
npx wrangler dev --test-scheduled
# in another terminal:
curl "http://localhost:8787/__scheduled?cron=0+17+*+*+*"
```
Check logs for `trim-trades` output.
## Nice-to-have (not blocking)
- [ ] End-to-end test of `wrangler dev --test-scheduled` documented with real output snippet in `docs/using-cron.md`.
- [ ] Decide on migration rollback story (currently forward-only). Either document "write a new migration to undo" explicitly, or add a `down/` convention.
- [ ] Tune `trim-trades` schedule if 17:00 UTC conflicts with anything — currently chosen as ~00:00 ICT.
- [ ] Consider per-environment D1 (staging vs prod) if a staging bot is added later.
+1 -1
View File
@@ -7,7 +7,7 @@
* These stubs satisfy the shape without doing any real IO. All init hooks
* are assumed read-only (or tolerant of missing state) at registration time.
* If a future module writes inside init(), update the matching stub to
* swallow writes or gate the write on a `process.env.REGISTER_DRYRUN` flag.
* swallow writes safely.
*/
/** @type {KVNamespace} */
+16 -11
View File
@@ -1,10 +1,12 @@
# Twentyq Module
A reverse-Akinator yes/no guessing game. The bot picks a secret object from a
hand-curated seed list, gives an opening hint, then judges every user input
with a Workers AI LLM (`@cf/google/gemma-4-26b-a4b-it`) via function calling.
Each turn the model returns `{ is_guess, answer, hint }`. Round ends on a
correct guess (`is it an organ?` matches secret) or `/twentyq_giveup`.
A reverse-Akinator yes/no guessing game. The bot picks a secret keyword from
a flat seed list. At round-start, a Workers AI LLM
(`@cf/google/gemma-4-26b-a4b-it`) generates a category + cryptic opening
hint. Each turn the same model judges the user's input and returns
`{ is_guess, answer, hint }` as one-line JSON (parsed out of response text —
no function calling, no tools array). Round ends on a correct guess or
`/twentyq_giveup`.
**Visibility: `public`** — commands appear in both `/help` and Telegram's
native `/` autocomplete menu.
@@ -60,16 +62,19 @@ AI request) + one request per turn, well under the cap for normal play volume.
## Architecture
- `seeds.js``SEEDS` array + `getRandomSeed(rng)`. Targets are lowercased.
- `seeds.js` flat `SEEDS` string array of target keywords + `getRandomSeed(rng)`.
- `state.js` — KV persistence for game + stats. Subject = user id (DM) or
chat id (group). 7-day TTL on the active round.
- `prompts.js``buildSystemPrompt(state)` injects secret + history;
`ANSWER_FUNCTION_SCHEMA` declares the `submit_answer` tool.
- `prompts.js``buildStartRoundPrompt(target)` (opens a round) and
`buildSystemPrompt(state)` (per-turn judging). Both include HINT STYLE
rules to keep hints cryptic.
- `validate-input.js` — pre-AI regex check; rejects open-ended starters,
empty/oversized input. Saves Neurons.
- `ai-client.js` — wraps `env.AI.run`, parses both Cloudflare-traditional and
OpenAI-style tool-call shapes, normalizes payload, redacts the secret from
hints (defense-in-depth). `UpstreamError` wraps any failure.
- `ai-client.js` — wraps `env.AI.run`. `generateRoundStart(env, target)`
produces `{category, initialHint}`; `judge(env, state, userInput)` produces
`{is_guess, answer, hint}`. Both parse one-line JSON from response text,
redact the secret word from any generated hint, and throw `UpstreamError`
on upstream failure.
- `render.js` — five Telegram-HTML formatters; all user-derived text
HTML-escaped.
- `handlers.js` — three command entry points + subject resolver + repeat
+1 -2
View File
@@ -126,8 +126,7 @@ export function redactSecret(hint, target) {
if (!target) return hint;
const escaped = target.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
const re = new RegExp(`\\b${escaped}\\b`, "ig");
const out = hint.replace(re, "(redacted)");
return out.length > 0 ? out : "the hint was redacted to avoid revealing the answer";
return hint.replace(re, "(redacted)");
}
/**
+5 -5
View File
@@ -1,11 +1,11 @@
/**
* @file Twentyq module — reverse-Akinator yes/no guessing game.
*
* Bot picks a secret object from a hand-curated seed list (./seeds.js) and
* gives an initial hint. Each user input is judged by Workers AI
* (@cf/google/gemma-4-26b-a4b-it) via function calling — the model returns
* { is_guess, answer, hint }. Round ends on a correct guess or /twentyq_giveup.
* Unlimited turns. Per-subject state in KV (user id in DMs, chat id in groups).
* Bot picks a secret keyword from ./seeds.js. Workers AI
* (@cf/google/gemma-4-26b-a4b-it) generates {category, initialHint} at
* round start and emits a one-line JSON {is_guess, answer, hint} per turn.
* Round ends on a correct guess or /twentyq_giveup. Unlimited turns.
* Per-subject state in KV (user id in DMs, chat id in groups).
*
* `init` captures both the prefixed KV store AND the raw env so handlers can
* reach env.AI per request without changing the dispatcher contract.
+3 -3
View File
@@ -137,6 +137,7 @@ describe("twentyq/handlers", () => {
mockJudgement(ai, { is_guess: false, answer: "yes", hint: "yes hint" });
const ctx = makeCtx(1, "private", "/twentyq is it big?");
await handleTwentyq(ctx, { db, env });
expect(ai.run).toHaveBeenCalledTimes(2); // roundstart + judge
expect(ctx.reply).toHaveBeenCalledTimes(2); // intro + turn
expect(ctx.replies[0].text).toMatch(/I'm thinking/);
expect(ctx.replies[1].text).toMatch(/Yes/);
@@ -164,6 +165,7 @@ describe("twentyq/handlers", () => {
const ctx = makeCtx(99, "group", "/twentyq is it big?");
ctx.chat.id = 12345;
await handleTwentyq(ctx, { db, env });
expect(ai.run).toHaveBeenCalledTimes(2); // roundstart + judge
// Game saved under chat id (12345), not user id (99)
expect(await loadGame(db, 12345)).not.toBeNull();
expect(await loadGame(db, 99)).toBeNull();
@@ -190,11 +192,9 @@ describe("twentyq/handlers", () => {
});
describe("handleStats", () => {
it("renders stats summary", async () => {
await saveGame(db, 1, sampleGame());
it("renders empty-stats message when no rounds finished", async () => {
const ctx = makeCtx(1);
await handleStats(ctx, { db });
// No games played yet -> "no twentyq games"
expect(ctx.replies[0].text).toMatch(/no.*games/i);
});
});
+5 -6
View File
@@ -25,12 +25,11 @@ binding = "DB"
database_name = "miti99bot-db"
database_id = "261b54e7-0fdb-4fe7-8ed9-2e8a8bcf459c"
# Workers AI inference binding used by semantle + doantu for
# @cf/baai/bge-m3 multilingual text embeddings. Accessed as `env.AI`
# in handlers. Included on the Workers Free plan: 10,000 Neurons/day at
# no charge (hard-stops — no billing on Free plan).
# bge-m3 is 1075 Neurons per M input tokens → ~0.002 N/guess (2 short
# words), ~4.6M guesses/day within the cap.
# Workers AI inference binding, accessed as `env.AI` in handlers.
# Used by:
# - semantle / doantu → @cf/baai/bge-m3 multilingual embeddings
# - twentyq → @cf/google/gemma-4-26b-a4b-it text generation
# Workers Free plan: 10,000 Neurons/day, hard-stops (no billing on Free).
# Pricing: https://developers.cloudflare.com/workers-ai/platform/pricing/
[ai]
binding = "AI"