Add Telegram-sourced pipeline service with config, store, bot endpoints, and tests

[yl124915300-dot]

Motivation

• Provide a minimal, deployable service that fetches Telegram messages, analyzes them, generates drafts, and exposes a review/publish workflow.
• Support both mock and live modes so the pipeline can run without Telegram/OpenAI credentials and degrade gracefully when config is missing.

Description

• Introduce configuration handling via src/config.js with typed helpers and auto mode detection, and update package.json (version bump, test script, and optionalDependencies for telegram).
• Implement the core service in src/server.js with pipeline orchestration (runPipeline), simple LLM integration fallback (generateDraft), review state machine, bot command handling, HTTP API endpoints (/health, /pipeline/run, /pending, /approve, /reject, /publish, /bot/webhook, etc.), and periodic auto-run timer.
• Add persistent JSON-backed storage in src/store.js (JsonStore) with safe flush/backup and upsert helpers, and Telegram client integration in src/telegram.js with a mock fallback and sendBotMessage utility.
• Add docs and housekeeping: update top-level README.md, add detailed ai-hot-comments-render-oneclick-v2/README-RENDER.md, add .gitignore entries, and add tests under ai-hot-comments-render-oneclick-v2/test/ (smoke.test.js, store.test.js).

Testing

• Added unit tests test/smoke.test.js and test/store.test.js to validate config.mode and JsonStore.upsertMessages behavior.
• Ran the test suite with npm test (which executes node --test test/*.test.js) and the tests passed.

---

Codex Task