asxpi/tg-channel-analysis
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Add multi-user features: i18n, payments, multi-model AI, focus areas

Browse source 
Replace single report type with composable analysis: depth levels
(basic/standard/full) + focus area multi-select (psychology, business,
marketing, content, audience, sentiment). Multi-step inline keyboard
flow guides users through selection.

- i18n: English + Russian, auto-detect from Telegram, /lang override
- AI providers: Anthropic + OpenRouter via AIClient abstraction
- Telegram Stars payments with per-depth pricing and free trial
- SQLite (aiosqlite) for users, analyses, payments tracking
- User middleware for auto-registration and language detection
- Report persistence: save .md locally, offer file download
- New commands: /features, /prices, /lang
- Composable prompt system: depth modifiers + focus area fragments
This commit is contained in:
Sergei Poljanski 2026-02-23 01:18:12 +02:00
commit c30b7e4675
27 changed files with 1155 additions and 280 deletions
Download patch file Download diff file Expand all files Collapse all files

35
CLAUDE.md
View file

@ -4,40 +4,55 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
## Overview
Telegram Channel Analyzer Bot — fetches public channel history via Telethon, analyzes with Claude AI (chunked summarization pipeline), delivers reports via aiogram bot.
Telegram Channel Analyzer Bot — multi-user public bot with payments, multi-language (EN/RU), multi-provider AI (Anthropic + OpenRouter). Fetches public channel history via Telethon, analyzes with configurable AI models (chunked summarization pipeline), delivers HTML reports via aiogram bot.
## Stack
- **Python 3.12**, async throughout
- **aiogram 3** — Telegram bot interface (commands, inline keyboards, progress messages)
- **aiogram 3** — Telegram bot interface (commands, inline keyboards, payments, progress messages)
- **Telethon** — userbot client for reading public channel history
- **anthropic** (AsyncAnthropic) — Claude API with streaming
- **httpx** — OpenRouter API calls (transitive dep of anthropic)
- **aiosqlite** — SQLite database for users, usage, payments
- **pydantic-settings** — config from environment variables
## Architecture
- Both aiogram and Telethon share one asyncio loop (no threads)
- Telethon client is attached to the bot instance in `__main__.py`
- Analysis pipeline: Fetch → Chunk (token-bounded) → Summarize each chunk → Synthesize final report
- Analysis flow: Channel → Depth selection → Focus areas (multi-select) → Model → Payment check → Pipeline
- Pipeline: Fetch → Chunk (token-bounded) → Summarize each chunk → Synthesize final report → Save .md → Send
- Rate limit handling: semaphore(1), 60s cooldown between chunks, retry with backoff on 429
- Output: Markdown→HTML conversion, split on section boundaries at 4000 chars
- User middleware auto-creates DB user, detects language, injects `lang`/`db_user` into handler data
- In-memory `_sessions` dict tracks multi-step analysis flow per user
## Key Files
- `bot/config.py` — all settings from env vars, `CLAUDE_MODEL` selects the model
- `bot/services/analyzer.py` — Claude API calls, adaptive thinking only for opus-4-6
- `bot/services/chunker.py` — `MAX_TOKENS_PER_CHUNK` and `CHARS_PER_TOKEN` control chunking
- `bot/prompts/` — prompt templates per report type (chunk_summary.py, synthesis.py)
- `bot/config.py` — all settings from env vars (Anthropic, OpenRouter, pricing, paths)
- `bot/models.py` — `Depth`, `FocusArea` enums, `AnalysisSession` dataclass
- `bot/i18n/` — `Lang` enum, `t()` lookup, all UI strings in `strings.py`
- `bot/db/` — aiosqlite engine, `user_repo`, `usage_repo`
- `bot/middleware/user_middleware.py` — auto-create user, detect lang
- `bot/services/ai_client.py` — `AIClient` ABC, `AnthropicClient`, `OpenRouterClient`, `get_ai_client()`
- `bot/services/analyzer.py` — orchestrates chunk analysis + synthesis with retry
- `bot/services/report_saver.py` — saves .md to `data/reports/`
- `bot/prompts/` — composable prompts: `depth.py`, `focus_areas.py`, `chunk_summary.py`, `synthesis.py`
- `bot/handlers/analyze.py` — multi-step flow (depth→focus→model→pay→run)
- `bot/handlers/payment.py` — pre_checkout handler
## Running
- Container-based: `Containerfile` + `compose.yml`
- `--login` flag for interactive Telethon session creation
- Session persists in `data/` volume
- Session persists in `data/` volume, DB at `data/bot.db`, reports at `data/reports/`
- `.env` file must not have inline comments (Podman/Docker limitation)
## Common Tasks
- To change chunk size: edit `MAX_TOKENS_PER_CHUNK` in `bot/services/chunker.py`
- To add a report type: add to `ReportType` enum, add prompts in both `prompts/` files
- To change model: set `CLAUDE_MODEL` env var; thinking params auto-adapt in `analyzer.py`
- To add a focus area: add to `FocusArea` enum, add prompt fragments in `prompts/focus_areas.py`, add i18n strings
- To add an AI model: add to `AVAILABLE_MODELS` env var (use `org/model` format for OpenRouter)
- To change pricing: set `PRICE_BASIC`/`PRICE_STANDARD`/`PRICE_FULL` env vars
- To change free trial count: set `FREE_ANALYSES` env var
- To add a language: add to `Lang` enum, add translations in `i18n/strings.py`