melod1n/tg-chat-bot
1
0
Fork 0
Code Issues Actions Packages Projects Releases Wiki Activity
126 Commits 7 Branches 0 Tags
6a770683ebd560d7b5c5f345ea8036640a84ae4b
Commit Graph

16 Commits

Author SHA1 Message Date
melod1n 82468cf61b chat with gemini now will use any first images 2026-02-12 14:48:13 +03:00
melod1n 9ba3d81a21 feat(ai): add configurable default provider + refactor telegram handlers 
- add AiProvider enum and DEFAULT_AI_PROVIDER env var (fallback: OLLAMA)
- route chat execution based on selected provider (Ollama/Gemini/Mistral/OpenAI)
- move inline query / callback / edited message / my_chat_member handlers into utils
- minor cleanup (command requirements placement, whitespace)
 2026-02-04 12:55:04 +03:00
melod1n a736f786c2 feat(openai): add chat streaming and model commands 
- add /openai (/chatgpt) chat command using OpenAI Responses API (streaming + incremental edits)
- add /openAIListModels, /openAIGetModel, /openAISetModel
- introduce base Command class and migrate non-chat commands to it
- wire OpenAI client + env vars (OPENAI_API_KEY, OPENAI_MODEL)
- bump deps (@google/genai, systeminformation, @types/node) and add openai
 2026-02-03 13:39:01 +03:00
melod1n 9e30086af2 fix(images): cache Telegram photos by unique_id and pass base64 to LLM commands 
Stop reading image files inside gemini/mistral commands; use pre-encoded image data from message parts

Rework loadImagesIfExists / loadImagesFromFileIds to return cached file_unique_ids and download only missing sizes

Encode cached images to base64 when assembling the reply chain

Make getPhotoMaxSize synchronous (returns PhotoSize), and map to URL only when needed

Await MessageStore.put / UserStore.put and prefetch single-image downloads on message receipt
 2026-01-29 20:58:08 +03:00
melod1n b74e0a0f3e refactor(bot): centralize runtime state; support albums + safer vision handling 
- make MessageStore.put() return StoredMessage and allow collectReplyChainText() to work with StoredMessage
- move muted users + answers loading into Environment (add Answers model + GEMINI_IMAGE_MODEL)
- extract message handling into processNewMessage() and add media-group (album) caching/downloading by unique_file_id
- for Ollama: check model capabilities before sending images; use replyToMessage helper consistently
- add /geminiGenImage command stub for Gemini image generation
 2026-01-29 20:01:30 +03:00
melod1n 486a7a4f11 update gemini chat to interactions api 2026-01-28 18:19:53 +03:00
melod1n 63c64dfa8e update error logging 
cutting all kind of prefixes for prompt making
 2026-01-28 16:40:29 +03:00
melod1n 458e56dcb4 update streaming for Gemini and Mistral and update list gemini models and list mistral models commands 2026-01-28 12:29:52 +03:00
melod1n 32baaebb93 update regex matching - now bot responds only for base commands (like /ping) and with mention (like /ping@panfilovi4_bot) and ignores command if it mentions other bot; 
Add ability to see, change and list gemini, mistral and ollama models
 2026-01-20 08:31:05 +03:00
melod1n ad8e3115db add gemini and mistral models to environment 2026-01-16 19:03:21 +03:00
melod1n 97582881ea optional usage of names in prompt 2026-01-16 16:14:40 +03:00
melod1n c491329d4f optional usage of names in prompt 2026-01-16 16:13:10 +03:00
melod1n 7047a5fa15 add user's names to messages 2026-01-16 16:04:10 +03:00
melod1n 326cfcafc8 /mistral command for chat with Mistral AI 2026-01-16 15:40:15 +03:00
melod1n 6eff5d17ea test support for images (ollama vision models) 2026-01-12 19:19:08 +03:00
melod1n df9471a7e4 refactor!: rewrite bot core; add AI (Ollama, Gemini), DB, new commands 2026-01-12 15:32:50 +03:00