v1.8.1: Fix

Fix a runtime error when room_id is empty.

v1.8.0: Custom help message and whitelist room

By design, the bot will automatically accept any invitation it receives. If a room_id is not set, the bot will work in the room it is in. In the past, only one room_id could be set; with this version, we can add multiple room_ids.

• Support custom help message
  #51
  Create a custom_help_message.txt and uncomment the field in docker compose file.
  
• Support several whitelist room_id
  You can add as many as room_id that you hope the bot works on.
  "room_id": ["!111111:matrix.qqs.tw", "!222222:matrix.qqs.tw"] for config.json
ROOM_ID="!aaaaaaa:XXX.XXX,!bbbbbb:XXX.XXX" for .env

V1.7.2 More commands in thread chatting

• Refactor gpt vision trigger method
• !pic, !help, and gpt vision in thread chat
• Add GPT Vision thread level response to context

Breaking

GPT Vision can be triggered when quote a image and mention bot with prompt in room level or quote a image with prompt in thread level.
You can read the document for reference.

Full Changelog: v1.7.1...v1.7.2

v1.7.1 Some improvements

• Fix official openai image generation(#46)
• Thread level chatting for Element Android(#47)

Full Changelog: v1.7.0...v1.7.1

v1.7.0: Thread level context

Feature: room level and thread level context or both #29

Full Changelog: v1.6.0...v1.7.0

v1.6.0: GPT Vision

https://github.com/hibobmaster/matrix_chatgpt_bot/blob/main/CHANGELOG.md

Notice: GPT Vision won't work in E2EE room due to how i implement it.

Full Changelog: v1.5.3...v1.6.0

v1.5.3

• Make gptbot more compatible by using non-streaming method

v1.5.2

• feat: Expose more stable diffusion webui api parameters https://github.com/hibobmaster/matrix_chatgpt_bot/wiki

v1.5.1

• fix: set timeout not work in image generation

v1.5.0

• Fix localai v2.0+ image generation
• Fallback to gpt-3.5-turbo when caculate tokens using custom model
• Support specific output image format(jpeg, webp, png) and size