feat: Allow continuing with AI response after interrupting

[Vanalite]

Describe Your Changes

• Partial / user-terminated llm responses should be saved in threads
• Allow to continue with AI response after interrupting
• This behavior only applies for llamacpp
• Other providers' behavior remain untouched
• If users switch models during the interruption, the response should be generated from the beginning to avoid cross models problems.

Fixes Issues

• Closes bug: Partial / user-terminated llm responses should be saved in threads #5571

Test plan

Screen.Recording.2025-10-01.at.5.44.50.PM.mov

• Check behaviors with other providers

Self Checklist

• Added relevant comments, esp in complex areas
• Updated docs (for bug fixes / features)
• Created issues for follow-up changes or refactoring needed

[github-actions]

Barecheck - Code coverage report

Total: 30.06%

Your code coverage diff: 0.15% ▴

Uncovered files and lines

File	Lines
web-app/src/containers/GenerateResponseButton.tsx	1-7, 9-12, 15-23, 26-29, 31-35, 37, 39, 41-43, 45, 47-51, 54-68, 70-76, 78-93, 95-101, 103
web-app/src/containers/ScrollToBottom.tsx	1-12, 14-17, 20-29, 31-33, 36-41, 44-46, 48-50, 52, 54-64, 66-72, 74-83, 85-90, 92-94, 96-97, 99, 101, 103
web-app/src/containers/StreamingContent.tsx	1-5, 12-15, 17-20, 24-27, 29-34, 37-40, 42-50, 52-54, 56-58, 61-63, 67-83, 85, 87
web-app/src/hooks/useAppState.ts	81, 105-114, 130-133, 149-157, 164-167, 169-172, 174-177
web-app/src/hooks/useChat.ts	47-57, 59-63, 67-73, 99-114, 116-119, 121-127, 130-138, 140-152, 154-160, 162-166, 168-174, 176-177, 179-181, 184, 186-188, 190-191, 194-200, 202-203, 205-210, 212-217, 219-222, 224-226, 229-231, 294-295, 299-303, 306-316, 318-327, 330-332, 335-337, 339-343, 349-358, 368-390, 393, 395, 397, 400, 403-408, 410-411, 416-418, 421-425, 428, 431-433, 435-445, 479-480, 482-485, 487-489, 491-511, 514-515, 517-520, 523-525, 527-530, 533-540, 543-557, 559-561, 581-582, 618, 637-638, 640, 646, 650-652, 655, 657-664, 679-690, 725-727, 730-731, 733-734, 736-754, 756-766, 768-773, 775-784, 788-790, 792-796, 828, 855-856, 865-866, 869, 871-890, 892-908, 918-919, 922, 925, 927-962, 968-969
web-app/src/hooks/useMessages.ts	68, 72-73, 75-76, 78-84, 87-94, 98-101, 115-116
web-app/src/services/messages/default.ts	18-19, 32-33, 45-47, 49-53, 55, 60-61

[qnixsynapse]

This needs to be rebased I think

[Vanalite]

Rebase to latest dev and implement the force restart if they swap models midway @qnixsynapse

[Copilot]

Pull Request Overview

This PR implements a "Continue with AI Response" feature that allows users to continue generating a stopped/interrupted AI response from where it left off. When a user stops an AI response mid-generation, the system now preserves it as a "Stopped" message, which can later be continued with the same or a different model.

Key changes:

• Added modifyMessage method to the message interface and services to support updating existing messages
• Implemented updateMessage state function for optimistic UI updates with persistence
• Enhanced chat functionality to support continuing from a stopped message via continueFromMessageId parameter

Reviewed Changes

Copilot reviewed 12 out of 13 changed files in this pull request and generated 12 comments.

Show a summary per file

File	Description
yarn.lock	Added type definitions for @types/node@^20.0.0 and @types/whatwg-mimetype@^3.0.2
web-app/src/services/messages/types.ts	Added modifyMessage method to MessagesService interface
web-app/src/services/messages/default.ts	Implemented modifyMessage method for DefaultMessagesService
web-app/src/locales/en/common.json	Added translation key for "Continue with AI Response"
web-app/src/hooks/useMessages.ts	Implemented updateMessage function with optimistic updates and persistence
web-app/src/hooks/useChat.ts	Major refactoring to support continuing interrupted messages, including helper functions for streaming and message finalization
web-app/src/hooks/useAppState.ts	Exported PromptProgress type for use in other modules
web-app/src/hooks/tests/useMessages.test.ts	Updated tests to verify optimistic update behavior
web-app/src/hooks/tests/useChat.test.ts	Added comprehensive tests for continue functionality
web-app/src/containers/StreamingContent.tsx	Added logic to hide streaming content when a stopped message exists
web-app/src/containers/ScrollToBottom.tsx	Added logic to detect partial responses and model mismatches, showing continue button
web-app/src/containers/GenerateResponseButton.tsx	Implemented logic to continue or regenerate based on partial response and model mismatch
core/src/types/message/messageInterface.ts	Added modifyMessage method documentation to MessageInterface

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.