-
Notifications
You must be signed in to change notification settings - Fork 31
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Main Update 24 Feb #372
Merged
Main Update 24 Feb #372
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* make download_items responsive * remove those <view>
* Testing basic integration with MAE * Add first MAE agent files to our repo * Support an additional agent (web search) * First take on proper interface with MAE service * Fix double parsing issue * Refactor MAE backend to have only one loop * Use markdown format to display the full response from WebSearch agent * Add a fake MAE backend to allow quick testing * Pass config values to MAE agent * More progress on papers agent * Add MAE paper agent files * Make sure Dora workflow keeps running after interaction * Pass configuration values from Moxin to all parts of Paper agent * Display partial progress for interactions with Paper Research agent * squashed agent ui progress * prettier agents in left panel * reuse agent button * restore agent at-click * rename some stuff * select agent from left panel * remove unused accessor from computed list for now * correct comment * simple send message to agent * agent reply bubble follow design better * fix agent autocomple hiding last message * remove leftover * remove leftover * Show headers in chaat left sidebar * Rename agents * Rework history chat left bar * Show agent icons * Restore chat options menu * Some adjustments * Use the proper initial letter for chats that interacted with agents * Minor refactor around chat entity name * Consider Agent conversation in all places where avatar is displayed * Basic support to stop agent completitions * Consider message edition for new Agents case Note: We don't have yet a good way to interact with Agents when editing an inline message. It would send the message to the current model or agent, instead of re-sending to the original entity at the moment of creating the message. We need to define the proper UX here. * Fix problem to show agents options when typing @ * Clear selected agent in some flows * Show short description in agents dropdown menu * filter agents by text * select agent on enter * handle more show and hide agent autocomplete cases * add missing `short_description` replace with the proper one later * highlight agent to be submitted on enter * move around the agents list with keyboard pressing enter to confirm the selection still requires filter input focus * reset agents keyboard index on hide * on agent select, focus prompt input * use the actual filtered list of agents to clamp keyboard movement * remove no longer necessary livehook * Update Agents short description * deferred focus workaround for agent search input * fix strange bug where keyboard selection of agents hides items except the first one * fix focus of prompt input after clicking with mouse an agent * improve agents autocomplete cursor handling * hide agents autocomplete on focus lost * deselect agent with keyboard * group some widget actions * remove at if agent selected * replaced todo with workaround comment * Improve chat card title edition * Show agents in main models selector * Improve Agent selection in the main dropdown * Wrap up the entity that is associated to a chat session * Fix oustanding issue selecting Agent * dedup ChatAgentAvatar dsl * take adventage of previous commit at prompt_input * agents in landing sketch * better cards in landing agents list * start chat with agent from any screen doing the same we were doing for models * Support markdown links * Improve links in Markdown formatted messages * Improve agents in main model/agents selector * Remove agent selection in the prompt on model/agent change * agents list inside scrollable area of landing * upgrade agent avatars to 150x150 * agent card improved design * flat agent list at landing * Fix problem with modals in chat history * Make sure chat message is sent to chat's active agent * Fix leftovers * draw_text to draw_label in prompt input * move drag_scrolling inside portal list * Bump makepad (tmp branch) * use fake mae backend if MAE_BACKEND=fake * mirrored chat * battle screen layout from scratch * introduce a safer way to link data to dyn widgets * keep agent battle screen hidden if not enough agents * started implementing agent selector * agent battle prompt input basic visuals * agent selector selecting * agent selector visuals * overlay agent selector * animated agent selector * handle prompt submit * draw_label -> draw_text * almost isolated mae interface * mae automatic isolation * handle enter in prompt input * remove unused use * isolated fake delay * clear prompt input on submit * scroll to bottom * message loading animation * battle screen further styling * identify loading messages * match better prompt input style * quickly add icons to agent selector * battle to arena * empty state * a little bit more dummy agent selector + some refactors + renames + other small stuff * fix scroll bar appearing inside messages * use bare text for user messages and markdown for agent ones * handle markdown links at app level * battle screen remove & replace widgets for v2 * battle screen loaded from battle sheet * vote buttons * rounds passing * small details * round counter * insert code screen * simulate send battle sheet * removed some unused files * refactor * icon in spinner * more fake icons * Migrate mae backend to new MoFa HTTP API * Some fixes * Fix formatting * Remove not longer used options * Mae -> MoFa rename * Remove Arena code * Adjust web interface based in latest MoFa developments * Add MoFa settings fields * Simplify to have only one agent per MoFa installation * Add visual indication to check MoFa server status * Tweaks to improve the MoFa availability check * Catching up with Makepad * Improve MoFa adddress edit in Settings * Fix naming issue in GH release config * Update release config * Leave only one Example Agent * Adjust positions in the chat input * Fix links styling issues * Silent missing View animation warning * Ignore trailing slash char when configuring MoFa address * Fix first app usage for the MoFa case * Fix crash after removing a model file previouly used * Make sure models and agent list selector is refresh when models are deleted * Bump makepad * Remove outdated handling of history panel toggle * remove conflict resolution signal leftovers * fix start chat with agent from landing * change mofa backend runtime env ff for the positive * quick mofa hide * Fix crashes when mofa is hidden * Add spacing to mofa settings * Set localhost as default mofa url * Cleanup --------- Co-authored-by: noxware <7684329+noxware@users.noreply.github.com> Co-authored-by: Julian Montes de Oca <julian@wyeworks.com>
Catch up with Makepad's theming changes
Remember model used at creation time of a model
* try mixing models and agents in autocomplete * loading model from at selection * wip maybe model selector not the same as invoked model * unify more chat entities and fix keyboard selection of entity * fix model button dsl with latest makepad syntax * address fragile todo * remove agent button tighly coupled properties * unify model button into agent button * remove ChatHandler and favor chatentityid * remove previously necessary steps to convert chat entity id to chat handler * remove unnecessary owned ChatEntity type * avoid unnecessary premature string allocations and cloning * tidy many details * remove ChatEntityRef::from_chain in favor of explicit sorting * move chat entity types to their own file * hide loader indicator for at model messages * take adventage of list items() to dedup handling of prompt input items * remove ChatEntityRef::from as enum variants can be mapped directly * allow set_agent directly in entity button * fix type name after rebase * Add autocomplete section headers + text updates - Added section labels for "Agents" and "Models" in the autocomplete list. - Updated the empty message to clarify search options. - Keyboard focus handling skips section headers. - Minor layout and padding adjustments --------- Co-authored-by: Julian Montes de Oca <julian@wyeworks.com>
* fix chat switching issues by reacting to the propety change directly * remove not longer necessary action reaction * remove ChatSelected action as not longer necessary * restore a suspecious line just in case
Fix chat switch streaming cancel
Signed-off-by: csh <458761603@qq.com>
Adds support for multiple concurrent MoFa servers, both locally and remotely.
* mofa fake multi servers * always insert default mofa address
* mofa fake multi servers * always insert default mofa address * use command text input for prompt input * update Cargo.toml to use command text input branch * back to rik
* fix loop redraw in chat history card * fix loop redraw when a message is sent * fix history card title update
* Bump version of windows SDK setup action in CI workflow * Bump Windows 10 SDK version to 20348
…338) * Fix the incompatibility issue of libwasmedge.0.dylib on macos x86_64 * Add arch arg * Use cfg(target_arch) to replace arg
…346) * Update WasmEdge version to 0.14.1 across all configurations and docs * Revert Windows wasn_nn plugin URLs to v0.14.0
* Bump Makepad version - update set_visible and set_text usage * Add missing spaces after (cx,...) for consistency * Revert AI autocomplete slop * Bump Makepad
* Remove redundant visibility checks * Update README with detailed setup instructions
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Merging a stable integration with MoFa