- Rename old "model_management" directory to "model_management_OLD" in order to catch
dangling references to original model manager.
- Caught and fixed most dangling references (still checking)
- Rename lora, textual_inversion and model_patcher modules
- Introduce a RawModel base class to simplfy the Union returned by the
model loaders.
- Tidy up the model manager 2-related tests. Add useful fixtures, and
a finalizer to the queue and installer fixtures that will stop the
services and release threads.
- ModelMetadataStoreService is now injected into ModelRecordStoreService
(these two services are really joined at the hip, and should someday be merged)
- ModelRecordStoreService is now injected into ModelManagerService
- Reduced timeout value for the various installer and download wait*() methods
- Introduced a Mock modelmanager for testing
- Removed bare print() statement with _logger in the install helper backend.
- Removed unused code from model loader init file
- Made `locker` a private variable in the `LoadedModel` object.
- Fixed up model merge frontend (will be deprecated anyway!)
- Update most model identifiers to be `{key: string}` instead of name/base/type. Doesn't change the model select components yet.
- Update model _parameters_, stored in redux, to be `{key: string, base: BaseModel}` - we need to store the base model to be able to check model compatibility. May want to store the whole config? Not sure...
- Replace legacy model manager service with the v2 manager.
- Update invocations to use new load interface.
- Fixed many but not all type checking errors in the invocations. Most
were unrelated to model manager
- Updated routes. All the new routes live under the route tag
`model_manager_v2`. To avoid confusion with the old routes,
they have the URL prefix `/api/v2/models`. The old routes
have been de-registered.
- Added a pytest for the loader.
- Updated documentation in contributing/MODEL_MANAGER.md
- Implement new model loader and modify invocations and embeddings
- Finish implementation loaders for all models currently supported by
InvokeAI.
- Move lora, textual_inversion, and model patching support into
backend/embeddings.
- Restore support for model cache statistics collection (a little ugly,
needs work).
- Fixed up invocations that load and patch models.
- Move seamless and silencewarnings utils into better location
- Cache stat collection enabled.
- Implemented ONNX loading.
- Add ability to specify the repo version variant in installer CLI.
- If caller asks for a repo version that doesn't exist, will fall back
to empty version rather than raising an error.
Unfortunately you cannot test for both a specific type of error and match its message. Splitting the error classes makes it easier to test expected error conditions.
The changes aim to deduplicate data between workflows and node templates, decoupling workflows from internal implementation details. A good amount of data that was needlessly duplicated from the node template to the workflow is removed.
These changes substantially reduce the file size of workflows (and therefore the images with embedded workflows):
- Default T2I SD1.5 workflow JSON is reduced from 23.7kb (798 lines) to 10.9kb (407 lines).
- Default tiled upscale workflow JSON is reduced from 102.7kb (3341 lines) to 51.9kb (1774 lines).
The trade-off is that we need to reference node templates to get things like the field type and other things. In practice, this is a non-issue, because we need a node template to do anything with a node anyways.
- Field types are not included in the workflow. They are always pulled from the node templates.
The field type is now properly an internal implementation detail and we can change it as needed. Previously this would require a migration for the workflow itself. With the v3 schema, the structure of a field type is an internal implementation detail that we are free to change as we see fit.
- Workflow nodes no long have an `outputs` property and there is no longer such a thing as a `FieldOutputInstance`. These are only on the templates.
These were never referenced at a time when we didn't also have the templates available, and there'd be no reason to do so.
- Node width and height are no longer stored in the node.
These weren't used. Also, per https://reactflow.dev/api-reference/types/node, we shouldn't be programmatically changing these properties. A future enhancement can properly add node resizing.
- `nodeTemplates` slice is merged back into `nodesSlice` as `nodes.templates`. Turns out it's just a hassle having these separate in separate slices.
- Workflow migration logic updated to support the new schema. V1 workflows migrate all the way to v3 now.
- Changes throughout the nodes code to accommodate the above changes.
Currently translated at 80.4% (1183 of 1470 strings)
Co-authored-by: Alexander Eichhorn <pfannkuchensack@einfach-doof.de>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/de/
Translation: InvokeAI/Web UI
Currently translated at 80.3% (1181 of 1470 strings)
translationBot(ui): update translation (German)
Currently translated at 80.1% (1178 of 1470 strings)
Co-authored-by: B N <berndnieschalk@gmail.com>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/de/
Translation: InvokeAI/Web UI
Currently translated at 80.0% (1176 of 1470 strings)
Co-authored-by: Alexander Eichhorn <pfannkuchensack@einfach-doof.de>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/de/
Translation: InvokeAI/Web UI
Currently translated at 95.6% (1406 of 1470 strings)
translationBot(ui): update translation (Italian)
Currently translated at 93.9% (1381 of 1470 strings)
Co-authored-by: Riccardo Giovanetti <riccardo.giovanetti@gmail.com>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/it/
Translation: InvokeAI/Web UI
* new workflow tab UI - still using shared state with workflow editor tab
* polish workflow details
* remove workflow tab, add edit/view mode to workflow slice and get that working to switch between within editor tab
* UI updates for view/edit mode
* cleanup
* add warning to view mode
* lint
* start with isTouched false
* working on styling mode toggle
* more UX iteration
* lint
* cleanup
* save original field values to state, add indicator if they have been changed and give user choice to reset
* lint
* fix import and commit translation
* dont switch to view mode when loading a workflow
* warns before clearing editor
* use folder icon
* fix(ui): track do not erase value when resetting field value
- When adding an exposed field, we need to add it to originalExposedFieldValues
- When removing an exposed field, we need to remove it from originalExposedFieldValues
- add `useFieldValue` and `useOriginalFieldValue` hooks to encapsulate related logic
* feat(ui): use IconButton for workflow view/edit button
* feat(ui): change icon for new workflow
It was the same as the workflow tab icon, confusing bc you think it's going to somehow take you to the tab.
* feat(ui): use render props for NewWorkflowConfirmationAlertDialog
There was a lot of potentially sensitive logic shared between the new workflow button and menu items. Also, two instances of ConfirmationAlertDialog.
Using a render prop deduplicates the logic & components
* fix(ui): do not mark workflow touched when loading workflow
This was occurring because the `nodesChanged` action is called by reactflow when loading a workflow. Specifically, it calculates and sets the node dimensions as it loads.
The existing logic set `isTouched` whenever this action was called.
The changes reactflow emits have types, and we can use the change types and data to determine if a change should result in the workflow being marked as touched.
* chore(ui): lint
* chore(ui): lint
* delete empty file
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>
Currently translated at 79.4% (1128 of 1419 strings)
translationBot(ui): update translation (German)
Currently translated at 78.1% (1107 of 1416 strings)
Co-authored-by: B N <berndnieschalk@gmail.com>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/de/
Translation: InvokeAI/Web UI
* remove thunk for receivedOpenApiSchema and use RTK query instead. add loading state for exposed fields
* clean up
* ignore any
* fix(ui): do not log on canceled openapi.json queries
- Rely on RTK Query for the `loadSchema` query by providing a custom `jsonReplacer` in our `dynamicBaseQuery`, so we don't need to manage error state.
- Detect when the query was canceled and do not log the error message in those situations.
* feat(ui): `utilitiesApi.endpoints.loadSchema` -> `appInfoApi.endpoints.getOpenAPISchema`
- Utilities is for server actions, move this to `appInfo` bc it fits better there.
- Rename to match convention for HTTP GET queries.
- Fix inverted logic in the `matchRejected` listener (typo'd this)
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>
Currently translated at 74.4% (1054 of 1416 strings)
translationBot(ui): update translation (German)
Currently translated at 69.6% (986 of 1416 strings)
translationBot(ui): update translation (German)
Currently translated at 68.6% (972 of 1416 strings)
Co-authored-by: B N <berndnieschalk@gmail.com>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/de/
Translation: InvokeAI/Web UI
Hardcode the options in the dropdown, don't rely on translators to fill this in.
Also, add a number of missing languages (Azerbaijani, Finnish, Hungarian, Swedish, Turkish).
Closes#5647
The alpha values in the UI are `0-1` but the backend wants `0-255`.
Previously, this was handled in `parseFIeldValue` when building the graph. In a recent release, field types were refactored and broke the alpha handling.
The logic for handling alpha values is moved into `ColorFieldInputComponent`, and `parseFieldValue` now just does no value transformations.
Though it would be a minor change, I'm leaving this function in because I don't want to change the rest of the logic except when necessary.
Closes#5616
Turns out the OpenAPI schema definition for a pydantic field with a `Literal` type annotation is different depending on the number of options.
When there is a single value (e.g. `Literal["foo"]`, this results in a `const` schema object. The schema parser didn't know how to handle this, and displayed a warning in the JS console.
This situation is now handled. When a `const` schema object is encountered, we interpret that as an `EnumField` with a single option.
I think this makes sense - if you had a truly constant value, you wouldn't make it a field, so a `const` must mean a dynamically generated enum that ended up with only a single option.
Currently translated at 40.6% (582 of 1433 strings)
translationBot(ui): update translation (Turkish)
Currently translated at 38.8% (557 of 1433 strings)
Co-authored-by: Ufuk Sarp Selçok <ilkel@live.com>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/tr/
Translation: InvokeAI/Web UI
The `getIntermediatesCount` query is set to `refetchOnMountOrArgsChange`. The intention was for when the settings modal opens (i.e. mounts), the `getIntermediatesCount` query is refetched. But it doesn't work - modals only mount once, there is no lazy rendering for them.
So we have to imperatively refetch, by refetching as we open the modal.
Closes#5639
* Port the command-line tools to use model_manager2
1.Reimplement the following:
- invokeai-model-install
- invokeai-merge
- invokeai-ti
To avoid breaking the original modeal manager, the udpated tools
have been renamed invokeai-model-install2 and invokeai-merge2. The
textual inversion training script should continue to work with
existing installations. The "starter" models now live in
`invokeai/configs/INITIAL_MODELS2.yaml`.
When the full model manager 2 is in place and working, I'll rename
these files and commands.
2. Add the `merge` route to the web API. This will merge two or three models,
resulting a new one.
- Note that because the model installer selectively installs the `fp16` variant
of models (rather than both 16- and 32-bit versions as previous),
the diffusers merge script will choke on any huggingface diffuserse models
that were downloaded with the new installer. Previously-downloaded models
should continue to merge correctly. I have a PR
upstream https://github.com/huggingface/diffusers/pull/6670 to fix
this.
3. (more important!)
During implementation of the CLI tools, found and fixed a number of small
runtime bugs in the model_manager2 implementation:
- During model database migration, if a registered models file was
not found on disk, the migration would be aborted. Now the
offending model is skipped with a log warning.
- Caught and fixed a condition in which the installer would download the
entire diffusers repo when the user provided a single `.safetensors`
file URL.
- Caught and fixed a condition in which the installer would raise an
exception and stop the app when a request for an unknown model's metadata
was passed to Civitai. Now an error is logged and the installer continues.
- Replaced the LoWRA starter LoRA with FlatColor. The former has been removed
from Civitai.
* fix ruff issue
---------
Co-authored-by: Lincoln Stein <lstein@gmail.com>
Currently translated at 60.8% (871 of 1432 strings)
Co-authored-by: Alexander Eichhorn <pfannkuchensack@einfach-doof.de>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/de/
Translation: InvokeAI/Web UI
* fix(ui): download image opens in new tab
In some environments, a simple `a` element cannot trigger a download of an image. Fetching the image directly can get around this and provide more reliable download functionality.
* use hook for imageUrlToBlob so token gets sent if needed
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
* redo top panel of workflow editor
* add checkbox option to save to project, integrate save-as flow into first time saving workflow
* remove log
* remove workflowLibrary as a feature that can be disabled
* lint
* feat(ui): make SaveWorkflowAsDialog a singleton
Fixes an issue where the workflow name would erroneously be an empty string (which it should show the current workflow name).
Also makes it easier to interact with this component.
- Extract the dialog state to a hook
- Render the dialog once in `<NodeEditor />`
- Use the hook in the various buttons that should open the dialog
- Fix a few wonkily named components (pre-existing issue)
* fix(ui): when saving a never-before-saved workflow, do not append " (copy)" to the name
* fix(ui): do not obscure workflow library button with add node popover
This component is kinda janky :/ the popover content somehow renders invisibly over the button. I think it's related to the `<PopoverAnchor />.
Need to redo this in the future, but for now, making the popover render lazily fixes this.
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>
Adds adds ctrl/meta + scroll to change brush size on canvas.
* changed hotkeys
* new hotkey as an additional
* lint fixed"
* added ctrl scroll and removed hotkey
* using
* added fix
* feedbck_changes
* brush size change logic
* feat(ui): also check for meta key when modifying brush size
* feat(ui): add comment linking to where brush size algo was determined
---------
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>
Currently translated at 10.5% (151 of 1426 strings)
translationBot(ui): update translation (Turkish)
Currently translated at 8.1% (116 of 1426 strings)
translationBot(ui): update translation (Turkish)
Currently translated at 6.6% (95 of 1426 strings)
Co-authored-by: Ufuk Sarp Selçok <ilkel@live.com>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/tr/
Translation: InvokeAI/Web UI
It doesn't work now that the theme is external. I'm not sure how to fix it and not sure if it really did much (I don't think I ever got autocomplete...). Maybe it can be implemented in `@invoke-ai/ui-library`.
* dont show duplicate toasts if workflow actions fail due to auth
* dynamic order by options based on projectId
* add endpointName to authtoast to makeit unique per endpoint
* lint
* update toast logic to check based on endpoint name w type safety
* fix save as endpoit name
* lint
* fix type
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
* retain id through workflow state so that we correctly update or save new
* lint
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
* feat: ✨ "Remix Image" option on images
Adds a new "remix image" option where applicable, recalls all metadata except the seed
* refactor: 🚨 lint code
* feat: ✨ "Remix Image" option on images
Adds a new "remix image" option where applicable, recalls all metadata except the seed
* refactor: 🚨 lint code
* feat: ✨ add new remix hotkey to hotkeys modal
---------
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>
Currently translated at 60.0% (850 of 1415 strings)
Co-authored-by: Alexander Eichhorn <pfannkuchensack@einfach-doof.de>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/de/
Translation: InvokeAI/Web UI
Remove `trim()` from model identifier schema, which prevented parsed model identifiers from matching.
The root issue here is that model names are identifiers. This will be resolved in the model manager refactor.
Closes#5556
- Bump `@invoke-ai/ui` for updated styles
- Update regex to parse prompts with newlines
- Update styling of overlay button when prompt has an error
- Fix bug where loading and error state sometimes weren't cleared
We had a one-behind issue with recalling metadata items that had a model.
For example, when recalling LoRAs, we check against the current main model to decide whether or not the requested LoRA is compatible and may be recalled.
When recalling all params, we are often also recalling the main model, but the compat logic didn't compare against this new main model.
The logic is updated to check against the new main model, if one is being set.
Closes#5512
Currently translated at 97.3% (1365 of 1402 strings)
translationBot(ui): update translation (Italian)
Currently translated at 97.3% (1365 of 1402 strings)
translationBot(ui): update translation (Italian)
Currently translated at 97.3% (1365 of 1402 strings)
translationBot(ui): update translation (Italian)
Currently translated at 97.3% (1365 of 1402 strings)
Co-authored-by: Riccardo Giovanetti <riccardo.giovanetti@gmail.com>
Translate-URL: https://hosted.weblate.org/projects/invokeai/web-ui/it/
Translation: InvokeAI/Web UI
* resolved conflicts
* changed logo and some design changes
* feedback changes
* resolved conflicts
* changed logo and some design changes
* feedback changes
* lint fixed
* added translations
* some requested changes done
* all feedback changes done and replace links in settingsmenu comp
* fixed the gap between deps verisons & chnaged heights
* feat(ui): minor about modal styling
* feat(ui): tag app endpoints with FetchOnReconnect
* fix(ui): remove unused translation string
---------
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>
Co-authored-by: Kent Keirsey <31807370+hipsterusername@users.noreply.github.com>
Use each language's own language for their option in the language select. This falls back to the english translation if the language name isn't translated.
Per user feedback, this is preferrable to letting them expand when the window grows.
Also bumps `react-resizable-panels` now that one of my PRs is merged to fix an issue.
- Add various brand images, organise images
- Create favicon for docs pages (light blue version of key logo)
- Rename app title to `Invoke - Community Edition`
Add `FetchOnReconnect` tag, tagging relevant queries with it. This tag is invalidated in the socketConnected listener, when it is determined that the queue changed.
- Add checks to the "recovery" logic for socket connect events to reduce the number of network requests.
- Remove the `isInitialized` state from `systemSlice` and make it a nanostore local to the socketConnected listener. It didn't need to be global state. It's also now more clearly named `isFirstConnection`.
- Export the queue status selector (minor improvement, memoizes it correctly).
- Fixed a bug where after you load more, changing boards doesn't work. The offset and limit for the list image query had some wonky logic, now resolved.
- Addressed major lag in gallery when selecting an image.
Both issues were related to the useMultiselect and useGalleryImages hooks, which caused every image in the gallery to re-render on whenever the selection changed. There's no way to memoize away this - we need to know when the selection changes. This is a longstanding issue.
The selection is only used in a callback, though - the onClick handler for an image to select it (or add it to the existing selection). We don't really need the reactivity for a callback, so we don't need to listen for changes to the selection.
The logic to handle multiple selection is moved to a new `galleryImageClicked` listener, which does all the selection right when it is needed.
The result is that gallery images no long need to do heavy re-renders on any selection change.
Besides the multiselect click handler, there was also inefficient use of DND payloads. Previously, the `IMAGE_DTOS` type had a payload of image DTO objects. This was only used to drag gallery selection into a board. There is no need to hold onto image DTOs when we have the selection state already in redux. We were recalculating this payload for every image, on every tick.
This payload is now just the board id (the only piece of information we need for this particular DND event).
- I also removed some unused DND types while making this change.
There was a lot of convoluted, janky logic related to trying to not mount the context menu's portal until its needed. This was in the library where the component was originally copied from.
I've removed that and resolved the jank, at the cost of there being an extra portal for each instance of the context menu. Don't think this is going to be an issue. If it is, the whole context menu could be refactored to be a singleton.
