Commit Graph

11493 Commits

Author SHA1 Message Date
psychedelicious
242b2a0b59 tests: clean up tests after events changes 2024-05-20 16:22:37 +10:00
psychedelicious
86e201612f fix(mm): port changes into new model_install_common file
Some subtle changes happened between this PR's last update and now. Bring them into the file.
2024-05-20 16:12:45 +10:00
psychedelicious
f6d1e1be22 fix(ui): update event handling to match new types 2024-05-20 15:52:28 +10:00
psychedelicious
edf043d1d6 chore(ui): typegen 2024-05-20 15:37:25 +10:00
psychedelicious
70a21eda78 fix(events): remove user_id and project_id from error event 2024-05-20 15:34:30 +10:00
psychedelicious
452f4fe0e6 feat(events): add extra field to event payloads
This allows for arbitrary serializable data to be sent with events.
2024-05-20 15:34:03 +10:00
psychedelicious
87f2d04ddd fix(events): fix session processor event handling 2024-05-20 15:22:10 +10:00
psychedelicious
9c45cbe8f7 tidy(ui): remove old unused session subscribe actions 2024-05-20 15:19:17 +10:00
psychedelicious
5f7c852493 docs: clarify comment in api_app 2024-05-20 15:19:17 +10:00
psychedelicious
35ef02bdf7 fix(ui): denoise percentage 2024-05-20 15:19:17 +10:00
psychedelicious
7da283f433 chore(ui): typegen 2024-05-20 15:19:17 +10:00
psychedelicious
812cf277b8 feat(api): sort socket event names for openapi schema
Deterministic ordering prevents extraneous, non-functional changes to the autogenerated types
2024-05-20 15:19:08 +10:00
psychedelicious
182cb51bf0 fix(events): fix denoise progress percentage
- Restore calculation of step percentage but in the backend instead of client
- Simplify signatures for denoise progress event callbacks
- Clean up `step_callback.py` (types, do not recreate constant matrix on every step, formatting)
2024-05-20 15:19:08 +10:00
psychedelicious
64a3adfc64 chore(ui): typegen 2024-05-20 15:19:08 +10:00
psychedelicious
a48ef9f7a7 feat(events): remove payload registry, add method to get event classes
We don't need to use the payload schema registry. All our events are dispatched as pydantic models, which are already validated on instantiation.

We do want to add all events to the OpenAPI schema, and we referred to the payload schema registry for this. To get all events, add a simple helper to EventBase. This is functionally identical to using the schema registry.
2024-05-20 15:18:58 +10:00
psychedelicious
9aeabf10df docs: tidy comments in processor 2024-05-20 15:18:58 +10:00
psychedelicious
7b93cc8538 feat(ui): add missing socket events 2024-05-20 15:18:58 +10:00
psychedelicious
a2480c16e7 feat(events): add missing events
These events weren't being emitted via socket.io:
- DownloadCancelledEvent
- DownloadCompleteEvent
- DownloadErrorEvent
- DownloadProgressEvent
- DownloadStartedEvent
- ModelInstallDownloadsCompleteEvent
2024-05-20 15:18:58 +10:00
psychedelicious
b1e2dd222e feat(events): use builder pattern for download events 2024-05-20 15:18:58 +10:00
psychedelicious
1f92e9eec2 fix(events): dump events with mode="json"
Ensures all model events are serializable.
2024-05-20 15:18:58 +10:00
psychedelicious
fb402f3b46 chore: ruff 2024-05-20 15:18:57 +10:00
psychedelicious
0abc328ddf docs(events): update event docstrings 2024-05-20 15:18:57 +10:00
psychedelicious
cfa4e5f88e tests: move fixtures import to conftest.py 2024-05-20 15:18:57 +10:00
psychedelicious
24d0d4932d tests: update tests to use new events 2024-05-20 15:18:57 +10:00
psychedelicious
20db93b901 fix(mm): check for presence of invoker before emitting model load event
The model loader emits events. During testing, it doesn't have access to a fully-mocked events service, so the test fails when attempting to call a nonexistent method. There was a check for this previously, but I accidentally removed it. Restored.
2024-05-20 15:18:57 +10:00
psychedelicious
500a733d79 fix(ui): correct model load event format 2024-05-20 15:18:57 +10:00
psychedelicious
338d5f158b fix(events): add missing __event_name__ to EventBase 2024-05-20 15:18:57 +10:00
psychedelicious
63e4b224b2 feat(events): simplify event classes
- Remove ABCs, they do not work well with pydantic
- Remove the event type classvar - unused
- Remove clever logic to require an event name - we already get validation for this during schema registration.
- Rename event bases to all end in "Base"
2024-05-20 15:18:57 +10:00
psychedelicious
e9043ff060 fix(events): emit bulk download events in correct room 2024-05-20 15:18:57 +10:00
psychedelicious
c725851c64 chore(ui): tidy after rebase 2024-05-20 15:18:57 +10:00
psychedelicious
a1c4ef55d7 feat(ui): update UI to use new events
- Use OpenAPI schema for event payload types
- Update all event listeners
- Add missing events / remove old nonexistent events
2024-05-20 15:18:42 +10:00
psychedelicious
e25b39aca2 chore(ui): typegen 2024-05-20 15:15:41 +10:00
psychedelicious
32a02b3329 refactor(events): use pydantic schemas for events
Our events handling and implementation has a couple pain points:
- Adding or removing data from event payloads requires changes wherever the events are dispatched from.
- We have no type safety for events and need to rely on string matching and dict access when interacting with events.
- Frontend types for socket events must be manually typed. This has caused several bugs.

`fastapi-events` has a neat feature where you can create a pydantic model as an event payload, give it an `__event_name__` attr, and then dispatch the model directly.

This allows us to eliminate a layer of indirection and some unpleasant complexity:
- Event handler callbacks get type hints for their event payloads, and can use `isinstance` on them if needed.
- Event payload construction is now the responsibility of the event itself (a pydantic model), not the service. Every event model has a `build` class method, encapsulating this logic. The build methods are provided as few args as possible. For example, `InvocationStartedEvent.build()` gets the invocation instance and queue item, and can choose the data it wants to include in the event payload.
- Frontend event types may be autogenerated from the OpenAPI schema. We use the payload registry feature of `fastapi-events` to collect all payload models into one place, making it trivial to keep our schema and frontend types in sync.

This commit moves the backend over to this improved event handling setup.
2024-05-20 15:15:21 +10:00
psychedelicious
620ee2875e fix(ui): store hidden state of edges in workflows
This prevents a minor visual bug where collapsed edges between collapsed nodes didn't display correctly on first load of a workflow.
2024-05-20 11:36:47 +10:00
psychedelicious
5553588147 fix(ui): ensure invocation edges have a type 2024-05-20 11:36:47 +10:00
psychedelicious
1c29b3bd85 feat(ui): updated field type translations 2024-05-20 11:28:33 +10:00
psychedelicious
e88b807a13 docs(ui): update field type docs & comments 2024-05-20 11:28:33 +10:00
psychedelicious
9e55ef3d4b fix(ui): workflow migration field type
At some point, I made a mistake and imported the wrong types to some files for the old v1 and v2 workflow schema migration data.

The relevant zod schemas and inferred types have been restored.

This change doesn't alter runtime behaviour. Only type annotations.
2024-05-20 11:28:33 +10:00
psychedelicious
8062a47d16 fix(ui): use new field type cardinality throughout app
Update business logic and tests.
2024-05-20 11:28:33 +10:00
psychedelicious
dba8c43ecb feat(ui): explicit field type cardinality
Replace the `isCollection` and `isCollectionOrScalar` flags with a single enum value `cardinality`. Valid values are `SINGLE`, `COLLECTION` and `SINGLE_OR_COLLECTION`.

Why:
- The two flags were mutually exclusive, but this wasn't enforce. You could create a field type that had both `isCollection` and `isCollectionOrScalar` set to true, whuch makes no sense.
- There was no explicit declaration for scalar/single types.
- Checking if a type had only a single flag was tedious.

Thanks to a change a couple months back in which the workflows schema was revised, field types are internal implementation details. Changes to them are non-breaking.
2024-05-20 11:28:33 +10:00
psychedelicious
8ebf2ddf15 fix(ui): fix t2i adapter dimensions error message
It now indicates the correct dimension of 64 (SD1.5) or 32 (SDXL) - before was hardcoded to 64.
2024-05-20 11:23:14 +10:00
psychedelicious
f4625c2671 feat(ui): add canvas objects to metadat a for all canvas graphs 2024-05-20 10:32:59 +10:00
psychedelicious
c94742bde6 feat(ui): add canvas objects to metadata when saving canvas to gallery 2024-05-20 10:32:59 +10:00
psychedelicious
a34faf0bd8 chore(ui): typegen 2024-05-20 10:32:59 +10:00
psychedelicious
ecfff6cb1e feat(api): add metadata to upload route
Canvas images are saved by uploading a blob generated from the HTML canvas element. This means the existing metadata handling, inside the graph execution engine, is not available.

To save metadata to canvas images, we need to provide it when uploading that blob.

The upload route now has a `metadata` body param. If this is provided, we use it over any metadata embedded in the image.
2024-05-20 10:32:59 +10:00
psychedelicious
ba8bed6870 fix(ui): edge case resulting in no node templates when loading workflow, causing failure
Depending on the user behaviour and network conditions, it's possible that we could try to load a workflow before the invocation templates are available.

Fix is simple:
- Use the RTKQ query hook for openAPI schema in App.tsx
- Disable the load workflow buttons until w have templates parsed
2024-05-19 07:34:00 -07:00
psychedelicious
ca186bca61 fix(ui): missed node execution state for progress images 2024-05-19 20:14:01 +10:00
psychedelicious
e2f109807c fix(ui): delete edges when their source or target no longer exists 2024-05-19 20:14:01 +10:00
psychedelicious
281bd31db2 feat(nodes): make ModelIdentifierInvocation a prototype 2024-05-19 20:14:01 +10:00
psychedelicious
cea1874e00 perf(ui): memoize WorkflowName selectors 2024-05-19 20:14:01 +10:00