tt/ttstd-infinite-canvas
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(workflow): 工作流提供方 - 集成 RunningHub 与 ComfyUI Bridge

Browse Source
This commit is contained in:
yepwait
2026-08-26 15:02:24 +08:00
committed by ddCat
parent 6141bd688c
commit 86fc3fda2f
82 changed files with 11405 additions and 248 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
web/src/pages/canvas/use-canvas-generation-executor.ts
View File

@@ -136,17 +136,18 @@ export function useCanvasGenerationExecutor({
let rawGenerationContext: Awaited<ReturnType<typeof hydrateNodeGenerationContext>>;
// 视频文本只保留输入框内容；连接的媒体仍作为结构化参考传递。
const promptOnly = mode === "video";
const usesWorkflowProvider = Boolean(mode !== "text" && generationConfig.taskWorkflowProvider && generationConfig.taskWorkflowProvider !== "model");
try {
const baseContext = buildNodeGenerationContext(nodeId, nodesRef.current, connectionsRef.current, editingTextNode ? `请根据要求修改以下文本。\n\n原文：\n${sourceTextContent}\n\n修改要求：\n${prompt}` : generationPrompt, assets, promptOnly);
const requirements = generationModelRequirements(mode, baseContext, sourceNode, generationConfig, true);
generationConfig = buildGenerationConfig(effectiveConfig, sourceNode, mode, requirements);
const compatibilityError = modelCompatibilityError(generationConfig, generationConfig.model, requirements);
const compatibilityError = usesWorkflowProvider ? "" : modelCompatibilityError(generationConfig, generationConfig.model, requirements);
if (compatibilityError) throw new Error(`当前模型无法支持这组输入和参数：${compatibilityError}`);
const referenceLimits = modelGroupReferenceLimits(effectiveConfig, generationConfig.model, mode, requirements);
const referenceLimits = usesWorkflowProvider ? undefined : modelGroupReferenceLimits(effectiveConfig, generationConfig.model, mode, requirements);
rawGenerationContext = await hydrateNodeGenerationContext(baseContext, projectId, domainProjectId, mode, mode === "video" && Boolean(referenceLimits?.maxAudios), !promptOnly, referenceLimits);
const hydratedRequirements = generationModelRequirements(mode, rawGenerationContext, sourceNode, generationConfig);
generationConfig = buildGenerationConfig(effectiveConfig, sourceNode, mode, hydratedRequirements);
const hydratedCompatibilityError = modelCompatibilityError(generationConfig, generationConfig.model, hydratedRequirements);
const hydratedCompatibilityError = usesWorkflowProvider ? "" : modelCompatibilityError(generationConfig, generationConfig.model, hydratedRequirements);
if (hydratedCompatibilityError) throw new Error(`当前模型无法支持这组输入和参数：${hydratedCompatibilityError}`);
} catch (error) {
const errorDetails = generationErrorMessage(error);
@@ -337,7 +338,7 @@ function generationModelRequirements(
mode: CanvasNodeGenerationMode,
input: Pick<Awaited<ReturnType<typeof hydrateNodeGenerationContext>>, "textCount" | "imageCount" | "videoCount" | "audioCount" | "characterReferences">,
sourceNode: CanvasNodeData | undefined,
config: ReturnType<typeof useEffectiveConfig>,
config: ReturnType<typeof useEffectiveConfig>,
includeCharacterMinimum = false,
): ModelRequirements {
return {
@@ -350,7 +351,7 @@ function generationModelRequirements(
characterCount: includeCharacterMinimum ? input.characterReferences.length : 0,
},
videoOperation: sourceNode?.metadata?.videoEditOperation,
videoSeconds: config.videoSeconds,
options: modelRequestOptions(config, mode),
videoSeconds: config.videoSeconds,
options: config.taskWorkflowProvider === "model" ? modelRequestOptions(config, mode) : undefined,
};
}