Workflow conversion system and method for converting on-premise nodes to cloud nodes

Abstract

The present disclosure relates to a workflow conversion system and method for converting native nodes into cloud nodes. The conversion system comprises: a native workflow acquisition component, which acquires a native workflow to be used by a user based on a request of the user; a native node traversal component, which traverses and analyzes native nodes in the acquired native workflow, determines a native node without independence as a non-migratable native node by determining independence of the native node on local attributes, and determines a native node with independence as a migratable native node; a cloud node conversion component, which converts the migratable native node into a cloud node, wherein the converted cloud node retains its own node upstream and downstream relationships; and a workflow regeneration component, which combines the converted cloud node and native nodes in the native workflow that have not been converted into a new workflow based on the upstream and downstream relationships, so that the new workflow contains a local part deployed on a local device and a cloud part deployed in the cloud.

Description

Technical Field

[0001] This disclosure relates to the field of computer information processing, and more specifically, to a workflow conversion system and method for converting native nodes into cloud nodes. Background Technology

[0002] With the rapid development of artificial intelligence, models for generating images or videos from text are becoming increasingly common. However, due to insufficient local computing power, these models often run very slowly, resulting in a poor user experience. Therefore, how to utilize network computing power and public computing resources to quickly run models and obtain results has become a common user need.

[0003] ComfyUI, a commonly used workflow tool, lowers the barrier to entry for AI image generation through its graphical interface. However, while regular users may know that public computing power can be used to accelerate AI image generation, they may not easily understand which ComfyUI native nodes can be converted into cloud nodes and deployed in the cloud. Furthermore, some cloud node development companies have deployed numerous cloud nodes as plugins to accelerate AI image generation, but ordinary users are unaware of the correspondence between these native nodes and cloud nodes, making direct application difficult. Additionally, in some specific workflows, multiple native nodes can actually be deployed simultaneously on a single cloud node. If deployed as multiple cloud nodes, since cloud nodes operate via network calls, multiple cloud nodes may generate multiple network calls. This results in multiple separate calls to the same network computing power during the workflow's execution, even after conversion to cloud nodes, leading to decreased network call efficiency. Therefore, overall, both conventional workflow tools and their corresponding plugins are not user-friendly for ordinary users.

[0004] Therefore, there is a need for a workflow conversion system and method to convert native nodes into cloud nodes, enabling ordinary users to directly and clearly understand which native nodes can be converted into cloud nodes and to easily select cloud nodes.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] In view of this, this disclosure provides a workflow conversion system that enables ordinary users to easily achieve the AI ​​image generation process by directly using system conversion plugins of some workflow tools.

[0007] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part from practice of this disclosure.

[0008] According to one aspect of this disclosure, a workflow conversion system for converting native nodes into cloud nodes is proposed, comprising: a native workflow acquisition component, which acquires the native workflow to be used by the user based on the user's request; a native node traversal component, which traverses and analyzes the native nodes in the acquired native workflow, and determines native nodes that do not have independence from local attributes as non-migratable native nodes, and determines native nodes that have independence as migratable native nodes; a cloud node conversion component, which converts the migratable native nodes into cloud nodes, wherein the converted cloud nodes retain their original upstream and downstream relationships; and a workflow regeneration component, which combines the converted cloud nodes with the unconverted native nodes in the native workflow based on upstream and downstream relationships to form a new workflow, thereby making the new workflow include a local part deployed on a local device and a cloud part deployed in the cloud.

[0009] The workflow conversion system for converting native nodes to cloud nodes according to this disclosure also includes: a workflow visualization component, which is used to connect nodes in workflow order after the native node traversal component has traversed and analyzed the native nodes in the acquired native workflow, configure attributes for each node, and display the portability of the native nodes.

[0010] According to the workflow conversion system for converting native nodes to cloud nodes disclosed herein, the native nodes that do not have independence include: native nodes that rely on local dedicated hardware, native nodes that rely on local file paths, native nodes that are directly connected to intranet databases, native nodes that interact with legacy systems without open APIs, native nodes that communicate via shared memory, native nodes that perform file operations on fixed paths, and native nodes that call local services without containerization.

[0011] The workflow conversion system for converting native nodes to cloud nodes according to this disclosure further includes: a node fusion component, which is used to determine consecutively adjacent migrateable native nodes as fusionable nodes after the native node traversal component traverses and analyzes the native nodes in the obtained native workflow, so that the cloud node conversion component can select multiple adjacent fusionable nodes to convert into one cloud node when converting migrateable native nodes into cloud nodes.

[0012] The workflow conversion system for converting native nodes to cloud nodes according to this disclosure also includes a cloud node conversion guidance component. On the visualized workflow, which displays the migrateable native nodes, there is a drop-down menu that displays selectable cloud nodes. The selectable cloud nodes point to cloud nodes in the cloud node library corresponding to the migrateable native nodes, so that users can directly select cloud nodes to convert native nodes to cloud nodes.

[0013] According to another aspect of this disclosure, a workflow conversion method for converting native nodes into cloud nodes is also provided, comprising: obtaining the native workflow to be used by the user based on the user's request through a native workflow acquisition component; traversing and analyzing the native nodes in the acquired native workflow through a native node traversal component, determining native nodes that do not have independence from local attributes as non-migratable native nodes, and determining native nodes that have independence as migratable native nodes through a cloud node conversion component; converting the migratable native nodes into cloud nodes through a cloud node conversion component, wherein the converted cloud nodes retain their own upstream and downstream relationships as native nodes; and combining the converted cloud nodes with the unconverted native nodes in the native workflow based on upstream and downstream relationships through a workflow regeneration component to form a new workflow, thereby making the new workflow include a local part deployed on a local device and a cloud part deployed in the cloud.

[0014] The workflow conversion method for converting native nodes to cloud nodes according to this disclosure further includes: after traversing and analyzing the native nodes in the obtained native workflow through the native node traversal component, connecting the nodes in the workflow order through the workflow visualization component, configuring attributes for each node, and displaying the portability of the native nodes.

[0015] According to the workflow conversion method for converting native nodes to cloud nodes disclosed herein, the native nodes that do not have independence include: native nodes that depend on local dedicated hardware, native nodes that depend on local file paths, native nodes that are directly connected to intranet databases, native nodes that interact with legacy systems without open APIs, native nodes that communicate via shared memory, native nodes that perform file operations on fixed paths, and native nodes that call local services without containerization.

[0016] The workflow conversion method for converting native nodes to cloud nodes disclosed herein further includes: a node fusion component, used to determine consecutively adjacent migrateable native nodes as fusionable nodes after the native node traversal component traverses and analyzes the native nodes in the obtained native workflow, so that the cloud node conversion component can select multiple adjacent fusionable nodes to convert into one cloud node when converting migrateable native nodes into cloud nodes.

[0017] The workflow conversion method for converting native nodes to cloud nodes according to this disclosure further includes: displaying selectable cloud nodes in a drop-down menu of the visualized workflow as transferable native nodes through a cloud node conversion guidance component, wherein the selectable cloud nodes point to cloud nodes in the cloud node library corresponding to the transferable native nodes, so that users can directly select cloud nodes to convert native nodes to cloud nodes.

[0018] According to the workflow conversion system and method disclosed herein, by determining the independence of native nodes, migrateable native nodes are displayed. Through the cloud node library provided by this system for various native nodes that can be converted into cloud nodes, users are provided with selectable cloud nodes. Furthermore, based on the adjacency and connection relationships between native nodes, users are automatically provided with fusion guidance. This provides a user-friendly workflow conversion interface for ordinary users, thereby greatly facilitating users to improve the speed and efficiency of AI generation and processing by leveraging network or cloud computing power based on commonly used workflow tools, thus bringing practical convenience to users' smart lives.

[0019] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this disclosure. Attached Figure Description

[0020] The above and other objects, features, and advantages of this disclosure will become more apparent from the detailed description of exemplary embodiments with reference to the accompanying drawings. The drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0021] Figure 1 This is a block diagram illustrating a first embodiment of a workflow conversion system for converting native nodes into cloud nodes according to an exemplary embodiment of the present disclosure.

[0022] Figure 2 This is a block diagram illustrating a second embodiment of a workflow conversion system for converting native nodes to cloud nodes according to an exemplary embodiment of the present disclosure.

[0023] Figure 3 This is a block diagram illustrating a third embodiment of a workflow conversion system for converting native nodes into cloud nodes according to an exemplary embodiment of the present disclosure.

[0024] Figure 4 This is a block diagram illustrating a fourth embodiment of a workflow conversion system for converting native nodes to cloud nodes according to an exemplary embodiment of the present disclosure.

[0025] Figure 5 This is a flowchart illustrating a workflow conversion method for converting a native node into a cloud node according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0026] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, they are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted.

[0027] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0028] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0029] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0030] It should be understood that although the terms first, second, third, etc., may be used herein to describe various components, these components should not be limited by these terms. These terms are used to distinguish one component from another. Therefore, the first computing device discussed below may be referred to as the second computing device without departing from the teachings of this disclosure. As used herein, the term "and / or" includes all combinations of any and more of the associated listed items.

[0031] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of exemplary embodiments, and the modules or processes in the drawings are not necessarily necessary for implementing this disclosure, and therefore cannot be used to limit the scope of protection of this disclosure.

[0032] Figure 1 This is a block diagram illustrating an embodiment of a workflow conversion system 100 for converting native nodes to cloud nodes according to this disclosure, based on an exemplary embodiment. Figure 1As shown, the workflow conversion system 100 includes: a native workflow acquisition component 110, a native node traversal component 120, a cloud node conversion component 130, and a workflow regeneration component 140. In general, the workflow conversion system 100, which converts native nodes to cloud nodes, involves installing a conversion system or plugin on a workflow tool such as ComfyUI. This replaces some nodes in the native workflow with corresponding cloud nodes, thus achieving the "cloudification" of computing resources while maintaining the workflow's functionality.

[0033] Specifically, the native workflow acquisition component 110 retrieves the native workflow the user wants to use based on the user's request. Subsequently, the native node traversal component 120 traverses and analyzes the native nodes in the acquired native workflow. By determining the independence of native nodes from local attributes, native nodes lacking independence are identified as non-migratable, while those possessing independence are identified as migratable. Non-independent native nodes include: native nodes dependent on local dedicated hardware, native nodes dependent on local file paths, native nodes directly connected to intranet databases, native nodes interacting with legacy systems without open APIs, native nodes using shared memory communication, native nodes performing fixed-path file operations, and native nodes calling local services without containerization. Native nodes dependent on local dedicated hardware include industrial control equipment drivers, local dongles, or nodes with specific physical interfaces. Nodes using the registry or non-virtualized local services are also non-migratable. Other non-migratable nodes include real-time control nodes with millisecond-level responses, such as high-frequency trading engines; and nodes deeply coupled with the local system, such as nodes using shared memory communication.

[0034] Taking a ComfyUI workflow for generating images from text as an example, a native workflow contains 7 native ComfyUI nodes. The entire workflow includes: a LOADCHECKPOINT node for loading pre-trained model weights and configurations to ensure that the model has initialized parameters when generating images; and two CLIP Text Encode nodes for encoding text prompts into vectors to align with image features. It utilizes the CLIP (Contrastive Language-Image Pre-training) model to convert text into the model's input format; the Empty Latent node generates an empty latent space vector, which is typically used as the initial latent representation for generating images; the KSampler node samples from the latent space to generate multiple latent vectors, which are used to generate diverse images; the Variational Autoencoder (VAE) node decodes the latent space vector into an image; the VAE encodes the image into a latent space representation through an encoder, and then restores it to an image through a decoder; and the Save Image node saves the generated image to a specified path or format for subsequent processing or display. These nodes together constitute a complete workflow for generating images from text, realizing the transformation from text description to image generation. Through traversal analysis, the Save Image and Empty Latent nodes are non-transferable native nodes, while the other nodes are cloud-compatible.

[0035] Next, the cloud node conversion component 130 converts the migrateable native nodes into cloud nodes, where the converted cloud nodes retain their original upstream and downstream relationships. Optionally, the cloud node conversion component 130 can directly convert all migrateable native nodes into cloud nodes. For example, if converting native ComfyUI workflow JSON format nodes into JSON format cloud nodes, the first step is usually to understand the specific structure and function of the native ComfyUI workflow JSON nodes. This includes the node's input ports, output ports, parameter settings, and the logical processing relationships between them. Then, mapping to the cloud node structure: based on the function of the native JSON nodes, design the corresponding cloud node structure. Determine the required input parameters, output data format, and logical processing flow of the cloud nodes, ensuring they are consistent with the native nodes. Write the cloud node JSON description: based on the designed cloud node structure, write a new JSON format description file. This file should contain the cloud node's metadata (such as node ID, type, etc.), the definition of input and output ports, parameter settings, and their processing logic. These are mostly conventional technical methods and will not be elaborated further here. Specialized cloud node service providers pre-generate a database of cloud nodes, including common native nodes, as plugins for users to choose from. When providing this database, the cloud nodes, after being written, are imported into the target cloud platform or tool for testing. This verifies whether the cloud nodes can correctly receive input data, execute logical processing, and output expected results. These pre-written cloud nodes are typically optimized and tuned to improve performance, enhance error handling, or adjust parameter settings for optimal performance. Detailed documentation is written for these cloud nodes, recording their functions, usage methods, parameter descriptions, and any precautions. The optimized cloud nodes are deployed to the cloud platform or related systems to ensure users can easily access and use them. The specific writing methods are not detailed here; conventional methods can be used. Therefore, most cloud node conversion processes in current cloud node platforms are completed by selecting existing cloud nodes.

[0036] Finally, the workflow regeneration component 140 combines the transformed cloud nodes with the untransformed native nodes in the original workflow based on upstream and downstream relationships to form a new workflow. This results in a new workflow containing both a local portion deployed on the local device and a cloud portion deployed in the cloud. In this way, by replacing some nodes in a native workflow such as ComfyUI with corresponding cloud nodes, the "cloudification" of computing resources is completed while maintaining the workflow's functionality. In other words, the native workflow is converted into a workflow that can use cloud nodes on a cloud platform, while preserving its functionality.

[0037] Figure 2This is a block diagram illustrating a second embodiment of a workflow conversion system for converting native nodes to cloud nodes according to an exemplary embodiment of the present disclosure. Figure 1 compared to, Figure 2 The only addition is a workflow visualization component 150. In some cases, users need to manually select which native nodes should be converted to cloud nodes based on their preferences, but they lack the ability to read the code. Therefore, to facilitate user operation, this disclosed workflow conversion system for converting native nodes to cloud nodes also includes a workflow visualization component 150. This component connects the nodes in the obtained native workflow after the native node traversal component 120 traverses and analyzes them, configures attributes for each node, and displays the portability of the native nodes. This allows users to manually select the native nodes that can be migrated. Typically, for example, using ComfyUI, the workflow visualization component 150 accesses the node library based on the native workflow and drags and drops nodes from the node library corresponding to the native nodes of the workflow onto the canvas: select the required nodes from the node library (such as LOADCHECKPOINT, CLIP Text Encode, KSampler), drag the required nodes onto the canvas, and connect the node logic, that is, connect the nodes in the workflow order (e.g., text input → CLIP encoding → empty latent space → KSampler → VAE decoding → save image), ensuring the correct data flow direction (e.g., the output of KSampler is connected to the input of VAE Decode). The connections between nodes represent the data transfer path and are the core of workflow visualization. Then, configure the node parameters. After forming the basic workflow graph, the user can double-click a node in the canvas to open the parameter panel and set the properties according to their needs. For example, LOADCHECKPOINT: select the cloud storage or local model file path. CLIP Text Encode: input prompt words to define the generated content. KSampler: Adjusts generation parameters such as sampling steps and seed. Save Image: Sets the output path (e.g., cloud storage bucket s3: / / output / or local folder). This allows for pre-running and debugging after generating the visual workflow, providing real-time progress tracking, logs, and error messages, facilitating troubleshooting of node connections or parameter issues. Once the graphical interface is complete, the current workflow can be exported as a JSON file (containing all node configurations and connection relationships).

[0038] Figure 3 This is a block diagram illustrating a third embodiment of a workflow conversion system for converting native nodes to cloud nodes according to an exemplary embodiment of the present disclosure. Figure 2 compared to, Figure 3The only addition is a cloud node conversion guidance component 160. It should be noted that selecting cloud nodes is very difficult for ordinary users when they are making their own conversions. Therefore, the workflow conversion system for converting native nodes to cloud nodes disclosed herein also includes a cloud node conversion guidance component 160. On the visualized workflow displaying the migrateable native nodes, a drop-down menu is set to display selectable cloud nodes. These selectable cloud nodes point to cloud nodes in the cloud node library corresponding to the migrateable native nodes, allowing users to directly select a cloud node for the native node to cloud node conversion. In this way, users do not need knowledge or experience of the correspondence between native nodes and the cloud nodes to be converted. Thus, the cloud node conversion component 130, guided by the cloud node conversion guidance component 160, can easily complete the cloud node conversion by simply selecting the corresponding cloud node in the cloud node library, greatly simplifying and making AI image generation much easier for users. Typically, when selecting a cloud node in the cloud node library corresponding to the native node to be converted, the cloud node conversion component 130 will attach the upstream and downstream links of the native node to the selected cloud node.

[0039] Figure 4 This is a block diagram illustrating a fourth embodiment of a workflow conversion system for converting native nodes to cloud nodes according to an exemplary embodiment of the present disclosure. Figure 3 In comparison, only the cloud node fusion component 170 is added. It should be noted that cloud nodes are completed via network calls, and multiple cloud nodes may generate multiple network calls. However, for some serialized cloud nodes, their network calls can be merged and completed through another "fused" cloud node. Selecting a "fused" cloud node to replace multiple serialized cloud nodes requires considerable experience and is not user-friendly for ordinary users. Therefore, further, the workflow conversion system for converting native nodes to cloud nodes disclosed in this invention also includes the cloud node fusion component 170. This component is used to determine consecutively adjacent migrateable native nodes as fusionable nodes after the native node traversal component 120 traverses and analyzes the native nodes in the obtained native workflow. This allows the cloud node conversion component 130 to select multiple adjacent fusionable nodes to convert into a single cloud node when converting migrateable native nodes to cloud nodes. Alternatively, the cloud node fusion component 170 can also traverse and analyze the new workflow after the workflow regeneration component 140 generates the new workflow, obtain the adjacent connection relationship between cloud nodes deployed on the cloud platform in the new workflow, and fuse the cloud nodes with adjacent connection relationship to obtain the optimized new workflow again.

[0040] Figure 5 This is a flowchart illustrating a workflow conversion method for converting a native node to a cloud node according to an exemplary embodiment of the present disclosure. Figure 5As shown, in step S210, the native workflow acquisition component 110 acquires the native workflow to be used by the user based on the user's request. Subsequently, in step S220, the native node traversal component 120 traverses and analyzes the native nodes in the acquired native workflow. By determining the independence of native nodes from local attributes, native nodes that do not have independence are identified as non-migratable native nodes, and native nodes that have independence are identified as migratable native nodes. Native nodes that do not have independence include: native nodes that depend on local dedicated hardware, native nodes that depend on local file paths, native nodes that are directly connected to intranet databases, native nodes that interact with legacy systems without open APIs, native nodes that communicate via shared memory, native nodes that perform file operations on fixed paths, and native nodes that call local services without containerization. Native nodes that depend on local dedicated hardware include industrial control equipment drivers, local dongles, or nodes with specific physical interfaces. Nodes that are registry entries or non-virtualized local services are also non-migratable nodes. There are also real-time control nodes with millisecond-level responses, such as high-frequency trading engines; and nodes that are deeply coupled with the local system, such as nodes that communicate via shared memory.

[0041] Taking a ComfyUI workflow for generating images from text as an example, a native workflow contains 7 native ComfyUI nodes. The entire workflow includes: a LOADCHECKPOINT node for loading pre-trained model weights and configurations to ensure that the model has initialized parameters when generating images; and two CLIP Text Encode nodes for encoding text prompts into vectors to align with image features. It utilizes the CLIP (Contrastive Language-Image Pre-training) model to convert text into the model's input format; the Empty Latent node generates an empty latent space vector, which is typically used as the initial latent representation for generating images; the KSampler node samples from the latent space to generate multiple latent vectors, which are used to generate diverse images; the Variational Autoencoder (VAE) node decodes the latent space vector into an image; the VAE encodes the image into a latent space representation through an encoder, and then restores it to an image through a decoder; and the Save Image node saves the generated image to a specified path or format for subsequent processing or display. These nodes together constitute a complete workflow for generating images from text, realizing the transformation from text description to image generation. Through traversal analysis, the Save Image and Empty Latent nodes are non-transferable native nodes, while the other nodes are cloud-compatible.

[0042] Subsequently, in step S230, the cloud node conversion component 130 converts the migrateable native nodes into cloud nodes, where the converted cloud nodes retain their original upstream and downstream relationships. Optionally, the cloud node conversion component 130 can directly convert all migrateable native nodes into cloud nodes. For example, if converting native ComfyUI workflow JSON format nodes into JSON format cloud nodes, typically, firstly, it is necessary to understand the specific structure and function of the native ComfyUI workflow JSON nodes. This includes the node's input ports, output ports, parameter settings, and the logical processing relationships between them. Then, mapping to the cloud node structure: based on the function of the native JSON nodes, design the corresponding cloud node structure. Determine the input parameters, output data format, and logical processing flow required by the cloud nodes, ensuring they are consistent with the native nodes. Write the cloud node JSON description: based on the designed cloud node structure, write a new JSON format description file. This file should contain the cloud node's metadata (such as node ID, type, etc.), the definition of input and output ports, parameter settings, and their processing logic. These are mostly conventional technical methods and will not be elaborated further here. Specialized cloud node service providers pre-generate a database of cloud nodes, including common native nodes, as plugins for users to choose from. When providing this database, the cloud nodes, after being written, are imported into the target cloud platform or tool for testing. This verifies whether the cloud nodes can correctly receive input data, execute logical processing, and output expected results. These pre-written cloud nodes are typically optimized and tuned to improve performance, enhance error handling, or adjust parameter settings for optimal performance. Detailed documentation is written for these cloud nodes, recording their functions, usage methods, parameter descriptions, and any precautions. The optimized cloud nodes are deployed to the cloud platform or related systems to ensure users can easily access and use them. The specific writing methods are not detailed here; conventional methods can be used. Therefore, most cloud node conversion processes in current cloud node platforms are completed by selecting existing cloud nodes.

[0043] Finally, the workflow regeneration component 140 combines the transformed cloud nodes with the untransformed native nodes in the original workflow based on upstream and downstream relationships to form a new workflow. This results in a new workflow containing both a local portion deployed on the local device and a cloud portion deployed in the cloud. In this way, by replacing some nodes in a native workflow such as ComfyUI with corresponding cloud nodes, the "cloudification" of computing resources is completed while maintaining the workflow's functionality. In other words, the native workflow is converted into a workflow that can use cloud nodes on a cloud platform, while preserving its functionality.

[0044] However, in some cases, users need to choose which native nodes to convert to cloud nodes based on their preferences, but they lack the ability to read the code. Therefore, to facilitate user operation, the workflow conversion system disclosed herein for converting native nodes to cloud nodes also includes a workflow visualization component 150. This component connects the nodes in the obtained native workflow according to the workflow order after traversing and analyzing the native nodes in step S220, configures attributes for each node, and displays the portability of the native nodes. In this way, users can independently select the native nodes that can be migrated. Typically, for example, using ComfyUI, the workflow visualization component 150 accesses the node library based on the native workflow and drags and drops nodes from the node library corresponding to the native nodes of the workflow onto the canvas: select the required nodes from the node library (such as LOADCHECKPOINT, CLIP Text Encode, KSampler), drag the required nodes onto the canvas, and connect the node logic, that is, connect the nodes in the workflow order (e.g., text input → CLIP encoding → empty latent space → KSampler → VAE decoding → save image), ensuring the correct data flow direction (e.g., the output of KSampler is connected to the input of VAE Decode). The connections between nodes represent the data transfer path and are the core of workflow visualization. Then, configure the node parameters. After forming the basic workflow graph, the user can double-click a node in the canvas to open the parameter panel and set the properties according to their needs. For example, LOADCHECKPOINT: select the cloud storage or local model file path. CLIP Text Encode: input prompt words to define the generated content. KSampler: Adjusts generation parameters such as sampling steps and seed. Save Image: Sets the output path (e.g., cloud storage bucket s3: / / output / or local folder). This allows for pre-running and debugging after generating the visual workflow, providing real-time progress tracking, logs, and error messages, facilitating debugging of node connections or parameter issues. Once the graphical interface is complete, the current workflow can be exported as a JSON file (containing all node configurations and connection relationships).

[0045] It should be noted that selecting cloud nodes is very difficult for ordinary users when making their own conversions. Therefore, the workflow conversion system for converting native nodes to cloud nodes disclosed herein also includes a cloud node conversion guidance component 160. This component sets up a drop-down menu on the visualized workflow displaying the migrateable native nodes, showing selectable cloud nodes. These selectable cloud nodes point to cloud nodes in the cloud node library corresponding to the migrateable native nodes, allowing users to directly select a cloud node for the native-to-cloud node conversion. In this way, users do not need knowledge or experience of the correspondence between native nodes and the cloud nodes to be converted. Thus, the cloud node conversion component 130, guided by the cloud node conversion guidance component 160, can easily complete the cloud node conversion by simply selecting the corresponding cloud node in the cloud node library, greatly simplifying and making AI image generation much easier for users. Typically, when selecting a cloud node in the cloud node library corresponding to the native node to be converted, the cloud node conversion component 130 will attach the upstream and downstream links of the native node to the selected cloud node.

[0046] It should be noted that cloud nodes are completed through network calls. Multiple cloud nodes may generate multiple network calls. However, for some connected cloud nodes, their network calls can be merged and completed through another "fused" cloud node. Selecting a "fused" cloud node to replace multiple connected cloud nodes requires considerable experience and is not user-friendly for ordinary users. Therefore, further, the workflow conversion system for converting native nodes to cloud nodes disclosed herein also includes a node fusion component 170. After traversing and analyzing the native nodes in the acquired native workflow in step S120, it determines consecutively adjacent migrateable native nodes as fusionable nodes, so that when the cloud node conversion component 130 converts migrateable native nodes to cloud nodes, it can select multiple adjacent fusionable nodes to convert into one cloud node. Optionally, the node fusion component 170 can also traverse and analyze the new workflow after generating it in step S240, obtain the adjacent connected relationships between cloud nodes deployed on the cloud platform in the new workflow, and fuse the adjacent connected cloud nodes to obtain an optimized new workflow.

[0047] In summary, the workflow conversion system and method disclosed herein, by determining the independence of native nodes, displays migrateable native nodes, and provides users with selectable cloud nodes through the cloud node library for various native nodes that can be converted to cloud nodes, and automatically provides users with fusion guidance based on the adjacency and connection relationships between native nodes, thus providing ordinary users with a user-friendly workflow conversion interface. This greatly facilitates users in improving the speed and efficiency of AI generation and processing by leveraging network or cloud computing power based on commonly used workflow tools, thereby bringing practical convenience to users' smart lives.

[0048] Those skilled in the art will understand that the above modules can be distributed in the device as described in the embodiments, or they can be modified accordingly and placed in one or more devices that are unique to this embodiment. The modules in the above embodiments can be combined into one module, or they can be further divided into multiple sub-modules.

[0049] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, mobile terminal, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0050] Exemplary embodiments of this disclosure have been specifically shown and described above. It should be understood that this disclosure is not limited to the detailed structures, arrangements, or implementations described herein; rather, this disclosure is intended to cover various modifications and equivalent arrangements contained within the spirit and scope of the appended claims.

Claims

1. A workflow conversion system for converting native nodes to cloud nodes, comprising: a native workflow obtaining component, obtaining a native workflow to be used by a user based on a request of the user; a native node traversing component, traversing and analyzing native nodes in the obtained native workflow, determining a native node without independence as a non-migratable native node by determining independence of the native node to local attributes, and determining a native node with independence as a migratable native node; a cloud node converting component, converting the migratable native nodes to cloud nodes, wherein the converted cloud nodes retain node upstream and downstream relationships of the native nodes themselves; and a workflow regenerating component, combining the converted cloud nodes and native nodes in the native workflow that are not converted into a new workflow based on the upstream and downstream relationships, so that the new workflow contains a local part deployed on a local device and a cloud part deployed in the cloud.

2. The workflow conversion system that converts native nodes to cloud nodes of claim 1, further comprising: a workflow visualizing component, configured to connect nodes in order of a workflow and configure attributes for each node after the native node traversing component traverses and analyzes the native nodes in the obtained native workflow, and display migratability of the native nodes.

3. The workflow conversion system that converts native nodes to cloud nodes of claim 1, wherein, The native nodes without independence include: a native node dependent on local dedicated hardware, a native node dependent on a local file path, a native node directly connected to an intranet database, a native node interacting with a legacy system whose API is not opened, a native node communicating through shared memory, a native node performing file operations on a fixed path, and a native node calling a local service that is not containerized.

4. The workflow conversion system that converts native nodes to cloud nodes of claim 1 or 2, further comprising: a node fusing component, configured to determine consecutive adjacent migratable native nodes as fusible nodes after the native node traversing component traverses and analyzes the native nodes in the obtained native workflow, so that the cloud node converting component selects multiple adjacent fusible nodes to convert into one cloud node when converting the migratable native nodes to cloud nodes.

5. The workflow conversion system for converting native nodes to cloud nodes according to claim 2, further comprising a cloud node conversion guiding component, having a drop-down menu on a display of a migratable native node in the visualized workflow, displaying selectable cloud nodes pointing to cloud nodes in a cloud node library corresponding to the migratable native node, so that a user directly selects a cloud node for conversion of the native node to the cloud node.

6. A workflow conversion method for converting native nodes to cloud nodes, comprising: obtaining a native workflow to be used by a user based on a request of the user by a native workflow obtaining component; traversing and analyzing native nodes in the obtained native workflow by a native node traversing component, determining a native node without independence as a non-migratable native node by determining independence of the native node to local attributes, and determining a native node with independence as a migratable native node; converting the migratable native nodes to cloud nodes by a cloud node converting component, wherein the converted cloud nodes retain node upstream and downstream relationships of the native nodes themselves; and The converted cloud node and the native node not converted in the native workflow are combined into a new workflow by a workflow regeneration component based on upstream and downstream relationships, so that the new workflow contains a local part deployed on a local device and a cloud part deployed in the cloud.

7. The workflow conversion method of converting native nodes to cloud nodes of claim 6, further comprising: After the native node traversal component traverses and analyzes the native nodes in the acquired native workflow, a workflow visualization component connects the nodes in sequence and configures properties for each node, and displays the migratability of the native nodes.

8. The workflow conversion method of converting a native node to a cloud node of claim 6, wherein, The native nodes without independence include native nodes dependent on local special hardware, native nodes dependent on local file paths, native nodes directly connected to intranet databases, native nodes for interaction with legacy systems whose APIs are not open, native nodes for shared memory communication, native nodes for fixed path file operations, and native nodes for local service calls without containerization.

9. The workflow conversion method of converting a native node to a cloud node of claim 6 or 7, further comprising: A node fusion component is configured to determine that continuously adjacent migratable native nodes are fusible nodes after the native node traversal component traverses and analyzes the native nodes in the acquired native workflow, so that the cloud node conversion component selects multiple adjacent fusible nodes to be converted into one cloud node when converting the migratable native nodes into cloud nodes.

10. The workflow conversion method of converting native nodes to cloud nodes of claim 7, further comprising: A cloud node conversion guide component is configured to display selectable cloud nodes in a drop-down menu of the visualized workflow for the migratable native nodes, the selectable cloud nodes pointing to cloud nodes in a cloud node library corresponding to the migratable native nodes, so that a user directly selects a cloud node for conversion of the native node into the cloud node.

Images