Workflow Code Generation from Function Description Logic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Low-code development tools heavily depend on manual design, lacking intelligent generation and assistant capabilities, which affects development efficiency and fails to meet complex service requirements.
Innovation Solution
A code development platform that intelligently generates workflow code by mining and analyzing function description text to identify logical relationships between call statements, reducing developer dependency and automating trigger/action parameter filling, while supporting complex service logic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual design is used for workflow development, then developers have full control over the design process, but development efficiency is low and time costs are high
Solution Approach 1:
The system enables self-service by automatically generating workflow code from natural language descriptions. The code generation platform autonomously parses user input, identifies logical relationships between calls, and produces executable workflow code without requiring manual programming, thereby dramatically improving development efficiency and reducing time costs.
Solution Approach 2:
The patent introduces an intermediary natural language interface between the user and the code generation system. Users describe workflow requirements in natural language, and the platform translates these descriptions into structured code with automatic parameter filling, serving as a mediator that bridges human intent and machine execution while reducing direct manual coding requirements.
2Extent of automation
If low-code development tools are used, then development speed is improved, but intelligent generation capabilities are lacking and dependency on manual design remains high
Solution Approach 1:
The patent replaces manual mechanical coding processes with intelligent automated generation. The system uses natural language processing and code generation algorithms to automatically create workflow code, substituting the mechanical act of manual programming with intelligent automated systems that can understand and translate human language into executable code structures.
Solution Approach 2:
The system performs preliminary action by pre-defining code templates and parameter structures for common workflow patterns. When generating code, it automatically fills in parameters based on the natural language description and pre-established templates, reducing the need for manual configuration and accelerating the development process while maintaining system manageability.
3Productivity
If automatic code generation is implemented, then development time is reduced, but handling complex service logic such as multi-branch parallel and conditional workflows becomes challenging
Solution Approach 1:
The patent applies segmentation by breaking down complex workflow logic into discrete call statements and logical relationship components. The system separately identifies triggers, actions, parameters, and logical connections, then systematically reassembles them into complete workflow code, ensuring that even complex multi-branch and conditional logic is accurately generated through structured decomposition.
Solution Approach 2:
The system incorporates feedback mechanisms to verify generated code against the original natural language requirements. It checks whether the generated workflow logic accurately reflects the intended service requirements, particularly for complex scenarios involving parallel processing and conditional branches, allowing for correction and refinement to ensure reliability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
This application provides a code development method, applied to a code development platform, and including: receiving function description information that is of a target workflow and that is input by a user; obtaining, based on the function description information, a function description text including call statements, where the call statements include a trigger call statement used to start the target workflow and an action interface call statement used to execute the target workflow; then mining and analyzing the function description text to obtain a logical relationship between the call statements; and generating code of the target workflow based on the logical relationship between the call statements, and rendering a view of the target workflow based on the code of the target workflow. This method provides a capability of intelligently generating code of a workflow, reduces dependency on manual design of a developer, improves development efficiency, and can meet a service requirement.