Abstract: A program code component module implementing a portion of a program application is received. A trained machine learning model is used to automatically predict to which one among a plurality of program architecture layer classifications the program code component module belongs. An automatic analysis option is selected based on the predicted program architecture layer classification for the program code component module. The selected automatic analysis option is performed on the program code component module.

Abstract: A specification of a program code component module implementing a portion of a program application is received. A trained machine learning model is used to automatically predict to which one among a plurality of program architecture layer classifications the program code component module belongs. An automatic analysis option is selected based on the predicted program architecture layer classification for the program code component module. The selected automatic analysis option is performed on the program code component module.

Abstract: User interactions and states of a program development tool utilized to develop an application are tracked. Input features associated with the tracked user interactions and the states are provided to a trained machine learning model to determine a prediction result associated with whether a user is likely unable to proceed in the development of the application and likely needs assistance. In response to a determination that the prediction result at least meets a threshold, one or more resolution suggestions are automatically provided. The one or more resolution suggestions is at least one of: automatically selected based on at least a portion of the tracked user interactions and states, or automatically selected based at least in part the prediction result.

Abstract: In various embodiments, a process for constrained decoding and ranking of language models for code generation includes receiving a natural language input specifying a desired computer task. The process includes using a machine learning trained converter to convert the natural language input to an output in a computer language, including by, based on a specified grammar for the computer language, limiting eligible options for a token to include in the output in the computer language. The process includes providing the output in the computer language for computer execution.

Abstract: User interactions and states of a program development tool utilized to develop an application are tracked. Input features associated with the tracked user interactions and the states are provided to a trained machine learning model to determine a prediction result associated with whether a user is likely unable to proceed in the development of the application and likely needs assistance. In response to a determination that the prediction result at least meets a threshold, one or more resolution suggestions are automatically provided. The one or more resolution suggestions is at least one of: automatically selected based on at least a portion of the tracked user interactions and states, or automatically selected based at least in part the prediction result.

Abstract: In various embodiments, a process for constrained decoding and ranking of language models for code generation includes receiving a natural language input specifying a desired computer task. The process includes using a machine learning trained converter to convert the natural language input to an output in a computer language, including by, based on a specified grammar for the computer language, limiting eligible options for a token to include in the output in the computer language. The process includes providing the output in the computer language for computer execution.

Abstract: A repository of graph based visual programming language code instances is analyzed. A similar code portion pattern duplicated is detected among a group of graph based visual programming language code instances included in the repository of graph based visual programming language code instances including by using an index and tokenizing one or more graph nodes connected by one or more graph edges included in a flow corresponding to at least one graph based visual programming language code instance in the group of graph based visual programming language code instances. Within a visual representation of at least one of the group of graph based visual programming language code instances, elements belonging to the detected similar code portion pattern are visually indicated.

Abstract: A declarative specification of a search pattern for a graph visual programming language code is received. A repository of graph based visual programming language code instances is analyzed using one or more processors to identify at least a portion of a visual programming language code instance of the repository that matches the search pattern. An indication of at least the portion of the visual programming language code instance of the repository that matches the search pattern is provided.

Abstract: In various embodiments, a process for detecting duplicated code patterns in visual programming language code instances includes analyzing a repository of graph based visual programming language code instances and detecting a similar code portion pattern duplicated among a group of graph based visual programming language code instances included in the repository of graph based visual programming language code instances including by using an index and tokenizing a flow corresponding to at least one graph based visual programming language code instance in the group of graph based visual programming language code instance. The process includes visually indicating elements belonging to the detected similar code portion pattern within a visual representation of at least one of the group of graph based visual programming language code instances.

Abstract: In various embodiments, a process for detecting duplicated code patterns in visual programming language code instances includes analyzing a repository of graph based visual programming language code instances and detecting a similar code portion pattern duplicated among a group of graph based visual programming language code instances included in the repository of graph based visual programming language code instances including by using an index and tokenizing a flow corresponding to at least one graph based visual programming language code instance in the group of graph based visual programming language code instance. The process includes visually indicating elements belonging to the detected similar code portion pattern within a visual representation of at least one of the group of graph based visual programming language code instances.