GENERATION AND CONSUMPTION OF TEMPLATES BASED ON SOURCE CODE EXAMPLES

    公开(公告)号:US20230289151A1

    公开(公告)日:2023-09-14

    申请号:US17691746

    申请日:2022-03-10

    CPC classification number: G06F8/36

    Abstract: Generating a template based on source code examples includes identifying a set of related files within a source code project, and identifying a set of textual content portions that are each at least partially repeated across a subset of the related files. A set of templates is generated, each comprising at least one textual content portion from the set of textual content portions. Each template is associated with a set of selection criteria. The set of templates is exposed for automated consumption within a source code editor, based on the set of selection criteria associated with each template. Consuming a template includes identifying a user input indicating creation of a source code block or file within the source code editor, identifying attribute(s) of the source code block or file, and using the attribute(s) to identify a selection criterion associated with a particular template. The particular template is automatically presented.

    MULTI-MODAL PROGRAM INFERENCE
    3.
    发明公开

    公开(公告)号:US20230176829A1

    公开(公告)日:2023-06-08

    申请号:US17544502

    申请日:2021-12-07

    CPC classification number: G06F8/33 G06F40/40 G06F8/38 G06F40/30 G06F8/10

    Abstract: Embodiments use a multi-modal approach to generate software programs that match a solution program description. The solution program description may include natural language, input-output examples, partial source code, desired operators, or other hints. Some embodiments use optimized prompts to a pre-trained language model to obtain initial candidate programs. Maximal program components are extracted and then recombined variously using component-based synthesis. Beam search reduces a solution program search space by discarding some candidates from a given synthesis iteration. Relevance metrics, string similarity metrics, operator frequency distributions, token rareness scores, and other optimizations may be employed. By virtue of optimizations and the multi-modal approach, a solution program may be obtained after fewer iterations than by use of a language model alone. The multi-modal approach is domain agnostic, as illustrated by examples using regular expression and cascading style sheet selector domain specific languages.

    COPY-PASTE-UPDATE EDIT AUTOMATION

    公开(公告)号:US20230116149A1

    公开(公告)日:2023-04-13

    申请号:US17497923

    申请日:2021-10-09

    Abstract: Embodiments automate several aspects of document copy-paste updates. An enhanced editor submits context, such as a copied section, pasted section, nearby text, or parser information, to an automatic suggestion generator. The editor gets back a suggestion for automatically changing the pasted section, thus helping users avoid tedium and errors. For instance, string substitutions begun by the user can be automatically and easily completed within the pasted section. Refactoring between variable declarations and parameter lists is detected and completed on request. Situation-specific transforms based on code synthesis, word associations, temporal edit patterns, anchor target lists, regular expressions, or autocompletion are offered. Suggestions are given inside the user's current workflow to avoid breaks in focus. Suggestions can be refined automatically in response to implicit or explicit user feedback. Users are warned of unedited pasted sections. Code review is aided by highlighting pasted sections.

Patent Agency Ranking