Natural Language Code Search Engine with Predictive Writing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Programmers face challenges in efficiently retrieving context-specific source code during the coding process due to the lack of a rapid, customizable, and relational code location and retrieval mechanism, relying heavily on personal knowledge and laborious search methods.

Innovation Solution

A natural language search engine with a predictive writing tool is developed, utilizing a system that filters and extracts source code comments, adds labels for supervised learning, and bridges source code and natural language classifiers to provide context-specific search results.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If programmers use traditional search methods to retrieve source code, then they can access code repositories, but the search process is laborious and time-intensive

Engineering Contradiction:
Improvecoding speedVSAvoidtime for code retrieval
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent replaces traditional mechanical search methods (manual browsing, keyword searching) with an AI-based natural language processing system. The system uses machine learning models to understand programmer intent and retrieve relevant code snippets automatically, substituting the manual search mechanism with an intelligent automated system that processes queries semantically rather than through rigid keyword matching.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables programmers to query code repositories using natural language without requiring knowledge of the repository structure, search syntax, or code organization. The AI model autonomously interprets the query, searches the codebase, and returns relevant results, allowing the system to serve itself in understanding and responding to programmer needs without manual intervention or specialized search skills.

Inventive Principle:
Principle #25Self-service

2Adaptability or versatility

If programmers rely on personal knowledge to find context-specific code, then they can access their own experience, but this limits access to organizational knowledge and increases training time

Engineering Contradiction:
Improveaccess to context-specific codeVSAvoidorganizational knowledge
Core Design Contradiction:
Adaptability or versatilityVSLoss of information

Solution Approach 1:

The patent creates a universal search system that serves multiple functions: it accesses organizational code repositories, understands natural language queries, retrieves context-specific code snippets, and adapts to different programming domains. The AI model is trained on diverse codebases and can handle various query types, making it a multi-functional tool that replaces the need for individual programmers to have specialized knowledge of different code repositories and contexts.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The AI-based search system acts as an intermediary between programmers and organizational code repositories. Instead of programmers directly navigating complex codebases or relying on personal knowledge, the AI model mediates by understanding natural language queries, searching the code repository, and returning relevant results. This intermediary layer enables access to organizational knowledge without requiring programmers to have direct familiarity with the code structure or context.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Ease of operation

If general source code reference libraries are used, then code can be organized categorically, but the libraries lack organizational customization and context-specific relevance

Engineering Contradiction:
Improvecode organizationVSAvoidcontext-specific relevance
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent implements local quality by training the AI model on organization-specific codebases and contexts. Rather than using a single general-purpose search system, the model adapts to the specific coding styles, conventions, and contexts of different organizations or teams. This allows the search system to provide locally optimized results that are relevant to the specific organizational context while maintaining ease of use through natural language queries.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4046010B1Natural language search engine with a predictive writing tool for coding
Publication Date: 2023.08.02 GLOBANT ESPANA SA
  • EP4046010B1 patent drawingFigure 1
  • EP4046010B1 patent drawingFigure 2
  • EP4046010B1 patent drawingFigure 3

AI summary

A system including a source code classifier, a natural language classifier, and a search engine operative on a search engine index associating outputs of the source code classifier and the natural language classifier to match a natural language search query with at least one computer code section from a computer program.