OpenAI Model Wrappers, Graph Building Tools, and Speech Summary Graph Pipeline

[PeriniM]

This update introduces several significant enhancements to our web scraping and processing capabilities, designed to streamline the development of graph-based data extraction pipelines and leverage OpenAI's latest models for advanced text and image processing. Key additions and improvements include:

• OpenAI Model Wrappers: Implemented OpenAITextToSpeech and OpenAIImageToText classes as convenient wrappers around OpenAI's API. These wrappers simplify the process of converting text to speech and extracting textual content from images, respectively.

• Graph Building Module (builders): Introduced a new module named builders which contains classes aimed at facilitating the creation of graph-based pipelines using language model prompts. The centerpiece of this module is the GraphBuilder class, which not only assists in graph construction but also supports exporting the graph design in Graphviz format for visualization and reuse.

• Standard Graph Pipeline - SpeechSummaryGraph: Developed a new standard graph pipeline called SpeechSummaryGraph. This pipeline automates the extraction and summarization of web page content, then outputs both a text summary and an MP3 audio file with a synthesized voiceover of the summary. This feature harnesses the power of GPT-4 for content summarization and conversion to speech, providing a more accessible way to consume web content.

• Compliance with BaseNode Specifications: Updated the newly introduced nodes to ensure they meet the specifications required by BaseNode. This ensures consistency and compatibility within our graph-based processing framework.

• Enhanced Examples: Added new examples to demonstrate the practical applications of GPT-4 vision and text-to-speech capabilities within our framework. These examples serve as a guide for developers looking to integrate similar functionalities into their pipelines.

These enhancements collectively aim to provide a more robust, flexible, and user-friendly toolkit for developers working on web scraping and content processing projects, leveraging the latest advancements in AI and machine learning.

The following graph has been automatically generated by the GraphBuilder class.

[github-actions]

Dependency Review

✅ No vulnerabilities or license issues found.

Scanned Manifest Files

poetry.lock

• graphviz@0.20.1
• urllib3@2.2.1
• urllib3@2.2.0

pyproject.toml

• graphviz@0.20.1