The o1-reasoning Pipeline is a sophisticated framework designed to process user messages and interact seamlessly with the OpenAI API. It effectively decomposes user input into manageable steps, processes each step individually, and compiles a comprehensive final answer based on the results of these steps.
- Step-by-Step Processing: Analyzes and divides user input into smaller, actionable steps to ensure thorough understanding and processing.
- Dynamic Code Execution: Executes Python code snippets in real-time for computations, validations, and enhanced functionality.
- Comprehensive Logging: Offers detailed logging capabilities with color-coded log levels for easy tracking and debugging.
- Chain of Thoughts: Implements a structured reasoning approach to handle complex queries by planning and executing a sequence of thought steps.
- Flexible Tool Integration: Integrates various tools such as web search, news retrieval, maps search, and website content extraction to enrich responses.
- User Input Handling: Supports multiple types of user inputs including logical resolution, code generation, article creation, research, and generic queries.
Processes user queries requiring logical reasoning and step-by-step problem-solving.
Executes Python code snippets to validate solutions or perform computations in real-time.
Generates comprehensive articles based on user-specified topics with structured sections and detailed content.
Conducts thorough research on specified topics, summarizing findings with credible sources and data.
Handles general user questions, providing clear and accurate responses tailored to the user's language and requirements.
- UserInputDetails: Captures details about the user's input, including language, original question, desired outcome, and preliminary answers.
- TextWebSearchRequest: Handles text-based web search requests using DuckDuckGo.
- WebSiteContent: Retrieves and processes content from specified websites.
- WebAnswersRequest: Fetches instant answers related to the user's query.
- NewsSearchRequest: Searches for news articles relevant to the user's input.
- MapsSearchRequest: Performs map-related searches based on user specifications.
- CodeExecution: Manages the execution of Python code snippets provided by the user.
- Step & ChainOfThoughts: Structures the reasoning process into discrete steps for complex problem-solving.
- AnswerEvaluation: Evaluates the correctness of generated answers.
The Pipeline class orchestrates the entire process, handling user input extraction, tool execution, chain of thoughts processing, and final answer compilation. Key components include:
- Valves Configuration: Stores API keys, model settings, and execution limits.
- Logging Setup: Implements a custom color formatter for enhanced log readability.
- Tool Integration: Incorporates various tools for web content retrieval, news search, maps search, and code execution.
- Chain of Thoughts Execution: Breaks down complex queries into manageable steps, processes each step, and compiles the final answer.
- Answer Generation: Utilizes OpenAI's GPT models to generate coherent and contextually relevant responses based on processed data.
The pipeline uses a custom logging formatter to add colors to log levels, enhancing readability during development and debugging.
# Example of logging output
2024-09-18 12:00:00 [INFO] o1-reasoning: Pipeline startup initiated.
2024-09-18 12:00:01 [DEBUG] o1-reasoning: OpenAI client initialized.Contributions are welcome! Please follow these steps:
- Fork the repository.
- Create a new branch:
git checkout -b feature/YourFeature. - Commit your changes:
git commit -m 'Add YourFeature'. - Push to the branch:
git push origin feature/YourFeature. - Open a pull request.
- Victor Carvalho Tavernari
- This project is licensed under the MIT License.
Feel free to reach out if you have any questions or need further assistance!
- Email: victortavernari@gmail.com