on_boarding.md

graph LR
    RagaAICatalyst_Client["RagaAICatalyst Client"]
    Platform_Management_Services["Platform Management Services"]
    Internal_API_Completion["Internal API Completion"]
    Red_Teaming_Orchestration["Red Teaming Orchestration"]
    Core_Tracing_System["Core Tracing System"]
    Trace_Data_Management_Export["Trace Data Management & Export"]
    LLM_Framework_Integration["LLM Framework Integration"]
    External_LLM_Models["External LLM Models"]
    External_LLM_Frameworks["External LLM Frameworks"]
    Guardrail_Execution_Engine["Guardrail Execution Engine"]
    RagaAICatalyst_Client -- "Manages" --> Platform_Management_Services
    RagaAICatalyst_Client -- "Authenticates with" --> Internal_API_Completion
    Platform_Management_Services -- "Interacts with" --> Internal_API_Completion
    Platform_Management_Services -- "Utilizes" --> External_LLM_Models
    Internal_API_Completion -- "Serves" --> RagaAICatalyst_Client
    Internal_API_Completion -- "Serves" --> Platform_Management_Services
    Red_Teaming_Orchestration -- "Utilizes" --> External_LLM_Models
    Red_Teaming_Orchestration -- "Feeds results to" --> Platform_Management_Services
    Core_Tracing_System -- "Orchestrates" --> Trace_Data_Management_Export
    Core_Tracing_System -- "Monitors" --> External_LLM_Models
    Trace_Data_Management_Export -- "Sends data to" --> Internal_API_Completion
    Trace_Data_Management_Export -- "Receives data from" --> Core_Tracing_System
    LLM_Framework_Integration -- "Integrates with" --> External_LLM_Frameworks
    LLM_Framework_Integration -- "Sends data to" --> Trace_Data_Management_Export
    External_LLM_Models -- "Used by" --> Platform_Management_Services
    External_LLM_Models -- "Used by" --> Red_Teaming_Orchestration
    External_LLM_Frameworks -- "Integrated by" --> LLM_Framework_Integration
    Guardrail_Execution_Engine -- "Interacts with" --> External_LLM_Models
    Guardrail_Execution_Engine -- "Uses" --> Core_Tracing_System
    click RagaAICatalyst_Client href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/RagaAICatalyst_Client.md" "Details"
    click Platform_Management_Services href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/Platform_Management_Services.md" "Details"
    click Internal_API_Completion href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/Internal_API_Completion.md" "Details"
    click Red_Teaming_Orchestration href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/Red_Teaming_Orchestration.md" "Details"
    click Core_Tracing_System href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/Core_Tracing_System.md" "Details"
    click Trace_Data_Management_Export href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/Trace_Data_Management_Export.md" "Details"
    click LLM_Framework_Integration href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/LLM_Framework_Integration.md" "Details"
    click External_LLM_Models href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/External_LLM_Models.md" "Details"
    click External_LLM_Frameworks href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/External_LLM_Frameworks.md" "Details"
    click Guardrail_Execution_Engine href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main//RagaAI-Catalyst/Guardrail_Execution_Engine.md" "Details"

Loading

Component Details

Final architecture analysis for RagaAI-Catalyst, consolidating insights from the Control Flow Graph (CFG) and source code analysis.

RagaAICatalyst Client

The primary user-facing interface for the RagaAI Catalyst platform. It handles initial setup, user authentication, and orchestrates access to various platform functionalities, providing a unified entry point for users.

Related Classes/Methods:

• ragaai_catalyst.ragaai_catalyst (1:1)

Platform Management Services

A comprehensive suite of core services responsible for managing key aspects of the RagaAI Catalyst platform. This includes functionalities for dataset management, model evaluation, experiment tracking, guardrail configuration, prompt template management, and synthetic data generation.

Related Classes/Methods:

• ragaai_catalyst.dataset.Dataset (18:733)
• ragaai_catalyst.evaluation (1:1)
• ragaai_catalyst.experiment (1:1)
• ragaai_catalyst.guardrails_manager (1:1)
• ragaai_catalyst.prompt_manager (1:1)
• ragaai_catalyst.synthetic_data_generation (1:1)

Internal API Completion

The central backend service that processes all internal API requests and manages data persistence for the RagaAI Catalyst platform. It acts as the critical communication hub, facilitating data exchange and operations across various client-side and internal components.

Related Classes/Methods: None

Red Teaming Orchestration

Manages the entire lifecycle of red teaming exercises. This includes coordinating the generation of adversarial scenarios and test cases, facilitating interactions with LLMs for red teaming, evaluating conversational outcomes, and uploading the results back into the platform.

Related Classes/Methods:

• ragaai_catalyst.redteaming.red_teaming (1:1)
• ragaai_catalyst.redteaming.data_generator.scenario_generator (1:1)
• ragaai_catalyst.redteaming.data_generator.test_case_generator (1:1)
• ragaai_catalyst.redteaming.evaluator (1:1)
• ragaai_catalyst.redteaming.llm_generator (1:1)
• ragaai_catalyst.redteaming.upload_result (1:1)

Core Tracing System

The foundational component for all observability and tracing activities within RagaAI Catalyst. It orchestrates the lifecycle of traces, collects detailed system and component-specific information (e.g., LLM interactions, agent/tool executions, custom code), and prepares this data for further processing and export.

Related Classes/Methods:

• ragaai_catalyst.tracers.tracer (1:1)
• ragaai_catalyst.tracers.agentic_tracing.tracers.main_tracer (1:1)
• ragaai_catalyst.tracers.agentic_tracing.tracers.base (1:1)
• ragaai_catalyst.tracers.agentic_tracing.tracers.llm_tracer (1:1)
• ragaai_catalyst.tracers.agentic_tracing.tracers.agent_tracer (1:1)
• ragaai_catalyst.tracers.agentic_tracing.tracers.tool_tracer (1:1)
• ragaai_catalyst.tracers.agentic_tracing.tracers.custom_tracer (1:1)
• ragaai_catalyst.tracers.agentic_tracing.utils.file_name_tracker (1:1)
• ragaai_catalyst.tracers.agentic_tracing.utils.system_monitor (1:1)
• ragaai_catalyst.tracers.agentic_tracing.utils.llm_utils (1:1)
• ragaai_catalyst.tracers.utils.trace_json_converter (1:1)

Trace Data Management & Export

Responsible for defining the internal data structures used for traces, organizing all collected trace information, and managing the packaging, uploading, and exporting of this trace data (including associated code and calculated metrics) to the RagaAI platform.

Related Classes/Methods:

• ragaai_catalyst.tracers.agentic_tracing.data.data_structure (1:1)
• ragaai_catalyst.tracers.upload_traces (1:1)
• ragaai_catalyst.tracers.agentic_tracing.upload.trace_uploader (1:1)
• ragaai_catalyst.tracers.agentic_tracing.upload.upload_agentic_traces (1:1)
• ragaai_catalyst.tracers.exporters.ragaai_trace_exporter (1:1)
• ragaai_catalyst.tracers.exporters.raga_exporter (1:1)
• ragaai_catalyst.tracers.exporters.dynamic_trace_exporter (1:1)
• ragaai_catalyst.tracers.exporters.file_span_exporter (1:1)

LLM Framework Integration

Provides specialized mechanisms to seamlessly integrate RagaAI Catalyst's tracing capabilities with popular external LLM frameworks such as Langchain and LlamaIndex. This is achieved by implementing their respective callback or instrumentation mechanisms to automatically capture relevant events.

Related Classes/Methods:

• ragaai_catalyst.tracers.langchain_callback (1:1)
• ragaai_catalyst.tracers.llamaindex_callback (1:1)
• ragaai_catalyst.tracers.llamaindex_instrumentation.LlamaIndexInstrumentationTracer (367:423)

External LLM Models

Represents various Large Language Models (LLMs) hosted externally (e.g., OpenAI, Anthropic, Google GenAI, Vertex AI). These models are utilized by the RagaAI Catalyst platform for a wide range of tasks, including generative AI, analytical processing, and guardrail enforcement.

Related Classes/Methods: None

External LLM Frameworks

Represents popular open-source or commercial LLM frameworks (e.g., Langchain, LlamaIndex) that provide structured approaches for building applications with Large Language Models. RagaAI Catalyst integrates with these frameworks to offer enhanced tracing and monitoring capabilities.

Related Classes/Methods: None

Guardrail Execution Engine

A dedicated component responsible for executing defined guardrails. It applies specified rules to both inputs and outputs of LLM interactions and other processes, interacting directly with Large Language Models to enforce policies and ensure adherence to safety and performance criteria.

Related Classes/Methods:

• ragaai_catalyst.guard_executor (1:1)

FAQ