Skip to content

Latest commit

 

History

History
117 lines (96 loc) · 2.72 KB

architecture.md

File metadata and controls

117 lines (96 loc) · 2.72 KB

SPARC CLI Architecture

System Overview

+----------------+     +-------------+     +--------------+
|  CLI Interface |---->| Core Engine |---->| Tool System  |
+----------------+     +-------------+     +--------------+
        |                    |                   |
        v                    v                   v
+----------------+     +-------------+     +--------------+
|  Config System |     | LLM Clients |     | File System  |
+----------------+     +-------------+     +--------------+

Core Components

CLI Interface

  • Command parsing
  • Option handling
  • Output formatting
  • Interactive prompts

Core Engine

  • Task coordination
  • Mode selection
  • Provider management
  • Memory management

Tool System

  • Shell command execution
  • File operations
  • Expert knowledge queries
  • Human interaction
  • Web scraping and data extraction
    • Rate limiting controller
    • Response cache manager
    • Error recovery system
    • Data format handlers

Key Modules

LLM Integration (llm.py)

  • Provider initialization
  • Model configuration
  • API communication
  • Error handling

Tool Configuration (tool_configs.py)

  • Tool registration
  • Permission management
  • Mode-specific toolsets
  • Tool chain composition

Memory System (tools/memory.py)

  • State management
  • Context tracking
  • Key facts storage
  • Work logging

Shell Interface (tools/shell.py)

  • Command execution
  • Output capture
  • Safety controls
  • Cowboy mode

Integration Points

External Systems

  • LLM Provider APIs
  • File System
  • Shell Environment
  • Version Control

Internal Components

  • Tool <-> Memory
  • LLM <-> Tools
  • Config <-> Engine
  • Shell <-> Safety Controls
  • Scraper <-> Cache System

Scraping Architecture

+----------------+     +-------------+     +--------------+
|  Rate Limiter  |---->| HTTP Client |---->| Cache Layer  |
+----------------+     +-------------+     +--------------+
        |                    |                   |
        v                    v                   v
+----------------+     +-------------+     +--------------+
| Retry Manager  |     | Parser Hub  |     | Error Handler|
+----------------+     +-------------+     +--------------+

The scraping system implements:

  • Distributed rate limiting
  • Intelligent retry backoff
  • Response caching with TTL
  • Format-specific parsing
  • Graceful error recovery

Workflow

[User Input] -> [CLI] -> [Config] -> [Engine]
                                       |
[Output] <- [Console] <- [Tools] <- [LLM]
  1. User provides task description
  2. CLI processes command options
  3. Config system initializes environment
  4. Engine coordinates execution
  5. Tools perform operations
  6. Results stream to console

For usage instructions, see the Usage Guide.