Skip to content

TESTING.md

Latest commit

 

History

History
102 lines (81 loc) · 4.1 KB

TESTING.md

File metadata and controls

102 lines (81 loc) · 4.1 KB
Raw

Testing Guide for LangChain Parallel Web Integration

This document explains how to run the LangChain standard tests for both ChatParallelWeb and ParallelWebSearchTool integrations.

Overview

Both integrations follow the official LangChain standard testing framework:

  • ChatParallelWeb: Chat model integration with OpenAI-compatible interface
  • ParallelWebSearchTool: Search tool integration for web research capabilities

Architecture

Chat Model (ChatParallelWeb)

  • OpenAI Compatibility: Uses OpenAI-compatible interface with Parallel endpoint
  • Message Handling: Automatic message merging for API compliance
  • LangChain Integration: Full metadata support (lc_secrets, lc_attributes, etc.)

Search Tool (ParallelWebSearchTool)

  • Direct API Access: Direct integration with Parallel Search API
  • Input Validation: Comprehensive parameter validation
  • Metadata Collection: Search timing, result counts, processor information
  • Async Support: Full async/await support with proper event loop handling

Test Types

1. Unit Tests (tests/unit_tests/)

  • Purpose: Test components in isolation without external API calls
  • Network: Disabled (no external connections)
  • API Key: Uses mock/test values
  • Files:
    • test_chat_models.py - Chat model unit tests
    • test_search.py - Search tool unit tests

2. Integration Tests (tests/integration_tests/)

  • Purpose: Test components with real API calls
  • Network: Enabled (requires internet connection)
  • API Key: Requires valid PARALLEL_API_KEY environment variable
  • Files:
    • test_chat_models.py - Chat model integration tests
    • test_tools.py - Search tool integration tests

Running Tests

Unit Tests Only (Recommended for Development)

# Run unit tests with network disabled
poetry run pytest --disable-socket --allow-unix-socket --asyncio-mode=auto tests/unit_tests/ -v

Integration Tests (Requires API Key)

# Set your API key first
export PARALLEL_API_KEY="your-actual-api-key-here"

# Run integration tests
poetry run pytest --asyncio-mode=auto tests/integration_tests/ -v

All Tests

# Run all tests (integration tests will fail without API key)
poetry run pytest --disable-socket --allow-unix-socket --asyncio-mode=auto tests/ -v

Test Configuration

Chat Model Capabilities (ChatParallelWeb)

Capability Supported Notes
Basic Chat ✅ Full sync/async support
Streaming ✅ Real-time response streaming
System Messages ✅ OpenAI-compatible interface
Usage Metadata ❌ API doesn't return usage data
JSON Mode ❌ Not supported by API
Tool Calling ❌ Not supported by API
Structured Output ❌ Not supported by API
Image Inputs ❌ Not supported by API
Audio/Video Inputs ❌ Not supported by API

Search Tool Capabilities (ParallelWebSearchTool)

Capability Supported Notes
Objective-based Search ✅ Natural language search goals
Query-based Search ✅ Traditional keyword searches
Domain Filtering ✅ Include/exclude specific domains
Processor Selection ✅ Base (fast) and Pro (quality) modes
Async Support ✅ Full async/await functionality
Metadata Collection ✅ Search timing and result statistics
Input Validation ✅ Parameter validation and error handling

Current Test Status

Chat Model Tests

  • Unit Tests: 8+ tests covering initialization, serialization, environment setup
  • Integration Tests: 40+ tests covering chat functionality, streaming, message handling
  • Skipped Tests: Usage metadata, tool calling, structured output (not supported by API)

Search Tool Tests

  • Unit Tests: 10+ tests covering validation, error handling, metadata collection
  • Integration Tests: 5+ tests covering real API calls and response parsing
  • All Tests: Currently passing with proper API key configuration