ParallelWorkflow

Parallel Workflows with LangGraph

This directory contains examples of implementing parallel processing workflows using LangGraph. These examples demonstrate how to execute multiple independent tasks simultaneously and then combine their results.

Examples

1. Cricket Statistics Analysis

File: practice.ipynb

This example demonstrates parallel processing of cricket statistics to calculate various performance metrics for a batsman. The workflow calculates three independent metrics in parallel:

• Strike Rate: (Runs / Balls) * 100
• Balls per Boundary: Total balls / (Fours + Sixes)
• Boundary Percentage: ((Fours4) + (Sixes6)) / Runs * 100

Key Components:

• State Management: Uses TypedDict to track:

  • Runs, balls, fours, sixes
  • Calculated metrics (strike rate, balls per boundary, boundary percentage)
  • Summary string combining all metrics

• Parallel Processing:

  • Three independent calculation nodes run in parallel
  • Results are combined in a summary node
  • Demonstrates fan-out/fan-in pattern

2. Essay Evaluation System

File: practice2.ipynb

This example shows a more complex parallel workflow for evaluating essays using multiple criteria simultaneously. The system evaluates an essay on different aspects in parallel:

• Language and Structure
• Content and Argumentation
• Originality and Insight
• Grammar and Style

Key Components:

• Structured Output: Uses Pydantic models for consistent evaluation criteria
• Parallel Evaluation: Multiple evaluators analyze different aspects simultaneously
• Result Aggregation: Combines individual evaluations into a comprehensive review
• Google's Gemini Model: Leverages AI for sophisticated essay analysis

Setup

1. Install the required dependencies:

   pip install -r requirements.txt

2. For the essay evaluation example, you'll need a Google API key for the Gemini model:

   • Obtain an API key from Google AI Studio
   • Set it in the notebook or as an environment variable

Usage

Cricket Statistics Example

1. Open practice.ipynb
2. Modify the initial state with cricket statistics:

   initial_state = {
       "runs": 120,
       "balls": 100,
       "fours": 10,
       "sixes": 3
   }

3. Run the cells to see parallel processing in action

Essay Evaluation Example

1. Open practice2.ipynb
2. Set your Google API key if not using environment variables
3. Modify the essay text as needed
4. Run the workflow to see parallel evaluation results

Key Concepts

• Parallel Execution: Running multiple independent tasks simultaneously
• State Management: Maintaining and updating shared state across parallel nodes
• Fan-out/Fan-in Pattern: Splitting work into parallel tasks and combining results
• Structured Output: Using Pydantic models for consistent data validation
• AI Integration: Leveraging language models for content analysis

Dependencies

• Python 3.8+
• langgraph
• langchain-google-genai (for essay evaluation)
• pydantic

Related Workflows

Check out other workflow examples in the parent directory for different LangGraph patterns and use cases, including conditional logic and iterative refinement.