Skip to content

index.md

Latest commit

 

History

History
90 lines (62 loc) · 3.31 KB

index.md

File metadata and controls

90 lines (62 loc) · 3.31 KB

Examples

The airflow-ai-sdk comes with several example DAGs that demonstrate how to use the various decorators and features. All examples can be found in the examples/dags directory of the repository.

Available Examples

1. LLM Calls - GitHub Changelog

View example code

Demonstrates using @task.llm to summarize GitHub commits with a language model.

Features demonstrated:

  • Using @task.llm decorator with a specific model (gpt-4o-mini)
  • Specifying a detailed system prompt for commit summarization
  • Transforming task input to LLM input (joining commits into a string)
  • Weekly scheduling for regular changelog generation

2. Structured Output - Product Feedback

View example code

Shows how to use @task.llm with Pydantic models to parse structured data from product feedback.

Features demonstrated:

  • Creating a custom Pydantic model for structured LLM output
  • PII masking in preprocessing
  • Using output_type parameter with a Pydantic model
  • Task mapping with expand() to process multiple feedback items
  • Conditional execution with AirflowSkipException

3. Agent Tasks - Deep Research

View example code

Illustrates how to use @task.agent with custom tools to perform deep research on topics.

Features demonstrated:

  • Creating a custom agent with tools
  • Using the @task.agent decorator
  • Creating custom tools (web content fetching)
  • Integrating with external APIs (DuckDuckGo search)
  • Using runtime parameters for dynamic DAG execution

4. Branching Tasks - Support Ticket Routing

View example code

Demonstrates how to use @task.llm_branch to route support tickets based on priority.

Features demonstrated:

  • Using @task.llm_branch for conditional workflow routing
  • Setting up branch tasks based on LLM decisions
  • Configuring allow_multiple_branches parameter
  • Processing DAG run configuration parameters
  • Detailed prompt engineering for classification tasks

5. Embedding Tasks - Text Embedding

View example code

Shows how to use @task.embed to create vector embeddings from text.

Features demonstrated:

  • Using @task.embed decorator
  • Specifying embedding model parameters
  • Generating vector embeddings from text
  • Configuring encoding parameters (normalization)
  • Task mapping to process multiple texts in parallel

Running the Examples

The examples can be run in a local Airflow environment with the following steps:

  1. Clone the examples repository:

    git clone https://github.com/astronomer/ai-sdk-examples.git

  2. Navigate to the examples directory:

    cd ai-sdk-examples

  3. Start the Airflow environment:

    astro dev start

For more information on the individual examples, refer to the code comments in each example file.