Skip to main content

​
🛠️ How to Build Workflows

After successfully installing the workflowcommon Python SDK, it’s now time to build the workflow:
  1. ➡️ Import nodes and builder
  2. ➡️ Create the workflow builder
  3. ➡️ Define and add nodes
  4. ➡️ Connect the nodes
  5. ⏱️ Build and visualize the pipeline
  6. ⏱️ Execute the pipeline
  7. ⏱️ Get outputs and results

​
✍️ Example 1: Get Text from a File → Analyze with LLM → Upload Result

  1. ➡️ Import nodes and builder 
    # Import workflow builder
from workflowcommon.builder import FlowBuilder

# Import nodes
from workflowcommon.nodes import (
    S3LoadFile,
    LLM,
    S3UploadFile
)
  2. ➡️ Create the workflow builder 
    # Create workflow builder
builder = FlowBuilder()
  3. ➡️ Define and add nodes 
    # Load file from S3
s3_loader = builder.add_node(
    S3LoadFile.Node(
        name="S3LoadFile",
        params=S3LoadFile.Params(
            output_field="file_content",
            user_id="user123",
            file_key="my_txt_file.txt",
            file_type="text",
        )
    )
)

# Analyze with LLM
llm = builder.add_node(
    LLM.Node(
        name="LLM",
        params=LLM.Params(
            input_field="file_content.data",
            output_field="response",
            llm_provider="Mistral",  # or "OpenAI", "Anthropic"
        )
    )
)

# Upload result to S3
s3_uploader = builder.add_node(
    S3UploadFile.Node(
        name="S3UploadFile",
        params=S3UploadFile.Params(
            input_field="response.content",
            user_id="user123",
            file_key="my_md_file.md",  # upload results to Markdown file
        )
    )
)
  4. ➡️ Connect the nodes 
    # Connect the flow in the correct order
s3_loader.connect_to(llm)
llm.connect_to(s3_uploader)

​
📄 Example 2: Load a PDF and Extract Text

  1. ➡️ Import nodes and builder 
    # Import workflow builder
from workflowcommon.builder import FlowBuilder

# Import nodes
from workflowcommon.nodes import (
    S3LoadFile,
    ExtractTextPDF
)
  2. ➡️ Create the workflow builder 
    # Create workflow builder
builder = FlowBuilder()
  3. ➡️ Define and add nodes 
    # Load PDF file from S3
s3_loader = builder.add_node(
    S3LoadFile.Node(
        params=S3LoadFile.Params(
            output_field="pdf_base64",
            user_id="user123",
            file_key="my_pdf_file.pdf",
            file_type="pdf"
        )
    )
)

# Extract text from PDF
pdf_text_extractor = builder.add_node(
    ExtractTextPDF.Node(
        name="ExtractTextPDF",
        params=ExtractTextPDF.Params(
            input_field="pdf_base64",
            output_field="pdf_text"
        )
    )
)
  4. ➡️ Connect the nodes 
    # Connect the flow in the correct order
s3_loader.connect_to(pdf_text_extractor)

​
🧠 Tips for Designing Workflows

  • Use .connect_to() to link nodes in the correct order
  • Specify input_field and output_field to pass data between nodes (or not depending on node type)
  • Call pipeline.execute() to run the pipeline sequentially
  • Use display(pipeline) in a notebook to see a visual graph

​
🔍 More Ideas

Here are other workflows you can build with the SDK:
  • Load and extract text from PDF → Summarize with LLM → Store summary in graph memory
  • Generate Cypher queries from user prompts using Text2Cypher node
  • Build a graph from extracted entities and write it into TuringDB
  • Loop over a list of companies and fetch financial data using ForEach node
For a full list of supported nodes and their parameters, see the Node Reference. Build fast, flexible, and intelligent workflows with TuringDB + workflowcommon to supercharge your data, AI, and graph analytics stack.
⌘I