hashimportai 0.1.0

Manifest-driven workflow execution SDK for HashImport

hashimportai

Python SDK for executing HashImport workflows using hashlets.

Installation

pip install hashimportai

For development or to install from source, see PUBLISHING.md.

Quick Start

Complete Example with Subscribers

import asyncio
from hashimportai import WorkflowRunner

# Define subscriber callbacks for each action
def on_vault_linker(action_name, response):
    print(f"[Subscriber] {action_name} completed: {response}")

def on_upload_pdf(action_name, response):
    print(f"[Subscriber] {action_name} completed: {response}")
    if response and response.get("data"):
        result = response.get("data", {}).get("result", {})
        if isinstance(result, dict) and "linked_output" in result:
            print(f"  -> Uploaded file: {result['linked_output']}")

def on_esign_pdf(action_name, response):
    print(f"[Subscriber] {action_name} completed: {response}")

def on_esign_signed_url(action_name, response):
    print(f"[Subscriber] {action_name} completed: {response}")
    if response and response.get("data"):
        result = response.get("data", {}).get("result", {})
        if isinstance(result, dict) and "url" in result:
            print(f"  -> Signed URL: {result['url']}")

# Create subscribers dictionary mapping action names to callbacks
subscribers = {
    "vault_linker": on_vault_linker,
    "upload_pdf": on_upload_pdf,
    "esign_pdf": on_esign_pdf,
    "esign_signed_url": on_esign_signed_url,
}

async def main():
    # Option 1: Pass credentials directly (recommended)
    runner = WorkflowRunner(
        workflow_id="69576bd6af6515a54826291c",
        multi_part_actions=["upload_pdf"],
        subscribers=subscribers,
        api_key="your-api-key",
        base_url="http://localhost:6003/api/v1"
    )
    
    form_data = {
        "vault_linker": {"hashlet_id": "694d1e4bb731610cf532267b"},
        "upload_pdf": {"file": "tmp/document.pdf"},  # Relative to current working directory
        "esign_pdf": {
            "filename": "is_link",  # Will be auto-populated from upload_pdf output
            "creator_id": "1234567890",
            "purpose": "test",
            "email": "[e-mail]",
            "company_full_name": "John Doe",
            "location": "123 Main St, Anytown, USA",
            "sig_x1": 100,
            "sig_y1": 100,
            "sig_x2": 200,
            "sig_y2": 200,
            "sign_page": 1,
        },
        "esign_signed_url": {"filename": "document.pdf"},
    }
    
    await runner.run("hashhub-fc77bf01", form_data)

asyncio.run(main())

Option 2: Use Environment Variables

export HASHIMPORTAI_API_KEY="your-api-key"
export HASHIMPORTAI_BASE_URL="http://localhost:6003/api/v1"

import asyncio
from hashimportai import WorkflowRunner

async def main():
    # Credentials loaded from environment variables
    runner = WorkflowRunner(
        workflow_id="your-workflow-id",
        multi_part_actions=["upload_pdf"],
        subscribers=subscribers  # Optional: can be None or empty dict
    )
    
    form_data = {
        "upload_pdf": {"file": "document.pdf"}
    }
    
    await runner.run("hashhub-xxx", form_data)

asyncio.run(main())

Configuration

API Key and Base URL

You can provide credentials in two ways:

1. Direct parameters (recommended for scripts):

   runner = WorkflowRunner(
       workflow_id="...",
       multi_part_actions=["..."],
       api_key="your-api-key",
       base_url="http://localhost:6003/api/v1"
   )
   

2. Environment variables (recommended for production):

   export HASHIMPORTAI_API_KEY="your-api-key"
   export HASHIMPORTAI_BASE_URL="http://localhost:6003/api/v1"
   

Subscribers (Callbacks)

Subscribers allow you to handle responses from each action:

subscribers = {
    "action_name": lambda action, response: print(f"{action} completed: {response}")
}

Each callback receives:

• action_name: The name of the action that completed
• response: The response data from the action execution

File Paths

File paths in form_data are resolved relative to your current working directory, not the SDK directory:

• {"file": "document.pdf"} → looks for ./document.pdf in your current directory
• {"file": "tmp/document.pdf"} → looks for ./tmp/document.pdf in your current directory
• {"file": "/absolute/path/to/file.pdf"} → uses absolute path as-is

Auto-linking Actions

Use "is_link" as a value to automatically link outputs between actions:

form_data = {
    "upload_pdf": {"file": "document.pdf"},
    "esign_pdf": {
        "filename": "is_link",  # Will be replaced with output from upload_pdf
        # ... other fields
    }
}

The SDK automatically extracts the linked_output from the previous action and replaces "is_link" with the actual filename.