AI-Harness

AI-Harness is a Python package designed to showcase and utilize the capabilities of the AI Harness platform. This SDK provides a set of tools and functionalities to interact with the AI Harness ecosystem and integrate AI models seamlessly.

Installation

You can install AI-Harness-sdk using pip:

pip install ai-harness-sdk

Documentation

The documentation can be found on https://docs.ai-harness.com

Usage

# Import necessary modules/class
from dotenv import load_dotenv
from ai_harness import Documents

# Load environment variables from a .env file
load_dotenv();

# Initialize a Documents instance
doc_instance = Documents()

# To Create a collection in the AI Harness platform
doc_instance.create_collection(collection_name="Your Collection Name")

# Upload documents to the created collection
doc_instance.upload_documents(
    doc_path=r"document_path",  # Specify the path to the documents
    collection_id="",  # Provide the ID of the target collection
    ingest_with_google="false"  # Indicate whether to ingest documents using Google API or open source pdf loader
)

Applet Usage

# import applets from ai_harness
from ai_harness import applet_name

Conversation Applet

# run a Conversation applet
result = Conversation(prompt="Your Prompt").run()

Place Making Applet

# run a Place Making applet
result = PlaceMaking(site_name="write site name here").run()

Query Generator Applet

# run a Query Generator applet
result = QueryGenerator(
    prompt="Your Prompt", 
    dialect="dialect name",
    schema="schema",
    db_type="db type(relational or non-relational)",
    additional_filters="any additional filters"
    ).run()

Site Analysis Applet

# run a Site Analysis applet
result = SiteAnalysis(
    lat="latitude of site", # either use coordinates or site ID
    long="longitude of site",
    siteId="URA site id of site", # using URA site ID 
    properties=["stations","market_analysis","competitive_analysis","airport","seaport","zones","history","demographics"], #  list of features that can be used 
    market_analysis="residential", # market analysis (residential , commercial or industrial) 
    competitive_analysis="industrial" # competitive analysis (residential , commercial or industrial) 
    ).run()

JSON TO JSON applet

# run JSON TO JSON applet
result = JsonToJson(
    input_json="input json", 
    output_json="output json", 
    ).run()

QA Retrieval Applet

# run QA Retrieval applet
result = QaRetrieval(
    prompt="prompt",
    document_id="id of document uploaded on ai harness", 
    ).run()

DatasetAI Applet

# run DatasetAI applet
dataset_obj = DatasetAI()
dataset = dataset_obj.create_dataset(name="name of dataset", description="dataset description")
dataset_obj.upload_collection_csv(
    file_path="path to csv file",
    date_format="date format",
    collection_name="name of collection",
    dataset_id="id of dataset made on ai harness",
)
result = DatasetAI(
    prompt="prompt",
    dataset_id="id of dataset made on ai harness", 
    ).run()