knowledgeai-client 0.5.1

A python client for the Avvia Intelligence - Knowledge AI Rest API

Avvia Intelligence - KnowledgeAI Python Client

Version: 0.5.1, (c) 2024 Arvato Systems

This library is designed to provide easy access to the KnowledgeAI backend service. It enables users to manage projects, upload documents, and perform various tasks programmatically. The client offers an intuitive interface for seamless integration into Python applications.

Installation

pip install knowledgeai-client

poetry add knowledgeai-client

Usage

from knowledgeai import KnowledgeAIClient

# Initialize the KnowledgeAIClient
client = KnowledgeAIClient(url="https://your-knowledgeai-service.com", api_key="your_api_key")

# Uploads all files in a directory to the specified project
client.upload_directory(project_id="project_id", directory="/path/to/directory", roles="admin,user")

# Uploads multiple documents to the specified project
client.index_documents(project_id="project_id", files=["/path/to/file1.txt", "/path/to/file2.txt"], roles="user,admin")

# Indexes the specified URL to the specified project
client.index_urls(project_id="project_id", index_urls=["https://www.example.com"], roles="user")

# Lists all projects
projects = client.list_projects()
for project in projects:
    print(project.name)

# Gets the project with the specified ID
project = client.get_project(project_id=1)
print(project.name)

# Creates a new project with the specified name
project = client.create_project(project_name="New Project", prompt="x", language=IsoLanguage.ENGLISH, llm=LLM.OPENAI)
print(project.name)

# Updates the specified project
project = client.get_project(project_id=1)
project.name = "New Project Name"
response = client.update_project(project)

# Updates the configuration of the specified project
project = client.update_project_configuration(project_id=1, key=ProjectConfigKey.WELCOME_MESSAGE, value="value")

# Asks a question to the specified project
response = client.ask(project_id=1, question="What is the capital of Germany?")
print(response.answer)

# Retrieves documents for the specified project
references = client.retrieve(project_id=1, query="What is the capital of Germany?", retrieval_type=Retriever.default)
for reference in references:
    print(reference.content)

Documentation

The KnowledgeAIClient facilitates seamless communication with the Avvia Intelligence KnowledgeAI service, allowing users to manage projects, upload documents, and perform various other tasks programmatically. It provides an intuitive interface to interact with the KnowledgeAI platform, enabling easy integration into Python applications.

Class: KnowledgeAIClient

Constructor: __init__(url: str, api_key: str, timeout: int = 30) -> None

Initializes a new instance of the KnowledgeAIClient class.

• url (str): The URL of the KnowledgeAI service.
• api_key (str): The API key for authentication.
• timeout (int, optional): The timeout for API requests in seconds. Defaults to 30.

Method: upload_directory(project_id: int, directory: str, roles: str = "user") -> None

Convenience method to upload all files in a directory to the specified project.

• project_id (int): The ID of the project to upload the documents to.
• directory (str): The path to the directory containing the files to upload.
• roles (str, optional): The roles to assign to the uploaded documents. Comma-separated list. Defaults to "user".

Method: index_documents(project_id: int, files: List[str], roles: str = "user") -> None

Uploads multiple documents to the specified project.

• project_id (int): The ID of the project to upload the documents to.
• files (List[str]): A list of file paths to upload.
• roles (str, optional): The roles to assign to the uploaded documents. Comma-separated list. Defaults to "user".

Method: index_urls(project_id: int, index_urls: List[str], roles: Optional[str]) -> None

Indexes the specified URL to the specified project.

• project_id (int): The ID of the project to index the URL to.
• index_urls (List[str]): The URLs to index.
• roles (str, optional): The roles to assign to the indexed document. Comma-separated list. Defaults to "user".

Method: list_projects() -> List[Project]

Lists all projects.

Returns:

• List[Project]: A list of projects.

Method: get_project(project_id: int, with_documents: bool = False) -> Project

Gets the project with the specified ID.

• project_id (int): The ID of the project to get.
• with_documents (bool, optional): Whether to include the documents in the response.

Returns:

• Project: The project with the specified ID.

Method: create_project(project_name: str, prompt: str, language: IsoLanguage = IsoLanguage.ENGLISH, llm: LLM = LLM.OPENAI) -> Project

Creates a new project with the specified name.

• project_name (str): The name of the project to create.
• language (IsoLanguage, optional): The language of the project. Defaults to IsoLanguage.ENGLISH.
• llm (LLM, optional): The language model to use. Defaults to LLM.OPENAI.
• prompt (str, optional): The prompt to use for the project.

Returns:

• Project: The created project.

Method: update_project(project: Project) -> Response

Updates the specified project.

• project (Project): The project to update.

Returns:

• Response: The updated project.

Method: update_project_configuration(project_id: int, key: ProjectConfigKey, value: str) -> Project

Updates the configuration of the specified project.

• project_id (int): The ID of the project to update the configuration for.
• key (str): The key of the configuration to update.
• value (str): The value to update the configuration to.

Returns:

• Project: The updated project.

Method: ask(project_id: int, question: str) -> AskResponse

Asks a question to the specified project.

• project_id (int): The ID of the project to ask the question to.
• question (str): The question to ask.

Returns:

• AskResponse: The response to the question.

Method: retrieve(project_id: int, query: str, retrieval_type: Retriever = Retriever.default) -> List[RetrievedDocument]

Retrieves documents for the specified project.

• project_id (int): The ID of the project to retrieve documents from.
• query (str): The query to retrieve documents for.
• retrieval_type (Retriever, optional): The type of retrieval to use.

Returns:

• List[RetrievedDocument]: The retrieved documents.

Schema Description

Before diving into the usage of KnowledgeAIClient, it's important to understand the schema used in the API. Here are the key classes and enums used:

ProjectConfigKey (Enum)

Defines keys for project configuration settings such as BOTSERVICE_APP_ID, BOTSERVICE_APP_PASSWORD, BOTSERVICE_SECRET, and WELCOME_MESSAGE.

Plan (Enum)

Represents different plans available such as FREE, BASIC, PREMIUM, and ENTERPRISE.

IsoLanguage (Enum)

Defines ISO language codes for languages such as GERMAN and ENGLISH.

LLM (Enum)

Represents different language models like OPENAI, GOOGLE, and ANTHROPIC.

AskResponse (Model)

Represents a response to a question with attributes question, answer, source_paragraphs, and source_documents.

Retriever (Enum)

Defines different types of retrievers such as default, multiquery, and compression.