OllamaReader

Overview

The OllamaReader class is designed to interface with hosted Ollama models to extract relationships between entities in a given text. It extends the BaseLLMReader class and uses the requests library to communicate with the API.

Installation

1. Install the requests library:

   pip install requests
   

Example Usage

Here's an example demonstrating how to use the OllamaReader class:

from euler.llm_readers import OllamaReader

# Initialize the OllamaReader
ollama_reader = OllamaReader(
    base_url="http://localhost:11434", 
    model="qwen2:0.5b",           
    temprature=0.7
)

# Read text and extract relationships
text = """Thomas Alva Edison was an American inventor and businessman. He developed many devices in fields such as electric power generation, mass communication, sound recording, and motion pictures."""
response = ollama_reader.read(text)
print(response.text if response else "No response received.")

Using LLMReaderFactory

The LLMReaderFactory class can be used to easily initialize different LLM readers, including OllamaReader.

from euler.llm_readers import LLMReaderFactory

# Initialize the OllamaReader using LLMReaderFactory
ollama_reader = LLMReaderFactory.get_llm_reader(
    reader_type='ollama',
    base_url="http://localhost:11434", 
    model="qwen2:0.5b",           
    temprature=0.7
)

# Read text and extract relationships
text = """Thomas Alva Edison was an American inventor and businessman. He developed many devices in fields such as electric power generation, mass communication, sound recording, and motion pictures."""
response = ollama_reader.read(text)
print(response.text if response else "No response received.")

Classes and Methods

OllamaConfig

The OllamaConfig class is a configuration model that holds the API base URL, model name, temperature, context window, and an optional default prompt.

Attributes: - base_url (str): The base URL for the hosted model. - model (str): The hosted Ollama model to use. - temprature (float): The sample temperature. - context_window (int): The maximum number of tokens in the context. - default_prompt (Optional[str]): A default prompt to use if no custom prompt is provided.

OllamaResponse

The OllamaResponse class represents the response from the Ollama API.

Attributes: - text (str): The response text generated by the Ollama model.

OllamaReader

The OllamaReader class interfaces with the hosted Ollama models to extract relationships between entities in a given text.

Methods: - __init__(self, base_url: str, model: str, temprature: float = 0.75, default_prompt: Optional[str] = None): Initializes the reader with the base URL, model, temperature, and optional default prompt. - read(self, text: str, custom_prompt: Optional[str] = None) -> Optional[OllamaResponse]: Reads the text and returns the extracted relationships. - stream_complete(self, prompt: str, **kwargs: Any) -> Optional[OllamaResponse]: Sends a request to the Ollama API and processes the streamed response.

This documentation provides an overview, example usage, and detailed description of the OllamaReader class and its associated classes and methods.