image-text-structurizer

A Python package to process user-provided text descriptions of images and extract structured, validated outputs using pattern matching and a language model. Ensures that the output adheres to a predefined format such as XML-like tags, facilitating consistent and reliable extraction of image metadata or summaries without directly handling image data.

Installation

Install via pip:

pip install image_text_structurizer

Usage Example

from image_text_structurizer import image_text_structurizer

response = image_text_structurizer(user_input="A scenic landscape with mountains and a river")
print(response)

Function Parameters

• user_input (str): The text description of the image to process.
• llm (Optional[BaseChatModel]): An instance of a language model. Defaults to ChatLLM7 from langchain_llm7.
• api_key (Optional[str]): API key for accessing the ChatLLM7 model. If not provided, it attempts to fetch from environment variable LLM7_API_KEY. You can also set it directly in code.

Supported Language Models

This package uses ChatLLM7 by default, which can be imported from langchain_llm7. Developers can pass other models, such as:

• OpenAI Chat models
• Anthropic models
• Google Generative AI models

by creating an instance and passing it to the image_text_structurizer function.

Examples:

Using OpenAI:

from langchain_openai import ChatOpenAI
from image_text_structurizer import image_text_structurizer

llm = ChatOpenAI()
response = image_text_structurizer(user_input="A sunset over the ocean", llm=llm)

Using Anthropic:

from langchain_anthropic import ChatAnthropic
from image_text_structurizer import image_text_structurizer

llm = ChatAnthropic()
response = image_text_structurizer(user_input="A forest with tall trees and fog", llm=llm)

Using Google Generative AI:

from langchain_google_genai import ChatGoogleGenerativeAI
from image_text_structurizer import image_text_structurizer

llm = ChatGoogleGenerativeAI()
response = image_text_structurizer(user_input="A city skyline at night", llm=llm)

Notes

• The package relies on the ChatLLM7 model from langchain_llm7.
• Default rate limits are suitable for most use cases, but higher limits can be obtained by registering for an API key at https://token.llm7.io/.
• To pass your own API key, set the environment variable LLM7_API_KEY or pass it directly:

response = image_text_structurizer(user_input="Example text", api_key="your_api_key")

Support & Contributions

• For issues, open an issue on the GitHub repository.

Author

• Eugene Evstafev
  Email: hi@eugene.plus
  GitHub: https://github.com/chigwell