hephAIstus 1.2.0

Local AI tools and models for DevOps Engineers

hephAIstus - God of the Forge

hephaistus is a Python package that has a variety of different AI models to assist DevOps Engineers. All of these models run on your local computer and do not get saved online at all. This will allow the user to safely and securely discuss any topic with the AI bot without any potential data leak.

These models use gpt4all and ollama.

Install

To install this package, run pip install hephAIstus.

hephAIstus' Forge

The first AI model is forge() and it uses gpt4all. To use it, simply run the following commands in a Python terminal.

from hephaistus import HephAIstus

heph = HephAIstus()
heph.forge()

This will then start a conversation with an AI bot. You can specify the AI model to use with the model_version parameter in forge. If no model is specified, then it will use the default.

After the conversation is finished, the user can save the results to a text file.

hephAIstus' Hammer

Alternatively, you can use ollama with hammer(). To use it, simply run the following commands in a Python terminal.

from hephaistus import HephAIstus

heph = HephAIstus()
heph.hammer()

This will then start a conversation with an AI bot. You can specify the AI model to use with the model_version parameter in forge. If no model is specified, then it will use the default.

To use one of the provided AI models (found in src/hephaistus/models), run make create_models and all relevant models will be initialized. Alternatively, see the below section (Create Models) about how to do this with HephAIstus.

After the conversation is finished, the user can save the results to a text file.

Create models

There are some models already setup in the src/hephaistus/models folder. These can be build automatically by running the following code.

from hephaistus import HephAIstus

heph = HephAIstus()
heph.create_hammers()

You can then see the available models on your local machine by running heph.list_hammers().

Continue chat

You can continue a previous conversation, from either forge() or hammer(), by running load_save(file_name). This will then load a previously saved conversation to be used by hammer().

Note, forge() does not allow a conversation to be continued like this.