maihem

Introduction

The maihem python package allows you to generate synthetic text data for training and evaluating your LLMs.

Getting Started

Installation

To install the API, run the following command:

pip install maihem

Obtaining your maihem API key

Get a free API key by subscribing to our product here: www.maihem.ai.

Setting API key

Before using the maihem package, you need to set your maihem API key as an environment variable. You can add it to your local bash script, or directly in your python code.

In local bash script

For Linux, open the .bashrc file in your home directory (for MacOs .bash_profile) and add the following line

export MAIHEM_API_KEY = '<your_maihem_api_key>'

Run the following command in the terminal to apply the changes

For Linux

source ~/.bashrc

For Mac

source ~/.bash_profile

In python code

import os

os.environ['MAIHEM_API_KEY'] = '<your_maihem_api_key>'

Generate synthetic data

Persona prompts

See run_examply.py for an example python script for persona prompt generation. The example code is also below

import os
import maihem as mh

os.environ['MAIHEM_API_KEY'] = 'a923c14d881247a7bad58b93d9595494'

# Parameter dictionary for persona
persona = {
    'intent': "credit card got blocked",
    'mood': "angry",
    'age': "30-40",
    'gender': "male",
    'ethnicity': "white",
    'disability': "none",
    'income': "high",
    'education': "college degree",
    'marital_status': "married",
    'children': "2",
    'employment': "employed",
    'housing': "rent",
    'occupation': "banker",
    'location': "New York",
    'sector_name': "retail banking",
    'customer_name': "John Doe",
  }

# Create data generator object
dg = mh.DataGenerator()

# Generate list of prompts for defined persona
data = dg.generate_prompts(persona, model_temperature=0.8, n_calls=3, n_prompts_per_call=2)
print(data)

Metadata

Release files for maihem 1.0.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for maihem 1.0.0
File	Size	Uploaded
maihem-1.0.0.tar.gz	4.4 kB	Dec 14, 2023	Details

Built distribution (wheel)

Table of built distributions (wheels) for maihem 1.0.0
File	Interpreter	ABI	Platform	Reset
maihem-1.0.0-py3-none-any.whl 5.1 kB Dec 14, 2023	Python 3	none	any	Details

Total release size: 9.4 kB

Release files / maihem-1.0.0.tar.gz

Download URL	maihem-1.0.0.tar.gz
Size	4.4 kB
Tags	Source
SHA-256 checksum How to use checksums	`efa07c856eb97f19773621e4ec524e662f88de885142ab1ec65ec9194fcd4fcd`
BLAKE2b-256 checksum How to use checksums	`d1cc3d38f931148ddbffb376fdb2c834d70a23a386d9548aea485d3c191e687b`
Upload date	Dec 14, 2023
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/4.0.2 CPython/3.11.6`

Release files / maihem-1.0.0-py3-none-any.whl

Download URL	maihem-1.0.0-py3-none-any.whl
Size	5.1 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`5ae8dd807ca6983449d56d1ff9be8a1ffb0db8603fe9367f9082ec4070f4379e`
BLAKE2b-256 checksum How to use checksums	`a783ce067d0e4e111afda2c1fc29edea084284673b516958101c00769dd8ddda`
Upload date	Dec 14, 2023
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/4.0.2 CPython/3.11.6`