WWW::MistralAI

Introduction

This blog post proclaims and describes the Raku package “WWW::MistralAI” that provides access to the machine learning service MistralAI, [MAI1]. For more details of the MistralAI’s API usage see the documentation, [MAI2].

Remark: To use the MistralAI API one has to register and obtain authorization key.

Remark: This Raku package is much “less ambitious” than the official Python package, [MAIp1], developed by MistralAI’s team. Gradually, over time, I expect to add features to the Raku package that correspond to features of [MAIp1].

This package is very similar to the packages “WWW::OpenAI”, [AAp1], and “WWW::PaLM”, [AAp2].

“WWW::MistralAI” can be used with (is integrated with) “LLM::Functions”, [AAp3], and “Jupyter::Chatbook”, [AAp5].

Also, of course, prompts from “LLM::Prompts”, [AAp4], can be used with MistralAI’s functions.

Installation

Package installations from both sources use zef installer (which should be bundled with the “standard” Rakudo installation file.)

To install the package from Zef ecosystem use the shell command:

zef install WWW::MistralAI

To install the package from the GitHub repository use the shell command:

zef install https://github.com/antononcube/Raku-WWW-MistralAI.git

Usage examples

Remark: When the authorization key, auth-key, is specified to be Whatever then the functions mistralai-* attempt to use the env variable MISTRAL_API_KEY.

Universal “front-end”

The package has an universal “front-end” function mistral-playground for the different functionalities provided by MistralAI.

Here is a simple call for a “chat completion”:

use WWW::MistralAI;
mistralai-playground('Where is Roger Rabbit?');

# [{finish_reason => length, index => 0, message => {content => Roger Rabbit is a fictional character from animation. He first appeared in the, role => assistant}}]

Another one using Bulgarian:

mistralai-playground('Колко групи могат да се намерят в този облак от точки.', max-tokens => 300, random-seed => 234232, format => 'values');

# To determine the number of groups in this cluster of points, I would need more information. Traditionally, groups in a set of data points are identified using clustering algorithms. The number of groups identified depends on the specific clustering algorithm used and the parameters set for that algorithm.
# 
# For example, if we use K-Means clustering, we need to specify the number of clusters (K) beforehand. The algorithm will then group the points into K clusters based on their proximity to the centroids. If we don't know how many clusters exist in the data, we can try different values of K and evaluate the results using measures like the Silhouette score.
# 
# Without applying a clustering algorithm to the given set of points or having some prior knowledge about the number of groups, it is not possible to give an answer to this question.

Remark: The functions mistralai-chat-completion or mistralai-completion can be used instead in the examples above. (The latter is synonym of the former.)

Models

The current MistralAI models can be found with the function mistralai-models:

*<id>.say for |mistralai-models;

# mistral-medium
# mistral-small
# mistral-tiny
# mistral-embed

Code generation

There are two types of completions : text and chat. Let us illustrate the differences of their usage by Raku code generation. Here is a text completion:

mistralai-completion(
        'generate Raku code for making a loop over a list',
        max-tokens => 120,
        format => 'values');

# Here's an example of how to make a loop over a list in Raku:
# 
# ```raku
# my @numbers = (1, 2, 3, 4, 5); # create a list of numbers
# 
# for ^@numbers -> $number { # use a for loop to iterate over the list
#     say "$number squared is: ", $number ** 2; # perform an operation on each number
# }
# ```
# 
# In the example above, we use a `for` loop with the `^@

Here is a chat completion:

mistralai-completion(
        'generate Raku code for making a loop over a list',
        max-tokens => 120,
        format => 'values');

# In Raku, you can use the `for` keyword to loop over a list. Here is an example of how to use it:
# 
# ```raku
# my @numbers = (1, 2, 3, 4, 5);
# 
# for ^@numbers -> $number {
#     say "Current number: $number";
# }
# ```
# 
# In this example, we define a list `@numbers` with the elements 1, 2, 3, 4, and 5. The `for` loop

Embeddings

Embeddings can be obtained with the function mistralai-embeddings. Here is an example of finding the embedding vectors for each of the elements of an array of strings:

my @queries = [
    'make a classifier with the method RandomForeset over the data dfTitanic',
    'show precision and accuracy',
    'plot True Positive Rate vs Positive Predictive Value',
    'what is a good meat and potatoes recipe'
];

my $embs = mistralai-embeddings(@queries, format => 'values', method => 'tiny');
$embs.elems;

# 4

Here we show:

  • That the result is an array of four vectors each with length 1536
  • The distributions of the values of each vector
use Data::Reshapers;
use Data::Summarizers;

say "\$embs.elems : { $embs.elems }";
say "\$embs>>.elems : { $embs>>.elems }";
records-summary($embs.kv.Hash.&transpose);

# $embs.elems : 4
# $embs>>.elems : 1024 1024 1024 1024
# +----------------------------------+-----------------------------------+----------------------------------+-----------------------------------+
# | 3                                | 0                                 | 1                                | 2                                 |
# +----------------------------------+-----------------------------------+----------------------------------+-----------------------------------+
# | Min    => -0.088867              | Min    => -0.140381               | Min    => -0.091675              | Min    => -0.126343               |
# | 1st-Qu => -0.022064              | 1st-Qu => -0.019684               | 1st-Qu => -0.021881              | 1st-Qu => -0.021805               |
# | Mean   => -0.0014019259251654148 | Mean   => 0.0001275218091905117   | Mean   => -0.0014658444561064243 | Mean   => -0.00016893696738407016 |
# | Median => -0.0015511512756347656 | Median => -0.00012609362602233888 | Median => -0.0017815             | Median => 0.0005223751068115234   |
# | 3rd-Qu => 0.019516               | 3rd-Qu => 0.021347                | 3rd-Qu => 0.019943               | 3rd-Qu => 0.021469                |
# | Max    => 0.10968                | Max    => 0.088867                | Max    => 0.111755               | Max    => 0.097534                |
# +----------------------------------+-----------------------------------+----------------------------------+-----------------------------------+

Here we find the corresponding dot products and (cross-)tabulate them:

use Data::Reshapers;
use Data::Summarizers;
my @ct = (^$embs.elems X ^$embs.elems).map({ %( i => $_[0], j => $_[1], dot => sum($embs[$_[0]] >>*<< $embs[$_[1]])) }).Array;

say to-pretty-table(cross-tabulate(@ct, 'i', 'j', 'dot'), field-names => (^$embs.elems)>>.Str);

# +---+----------+----------+----------+----------+
# |   |    0     |    1     |    2     |    3     |
# +---+----------+----------+----------+----------+
# | 0 | 1.000405 | 0.580262 | 0.736048 | 0.546892 |
# | 1 | 0.580262 | 1.000174 | 0.663682 | 0.534618 |
# | 2 | 0.736048 | 0.663682 | 1.000578 | 0.545032 |
# | 3 | 0.546892 | 0.534618 | 0.545032 | 0.999975 |
# +---+----------+----------+----------+----------+

Remark: Note that the fourth element (the cooking recipe request) is an outlier. (Judging by the table with dot products.)

Chat completions with engineered prompts

Here is a prompt for “emojification” (see the Wolfram Prompt Repository entry “Emojify”):

my $preEmojify = q:to/END/;
Rewrite the following text and convert some of it into emojis.
The emojis are all related to whatever is in the text.
Keep a lot of the text, but convert key words into emojis.
Do not modify the text except to add emoji.
Respond only with the modified text, do not include any summary or explanation.
Do not respond with only emoji, most of the text should remain as normal words.
END

# Rewrite the following text and convert some of it into emojis.
# The emojis are all related to whatever is in the text.
# Keep a lot of the text, but convert key words into emojis.
# Do not modify the text except to add emoji.
# Respond only with the modified text, do not include any summary or explanation.
# Do not respond with only emoji, most of the text should remain as normal words.

Here is an example of chat completion with emojification:

mistralai-chat-completion([ system => $preEmojify, user => 'Python sucks, Raku rocks, and Perl is annoying'], max-tokens => 200, format => 'values')

# 😠 Python sucks, 🌈 Raku rocks, 😓 Perl is annoying

Command Line Interface

Playground access

The package provides a Command Line Interface (CLI) script:

mistralai-playground --help

# Usage:
#   mistralai-playground [<words> ...] [--path=<Str>] [--mt|--max-tokens[=UInt]] [-m|--model=<Str>] [-r|--role=<Str>] [-t|--temperature[=Real]] [--response-format=<Str>] [-a|--auth-key=<Str>] [--timeout[=UInt]] [-f|--format=<Str>] [--method=<Str>] -- Command given as a sequence of words.
#   
#     --path=<Str>                Path, one of 'chat/completions', 'images/generations', 'images/edits', 'images/variations', 'moderations', 'audio/transcriptions', 'audio/translations', 'embeddings', or 'models'. [default: 'chat/completions']
#     --mt|--max-tokens[=UInt]    The maximum number of tokens to generate in the completion. [default: 100]
#     -m|--model=<Str>            Model. [default: 'Whatever']
#     -r|--role=<Str>             Role. [default: 'user']
#     -t|--temperature[=Real]     Temperature. [default: 0.7]
#     --response-format=<Str>     The format in which the response is returned. [default: 'url']
#     -a|--auth-key=<Str>         Authorization key (to use MistralAI API.) [default: 'Whatever']
#     --timeout[=UInt]            Timeout. [default: 10]
#     -f|--format=<Str>           Format of the result; one of "json", "hash", "values", or "Whatever". [default: 'Whatever']
#     --method=<Str>              Method for the HTTP POST query; one of "tiny" or "curl". [default: 'tiny']

Remark: When the authorization key argument “auth-key” is specified set to “Whatever” then mistralai-playground attempts to use the env variable MISTRAL_API_KEY.

Mermaid diagram

The following flowchart corresponds to the steps in the package function mistralai-playground:

References

Packages

[AAp1] Anton Antonov, WWW::OpenAI Raku package, (2023), GitHub/antononcube.

[AAp2] Anton Antonov, WWW::PaLM Raku package, (2023), GitHub/antononcube.

[AAp3] Anton Antonov, LLM::Functions Raku package, (2023), GitHub/antononcube.

[AAp4] Anton Antonov, LLM::Prompts Raku package, (2023), GitHub/antononcube.

[AAp5] Anton Antonov, Jupyter::Chatbook Raku package, (2023), GitHub/antononcube.

[MAI1] MistralAI, MistralAI platform.

[MAI2] MistralAI Platform documentation, MistralAI documentation.

[MAIp1] MistralAI, [https://github.com/mistralai/client-python), (2023), GitHub/mistralai.

One thought on “WWW::MistralAI”

  1. Pingback: 2023.52 Final Magic – Rakudo Weekly News ~

Leave a comment