bcg-x-official / artkit Goto Github PK

We plan to add an optional progress bar to the run method in the dependency, fluxus: BCG-X-Official/fluxus#23

It will then be available to the run method in artkit and we should introduce it in the "Building Your First ARTKIT Pipeline" tutorial, specifically in the section where we introduce the core ARTKIT functions.

Once the progress bar is added and we have updated the fluxus version in artkit, we will re-open this issue.

Google's Gemini supports text-to-image, but currently we only have a text-to-text connector. We should support text-to-image.

Details

The new class will live here: src/artkit/model/diffusion/gemini/

Be sure to update the docs:

• Connecting to Gen AI Models: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb

Please see the tutorial: Creating New Model Classes

Google's Gemini supports image-to-text, but currently we only have a text-to-text connector. We should support image-to-text.

Details

The new class will live here: src/artkit/model/vision/gemini/

Be sure to update the docs:

• Connecting to Gen AI Models: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb

Please see the tutorial: Creating New Model Classes

Issue

Currently, the README.rst file does not contain any displayed badges that represent the repository. Some are in pypi_description.rst but don't have full list (see below)

Additional details on the .rst file format are located here

Specifically, we'd like to add badges for the following:

• pypi
• anaconda version
• python versions
• code style (black)
• sphinx (doc build)
• license (apache)
• GitHub Action build status
• Contributor Convenant

Here's an example with all badges we want to replicate. Note: the example uses Azure DevOps to add the test coverage badge but this repo utilizes GitHub Actions so a different approach is required and will entail changes to the artkit-release-pipeline.yml. Given it's a more complex change, a separate issue documents the approach:

Solution

• Update Readme.rst and pypi_description.rst with the relevant badges
  • Add the following where the badges should be displayed (example)

.. Begin-Badges

|pypi| |conda| |python_versions| |code_style| |made_with_sphinx_doc| |license_badge| |Contributor_Convenant|

.. End-Badges

• • Add the targets for the URLs where the badges should point to (example)

.. Begin-Badges

.. |pypi| image:: https://badge.fury.io/py/artkit.svg
    :target: https://pypi.org/project/artkit/

.. |conda| image:: https://anaconda.org/bcg_gamma/gamma-facet/badges/version.svg
    :target: https://anaconda.org/BCG_Gamma/artkit

.. |python_versions| image:: https://img.shields.io/badge/python-3.10|3.11|3.12-blue.svg
   :target: https://www.python.org/downloads/release/python-3100/

.. |code_style| image:: https://img.shields.io/badge/code%20style-black-000000.svg
   :target: https://github.com/psf/black

.. |made_with_sphinx_doc| image:: https://img.shields.io/badge/Made%20with-Sphinx-1f425f.svg
   :target: https://bcg-x-official.github.io/facet/index.html

.. |license_badge| image:: https://img.shields.io/badge/License-Apache%202.0-olivegreen.svg
   :target: https://opensource.org/licenses/Apache-2.0

.. image:: https://github.com/BCG-X-Official/artkit/actions/workflows/artkit-release-pipeline.yml/badge.svg
    :target: https://github.com/BCG-X-Official/artkit/actions/workflows/artkit-release-pipeline.yml
    :alt: ARTKIT Release Pipeline

.. |Contributor_Convenant| image:: https://img.shields.io/badge/Contributor%20Covenant-2.1-4baaaa.svg 
   :target: CODE_OF_CONDUCT.md

.. End-Badges

• Update the RELEASE_NOTES.rst file:
  • Increment a minor version here

Problem

The tutorials in our Sphinx documentation are HTML pages built from Jupyter notebooks. Readers may want to download the Jupyter notebook so they can run the code, but it isn't clear where they can find the notebook.

Solution

Within each notebook tutorial, include a hyperlink to the Jupyter notebook on GitHub. From GitHub, the user can choose to download the notebook and explore related files (e.g., data or cache files which are used by the notebook).

Details

Place hyperlinks just beneath the main title of notebooks, e.g.:

# Building Your First ARTKIT Pipeline

[View notebook on GitHub](https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/building_your_first_artkit_pipeline.ipynb)

Add hyperlinks to all Jupyter notebooks which are included in the User Guide and Examples sections of the Sphinx documentation.

In the artkit GitHub repo, you can find these files within:

• sphinx/source/user_guide/
• sphinx/source/examples/

Verifying your changes

• Double check that you included correct hyperlinks for all notebook tutorials in the User Guide and Examples
• Note that some docs pages are built from RST files instead of notebooks (e.g., the index pages for the User Guide and Examples, the intro to Gen AI Testing and Evaluation in the User Guide). These pages do not need hyperlinks.

Issue

Currently, CacheDB.clear allows users to clear cached results which were accessed or created before a given datetime:

• https://github.com/BCG-X-Official/artkit/blob/1.0.x/src/artkit/model/cache/_cache.py#L369

However, it is not possible to clear results after a given datetime. This would be useful when I make a mistake midway through an analysis and I want to clear only the most recent results from my cache.

Solution

• Update the CacheDB class:
  • Add accessed_after and created_after optional arguments to the CacheDB.clear method
  • Update the dynamic SQL query accordingly
• Update the CachedGenAIModel class (uses CacheDB):
  • Add the accessed_after and created_after optional arguments to the CachedGenAIModel.clear_cache method

Details

Once implemented, be sure to:

• Incorporate tests for the new arguments in test_cache_clearing
• Mention this capability in the Cache Management tutorial

Describe the bug

A clear and concise description of what the bug is.

To Reproduce

Steps to reproduce the behavior:

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

Expected behavior

A clear and concise description of what you expected to happen.

Screenshots

If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: Mac/Windows
• Version [e.g. 22]

Additional context

Add any other context about the problem here.

Justification

Amazon Bedrock provides an extensive suite of Gen AI models via a single API. Supporting Amazon Bedrock will significantly enhance ARTKIT's versatility.

Details

Note that the Bedrock API supports multiple modalities, depending on the selected model. This issue is for LLMs. The new class will live in src/artkit/model/llm/aws/

Be sure to update the docs:

• README: https://github.com/BCG-X-Official/artkit?tab=readme-ov-file#supported-model-providers
• Contributing: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb
• Connecting to Gen AI Models: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb
• .env_example: https://github.com/BCG-X-Official/artkit/blob/1.0.x/.env_example

Please see the tutorial: Creating New Model Classes

As the https://bcg-x-official.github.io/artkit/user_guide/introduction_to_artkit/connecting_to_genai_models.html#Google link is empty with no connection result.

Justification

Amazon Bedrock provides an extensive suite of Gen AI models via a single API. Supporting Amazon Bedrock will significantly enhance ARTKIT's versatility.

Details

Note that the Bedrock API supports multiple modalities, depending on the selected model. This issue is for text-to-image diffusion models. The new class will live in src/artkit/model/diffusion/aws/

Be sure to update the docs:

• README: https://github.com/BCG-X-Official/artkit?tab=readme-ov-file#supported-model-providers
• Contributing: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb
• Connecting to Gen AI Models: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb
• .env_example: https://github.com/BCG-X-Official/artkit/blob/1.0.x/.env_example

Please see the tutorial: Creating New Model Classes

Justification

Cohere focuses on providing low-latency LLMs. They have an async endpoint we can use and free API keys.

Details

The new class will live in src/artkit/model/llm/cohere/

Be sure to update the docs:

• README: https://github.com/BCG-X-Official/artkit?tab=readme-ov-file#supported-model-providers
• Contributing: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb
• Connecting to Gen AI Models: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb
• .env_example: https://github.com/BCG-X-Official/artkit/blob/1.0.x/.env_example

Please see the tutorial: Creating New Model Classes

Justification

The Azure OpenAI API is compatible with OpenAI's API. With Azure OpenAI, one can set up their own deployments of the common GPT and Codex models. When calling the API, they need to specify the deployment they want to use.

Details

The new class will live in src/artkit/model/llm/azureopenai/

Be sure to update the docs:

• README: https://github.com/BCG-X-Official/artkit?tab=readme-ov-file#supported-model-providers
• Contributing: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb
• Connecting to Gen AI Models: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb
• .env_example: https://github.com/BCG-X-Official/artkit/blob/1.0.x/.env_example

Please see the tutorial: Creating New Model Classes

Describe the bug

A clear and concise description of what the bug is.

To Reproduce

Steps to reproduce the behavior:

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

Expected behavior

A clear and concise description of what you expected to happen.

Screenshots

If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: Mac/Windows
• Version [e.g. 22]

Additional context

Add any other context about the problem here.

Describe the bug

A clear and concise description of what the bug is.

To Reproduce

Steps to reproduce the behavior:

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

Expected behavior

A clear and concise description of what you expected to happen.

Screenshots

If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: Mac/Windows
• Version [e.g. 22]

Additional context

Add any other context about the problem here.

Describe the bug

A clear and concise description of what the bug is.

To Reproduce

Steps to reproduce the behavior:

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

Expected behavior

A clear and concise description of what you expected to happen.

Screenshots

If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: Mac/Windows
• Version [e.g. 22]

Additional context

Add any other context about the problem here.

Is your feature request related to a problem? Please describe.

Previously, classes in model.llm.base like ChatModel where reachable through artkit.api and its been popular to use ak.ChatModel for type notations in notebooks. However, this is no longer possible as from .model.llm.base import * is missing from api.py

Describe the solution you'd like

Selectively include classes from .base modules such as ChatModel : from .model.llm.base import ChatModel

Describe alternatives you've considered

Alternative is to import all packages from base, which might be too much

Describe the bug

artkit.api does not import CachedDiffusionModel and CachedVisionModel.

To Reproduce

Steps to reproduce the behavior:

1. Run the Connecting to Gen AI Models notebook: https://github.com/BCG-X-Official/artkit/blob/1.0.x/sphinx/source/user_guide/introduction_to_artkit/connecting_to_genai_models.ipynb
2. The OpenAI "text-to-image" and "vision" sections will fail

Expected behavior

CachedDiffusionModel and CachedVisionModel should be imported with artkit.api

Additional context

The fix is to update the following imports in src/artkit/api.py, which currently do not seem to import anything:

from .model.diffusion import *
from .model.vision import *

to:

from .model.diffusion.base import *
from .model.vision.base import *

Please can you provide few documentation regarding the same and the gemini url that is being used.

Is your feature request related to a problem? Please describe.

The only connectors that exist currently are for existing models with clearly defined APIs. There is no way for a user to specify any HTTP endpoint to get a response.

Describe the solution you'd like

I want to be able to use ARTKIT with any endpoint that accepts a payload.

Describe alternatives you've considered

None

Additional context

It seems like a version of the existing HuggingFace connector could be leveraged