decenter-ai/decenter-ai.streamlit.app

Decenteralized AI training platform for all

decenter-ai.streamlit.app's Issues

add: tensorflow to ml deps

task: add ruff

https://github.com/astral-sh/ruff

• Implement ruff
• Install vscode extensions
• Add ruff to readme.md: https://github.com/astral-sh/ruff#show-your-support

beta-test: streamlit.app

• beta test https://decenter-ai.streamlit.app
• add good demos to samples

drag: drop datasets

• show datasets in left sidebar after v1, v2.....
• support drag and drop of datasets.
• allow users to drag and drop the datasets from left sidebar to

• add scrolling bar to datasets to allow more datasets
• start with 2 datasets in the sidebar

proposal: security: chroot before executing scripts

Proposal:

For security, leverage chroot to prevent unauthorized access of train script on other directories

Implementation Plan

Outline the steps or phases required to implement the proposal.

1.Copy venv dir to temp_dir
2.chroot
3. Make the system call

Additional Information

Include any additional information or references that support the proposal.

proposal: Update the pr template

Proposal:

Tidy up the pr template

add: robot E2E tests

• Write new robot tests for the demo model train
• Write tests that cover v1, v2

beta-test: decenter-ai.streamlit.app v3 with modified kaggle nb

1. Report bugs
2. Add 5+ different examples to examples dir

fix: UI size glitch

Fix: it

The app should be in the center. Fix: https://decenter-ai.streamlit.app

proposal: proposal: feat: add "Request a feature" button

Summary

Add "Request a feature" button on the top right corner next to "Report a bug"

Background

We are currently expanding and concerned about the stability of our released application. We want to leverage our beta testers and public to continuously develop and release major and minor versions.

Proposed Solution

• Add "Request a feature" button on the top right corner

Implementation Plan

@Dineshjnld

1. Add a "Request a Feature" button with the following styling:

Button text: "Request a Feature"
Background color: #751919 (Red)
Text color: #ffffff (White)
2. Apply a hover effect to the button:

On hover, transition the background color to #2B3A4F (Darker color).

tensorflow: is not working

Bug description

poetry add tensorflow

How to reproduce

Expected behavior

No response

proposal: chroot to the data_dir in system call for running the training script to prevent unauthorized access of files by training script

proposal: chroot to the data_dir in system call for running the training script to prevent unauthorized access of files by training script

retreive model name from archive name

if the model name was not changed:

then the model name should be updated when archive is uploaded as decenter-model-wineequalizer

bug: fix sample outputs

Bug Description:

The output files names in samples/kaggle/outputs doesn't convey the corresponding inputs.

Versions Affected:

• v3
• v2
• v1

use : assert

Use assert for check of train script

https://blog.streamlit.io/building-robust-streamlit-apps-with-type-checking/

assert train_script is not None

instead of if not train_script: st.stop()

CI: robot tests

update:

• create .github/workflows/automated-robot-tests.yml
• refer: https://github.com/marketplace/actions/robot-framework or https://github.com/marketplace/actions/robot-framework-action
• implement them
• make sure the below code is in the workflow (.yml) file

on:
  pull_request:
  push: 

chore: refactor the variables to app instance

Implementation Plan

Outline the steps or phases required to implement the proposal.

1. git fetch origin
2. git checkout refactor/app-instance
3. GO to L69. Rest is left

Review all demos

• reviiew all demos in https://decenter-ai.streamlit.app
• fix the demos

decenter-compute: lilypad module

bug: V1--while uploading requirements.txt

Bug Description:

in V1 , after uploading the requirements.txt file an error occur " st.experimental_memo is deprecated. Please use the new command st.cache_data instead, which has the same behavior. More information in our docs."

Versions Affected:

• [] v3
• v2
  - [ x] v1

Steps to Reproduce:

1. uploading csv file, .py file is fine, but when we upload the requirements.txt file , the error message appears
   

Expected Behavior:

[Describe what you expected to happen when performing the steps above.]

Actual Behavior:

[Describe what actually happened when performing the steps above.]

Additional Information:

[Include any additional information that may be helpful in understanding and resolving the bug. This may include screenshots, code snippets, or logs.]

Environment:

• [] Decenter-AI
• [ x] Local
• Public Hosted

Ignore the following for Decenter-AI

• Python Version:
• Operating System:

Relevant Code Snippet:

install ml deps poetry: tensorflow

regex
tqdm
Pillow
tensorflow
tensorflow_addons

poetry add $dep --group ml

Research: on celery and how we can leverage parallel processing

https://docs.celeryq.dev/en/stable/getting-started/first-steps-with-celery.html

decenter-compute: fvm contracts

bug: fix the dupllicate rerendering of head() and toast, model updated

Bug Description:

[Describe the bug in detail here. Include any error messages, unexpected behavior, or other relevant information.]

Versions Affected:

• v3
• v2
• v1

Steps to Reproduce:

1. Train an manually inputed model

2. Wait for execution to complete
   Expected Behavior:

3. You see Welcome to Decenter Again

4. And entire rerendering of head() due to experimental rerender

Actual Behavior:

1. Don't rerender toasts, alerts

Additional Information:

[Include any additional information that may be helpful in understanding and resolving the bug. This may include screenshots, code snippets, or logs.]

Environment:

• Decenter-AI
• Local
• Public Hosted

Ignore the following for Decenter-AI

• Python Version: 3.10.0
• Operating System: Macos/Linux/Macos

Relevant Code Snippet:

proposal:switch page among versions

from streamlit_extras.switch_page_button import switch_page

if st.button("Page2"):
    switch_page("Page2")

Use this code to switch between versions

Checkout documentation in: https://extras.streamlit.app

Implement automated testing for our app

Refer: https://blog.devgenius.io/testing-streamlit-a1f1fd48ce8f

Req:

Wait Until Page Contains Test
Page Should Contain You selected v3.
Click Element //*[contains(text(),'v2','v1')]
sleep 1
Page Should Contain You didn't select comedy.
Close Browser

proposal: demo dropdown

Proposal:

Add a demo dropdown at top right

1. Choose b/ any kaggle input archives (dynamics, u should os.listdir() all the zip files in kaggle/inputs directory)

2. When choosing one, u should be able to inject the archive.

decenter-compute: bacalhau job

proposal: Refactor python_repl,venv_dir, starter_script

Proposal:

in the tite

Implementation:

similar to #44

proposal: create chore issue template

Proposal:

Summary

Create a simplified flexible chore issue template

Problem Statement

Clearly state the problem that the proposal aims to address.

Benefits

Explain the benefits and potential impact of implementing the proposal.

Implementation Plan

Outline the steps or phases required to implement the proposal.

Additional Information

Include any additional information or references that support the proposal.

hotfix: all demos not working

Bug description

• remove all models that not working

How to reproduce

Expected behavior

No response

Bug: duplicate decenter in archive name when downloaded; Invalid model name

Duplicate decenter in the downloaded archive name

Possible Solution:

• Try to validate model names: ensure its suffixed with decenter once.

Cuz u r just using model name in the downloaded file

chore: update the model archives

• ensure that archive names match the training script: headbrain.zip->"headbrain.ipynb"
• Add 5 more demos. Including a python script demo.
• Add previous projects
• Max 15 demos

https://github.com/ypeleg/llama#llama-simple-using-huggingface

Replace the robot on the streamlit app with the one of the robots on the landing page

proposal: feat: add report bug button

Summary

Add "Report an Issue" button on the top right corner

Background

We are currently expanding and concerned about the stability of our released application. We want to leverage our beta testers and public to continuously release patch versions.

Proposed Solution

• Add "Report an Issue" button on the top right corner

Implementation Plan

@Dineshjnld

1. Add a "Report an issue" Red button
2. Styling: color: #751919. on hover it should transition to #2B3A4F
3. Onclick: it should open this is link in the new tab: https://github.com/DeCenter-AI/decenter-ai.streamlit.app/issues/new?assignees=&labels=bug&projects=&template=00-bug.md&title=bug%3A++

create a navigation drop box for versions after header

Update:

Overview:

Dropdown for versions

https://docs.streamlit.io/library/api-reference/widgets/st.selectbox

task: implement Multi page config

• Checkout https://st-pages-sections.streamlit.app
• use code based approach to reorder and customize v2, v1....

Review Latest Customer Relationship Management (CRM) developments

Review: labels: https://github.com/DeCenter-AI/decenter-ai.streamlit.app/labels

Review:

New-Issues Flows: https://github.com/DeCenter-AI/decenter-ai.streamlit.app/issues/new/choose

Review: https://github.com/DeCenter-AI/decenter-ai.streamlit.app/discussions/new/choose

Review: https://github.com/DeCenter-AI/decenter-ai.streamlit.app/discussions/new?category=q-a

proposal: Add stdout, stderr into outputs zip

Proposal:

Add the output file in output zip folder

Implementation Plan

Outline the steps or phases required to implement the proposal.

1. Currently we are using st.warning to log the stderr of the subprocess call
2. Additonally we need to write the stderr with file name "stderr" in app.work_dir
3. Also write the stdout with the file name "stdout" in app.work_dir

Additional Information

bug: Bug: demo mode always off

Bug Description:

[Describe the bug in detail here. Include any error messages, unexpected behavior, or other relevant information.]

Versions Affected:

• v3
• v2
• v1

Steps to Reproduce:

1. Open decenter-ai.streamlit.app

Expected Behavior:

Should allow train, and training script should be visibile

Actual Behavior:

Additional Information:

[Include any additional information that may be helpful in understanding and resolving the bug. This may include screenshots, code snippets, or logs.]

Environment:

• Decenter-AI
• Local
• Public Hosted

Ignore the following for Decenter-AI

• Python Version:
• Operating System:

Relevant Code Snippet:

python-nb-execution: migrate from sys.executable to virtual env