Skip to content

LLM - Custom Step Generator & Document Studio Flows using Azure OpenAI #204

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 4 commits into
base: main
Choose a base branch
from

Conversation

bteleuca
Copy link

Please include answers to these questions as part of your pull request

In the GitHub webUI, use the Write tab to modify the Markdown text that is part of the pull request. For each question simply place an X inside the square brackets, [X], that represents your answer. Make sure there are no blanks inside the brackets, otherwise MarkDown doesn't render properly. Using the Preview tab while editing this form, you can see the formatted/rendered version of the message.

  • Q1: Confirm that you have the right to submit the code that is being contributed. Please consider the origin of your code and confirm you have the appropriate rights to make the submission subject to the Apache 2.0 license that applies to everything in this repository of custom steps. If so, follow the instructions for the Contributor Agreement (which is based on the industry-standard Developer Certificate of Origin (DCO)).
    • Yes, I have the right to submit the contributed code on behalf of myself, my company, or any other owner of the code. I have also attached my signed copy of the DCO to this message.
    • No
  • Q2: Confirm that your contribution does not include any personally identifiable information (PII), for example, in any examples used in your README file.
    • My contribution does NOT include PII data
    • My contribution includes PII data
  • Q3: Confirm your contribution does not include any encryption or other export-controlled technology.
    • My contribution does NOT contain encryption or other export-controlled technology
    • My contribution includes encryption or other export-controlled technology

ContributorAgreement.Bogdan.Teleuca.txt

Sorry, something went wrong.

I, Bogdan Teleuca <bogdan.teleuca@sas.com>, hereby add my Signed-off-by to this commit: eccafa3

Signed-off-by: Bogdan Teleuca <bogdan.teleuca@sas.com>
@bteleuca
Copy link
Author

Thank you for including this step in the public repo.

@Criptic Criptic self-assigned this Mar 25, 2025
@Criptic Criptic added the new custom step This pull request represents a new custom step label Mar 25, 2025
bteleuca added 2 commits April 9, 2025 15:40
I, Bogdan Teleuca <bogdan.teleuca@sas.com>, hereby add my Signed-off-by to this commit: 795fa3e

Signed-off-by: Bogdan Teleuca <bogdan.teleuca@sas.com>
@bteleuca
Copy link
Author

bteleuca commented Apr 9, 2025

Added a new custom step -> LLM - Document Flows with Azure OpenAI. Thanks for considering this one too.

@snlwih
Copy link
Collaborator

snlwih commented May 6, 2025

@bteleuca , sorry it took me so long to start the review process.

Your steps look great, I had seen your blog entries. I will not have time to try them myself. I am very happy you used self-explanatory names for UI controls instead of accepting the system-generated defaults 😊 This also makes the code behind the step easier to read.

Have a few requests for changing names of folders and steps, and some proposed (small) changes in your code generator.

Also a remark/hint for future contributions. When you have created a pull request and then modify the contents of the branch in your local/fork and then contribute that new step, it will be added to the already open pull request on that branch. Don’t worry, you’re not the only one that has run into that caveat. So when adding multiple new steps, create a dedicated branch for each step and contribute from that branch. That will generate separate and dedicated pull requests, that are easier to review.

Also have a tip for future contributions: When you have created a pull request and then modify the contents of the branch in your local/fork and then contribute that new step, it will be added to the already open pull request on that branch. Don’t worry, you’re not the only one that has run into that caveat. Therefore, when adding multiple new steps, create a dedicated branch for each step and contribute from that branch. That will generate separate and dedicated pull requests. Make things easier to manage and review.

Will add the details of my recommendations for change in a separate comment ...

@snlwih
Copy link
Collaborator

snlwih commented May 6, 2025

General

  • It seems both steps are specific to Azure OpenAI, right? Therefore, use consistent naming across both steps and use something like “… using Azure OpenAI” or “… with Azure OpenAI” for both. Of course, this also means that the root-folder name for each of your steps need to be changed.
  • We have a best practice of including examples and other extras in a dedicated extras subfolder. Could you create such a subfolder and move those files there?
  • Your screenshots should only show the tabs in your custom step when used in standalone mode (which differs from Preview mode). See step DQ - Standardize as an example. Currently your screenshots are showing the complete SAS Studio window.
  • When making these changes, could you please update the version date in both README.md and About tab?

Step specific: LLM - Custom Step Generator

  • Rename to something like: LLM - Custom Step Generator using Azure OpenAI
  • Page: pageOptions
    • Remove introductory text “Generate a SAS Studio custom step …”
  • Code generator
    • It seems your step requires the user to select a SAS Compute path for all the folder/file selections. Please see
      Check SAS Server or SAS Content location for a code snippet on how to approach that. You’re free to create a Python equivalent version of it (which I will then greatly add to the above mentioned page 😊)
    • Perhaps a personal preference, but you could make your code easier to read by calling the UI control something like envfile_folderSelector, and then extract the text after “sasserver:” into a envfile_folder variable?
    • Similar remarks for fileFolderSelector UI controls named messages and output_file. Personally, I would have called those msgfile_fileSelector and output_step_fileSelector or something similar and then use dedicated variables to store the extract the path.

Step specific: LLM - Document Flows with Azure OpenAI

  • Rename to something like: LLM - Document Studio Flows using Azure OpenAI
    • This to make it very clear what type of flows this is all about.
  • Page: pageOptions
    • Remove introductory text “Chose the flow ….”
  • Same suggestions around fileFolderSelector UI control naming and code to extract the path for a file or a folder as for your other custom step

@snlwih snlwih changed the title LLM - Custom Step Generator LLM - Custom Step Generator & Document Studio Flows using Azure OpenAI May 7, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
new custom step This pull request represents a new custom step
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants