-
Notifications
You must be signed in to change notification settings - Fork 518
Pending Changes for Active Stack #3372
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
base: develop
Are you sure you want to change the base?
Pending Changes for Active Stack #3372
Conversation
Important Review skippedAuto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the You can disable this status message by setting the Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
Documentation and Community
|
eb058fb
to
cede201
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There is one major part missing here still: When running/building a pipeline for which this config attribute is set, we actually need to use the temporary_active_stack
context manager to switch the stack for the run/build.
This will be in PipelineDefinion.build(...)
and PipelineDefinition._run(...)
.
@@ -46,6 +47,7 @@ class PipelineRunConfiguration( | |||
steps: Dict[str, StepConfigurationUpdate] = {} | |||
settings: Dict[str, SerializeAsAny[BaseSettings]] = {} | |||
tags: Optional[List[str]] = None | |||
stack: Optional[Stack] = None |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This should be an Optional[str]
instead. The Stack
object is not something that can be defined in a yaml file, which is usually how users specify attributes of this run configuration class. The value for stack
will be the name or ID of a stack, which will then be passed to temporary_active_stack
.
pipeline_id=pipeline_id, | ||
code_repository=code_repository, | ||
) | ||
with temporary_active_stack(stack_name_or_id=run_config.stack): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Read the value of stack from the PipelineRunConfiguration
and if available use that within the temporary_active_stack
context manager. Same below.
@schustmi Hi Michael, Thank you for the feedback on PR. I have processed through code and tried to address your comment to best of my understanding. Let me know if this sounds good! |
25a491e
to
0d059d3
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
One thing I forgot: Could you also add this logic to temporarily set a different stack on the PipelineDefinition.create_run_template
method?
As for testing this:
- The test should first register a second stack (should probably just be default local orchestrator/artifact store)
- The active stack should be the default one
- We then configure any pipeline (just a single-step, do nothing pipeline should be fine) to run on a different stack.
- The above bullet point can be done either via a config file, or by passing it directly to
test_pipeline.with_options(stack=...)
. These should probably be two different test cases. - It should then run the pipeline
run = test_pipeline()
, and check whether the run ran on the configured stack.assert run.stack.id == ...
- We can then also check that the stack gets correctly reset to the previously active stack:
assert Client().active_stack.id == ...
I think examples for most of these can be found in tests/unit/pipelines/test_base_pipeline.py
(e.g. running a pipeline, configuring from a file). Let me know if I can help in any other way
local_repo = code_repository_utils.find_active_code_repository() | ||
code_repository = build_utils.verify_local_repository_context( | ||
deployment=deployment, local_repo_context=local_repo | ||
_from_config_file = self._parse_config_file( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not a huge fan of you having to do this manually before the compilation step in many places, but I also don't see any easy way around this. I'll refactor this (the entire config file logic is very messy right now) after this PR is merged.
When zenml/src/zenml/cli/pipeline.py Line 363 in 0d059d3
I can also add a context manager but I believe this case is already handled. |
Yep, but that is only available when using the CLI to create a run template, but not when doing this from purely in code. I think the better solution would be to remove the context manager from the CLI function, and instead pass the |
@schustmi Added some unit tests as per your comment. When I ran locally, I did not see issues related to the tests I added. I don't think unit tests are auto-executed by CI here. I executed tests on my branch using: |
One last question. I did pass the
Here the create_run_template, but it is not possible (or required) to patch the However, zenml/src/zenml/pipelines/pipeline_definition.py Line 1053 in 271076b
Any suggestions? |
@schustmi Friendly ping for the clarification :) |
@schustmi It's been a while! Any feedback on the comment above? |
* Update issue template configurations and URLs * Auto-update of LLM Finetuning template * Update .github/ISSUE_TEMPLATE/feature_report.yml Co-authored-by: Alex Strick van Linschoten <strickvl@users.noreply.github.com> * Update .github/ISSUE_TEMPLATE/feature_report.yml Co-authored-by: Alex Strick van Linschoten <strickvl@users.noreply.github.com> --------- Co-authored-by: GitHub Actions <actions@github.com> Co-authored-by: Alex Strick van Linschoten <strickvl@users.noreply.github.com> (cherry picked from commit b098798)
@preyasshah9 Sorry for the late comment, I lost track of this among so many ongoing things. The code in the In general, the |
…zenml-io#3796) * Created service accounts with least privilege permissions * Update Vertex AI and AzureML permissions for security * Add new GCP permission for accessing GCS artifacts * Add GetLogEvents permission to custom SageMaker policy (cherry picked from commit d2068df)
* Update best practices for upgrading ZenML servers * Apply suggestions from code review Co-authored-by: Alex Strick van Linschoten <strickvl@users.noreply.github.com> * Update best practices for ZenML server upgrades * Update run templates in production for ZenML Pro * Add rescheduling pipelines for all users in UAT phase * Recreate run templates in production using code patterns * Update docs/book/how-to/manage-zenml-server/best-practices-upgrading-zenml.md Co-authored-by: Alexej Penner <thealexejpenner@gmail.com> * Update docs/book/how-to/manage-zenml-server/best-practices-upgrading-zenml.md Co-authored-by: Alexej Penner <thealexejpenner@gmail.com> * Add server migration steps for upgrading ZenML * Update best practices for upgrading ZenML server * Update ZenML upgrade best practices for smooth team workflow * Added assets * Update diagrams and images for best practices * Optimised images with calibre/image-actions * Added assets * Optimised images with calibre/image-actions * Update best practices steps with "Step" instead of "Rule". * Update image placeholder in best practices guide --------- Co-authored-by: Alex Strick van Linschoten <strickvl@users.noreply.github.com> Co-authored-by: Alexej Penner <thealexejpenner@gmail.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> (cherry picked from commit 470cbba)
* initial * more * interim * Update README.md to enhance clarity and detail on ZenML's capabilities - Revised section titles for improved readability, changing "Stack-agnostic: keep your code the same" to "Write Code Once, Run Anywhere" and "Built for platform teams" to "Automatic Versioning for MLOps and LLMOps". - Expanded descriptions to better articulate ZenML's functionality, including examples of code snippets for pipelines and steps. - Added visual ideas to illustrate key features and workflows, enhancing the documentation's engagement and usability. - Included new sections on integrating tools and building production-grade AI workflows, showcasing ZenML's versatility in MLOps and LLMOps contexts. * Revise README.md to clarify ZenML's role in production AI - Updated the section title from "The Hidden Complexity of Production AI" to "Beyond the Demo: The Reality of Production AI" to better reflect the content. - Enhanced the description of challenges faced when transitioning from POC to production, emphasising the complexities of evaluation pipelines, document processing, cost tracking, and compliance. - Added specific target audiences for ZenML, highlighting its benefits for platform teams, enterprises, and organisations in regulated industries. - Removed outdated sections to streamline the document and improve focus on ZenML's orchestration capabilities. * Update README.md to enhance clarity and functionality of ZenML pipelines - Modified the `my_training_pipeline` function to return the trained model, improving the pipeline's utility. - Updated the description of automatic versioning to include code formatting for `scikit-learn`, enhancing readability. - Replaced the `Client().active_step_run.log_metadata` call with `log_metadata` for logging evaluation costs, streamlining the logging process. - Renamed the `training_pipeline_with_integrations` to `llm_finetuning_pipeline` for better alignment with its purpose. - Added an `on_failure` hook to the `llm_trainer` step, improving error handling in the pipeline. - Ensured consistent formatting and clarity throughout the README for better user understanding. * update README(s) * More README updates * README improvements * remove zenml go * Small update for LLM support * update header * use other version's line * improve secrets section * Remove excess * Remove decision maker section * Footer changes * Update README * Optimised images with calibre/image-actions * Update the use case example so that it all connects * Update README with improved MLOps messaging * Add version 0.83.1 release notes to README * Update latest release link in README.md * Update README with new content about AI agent development * Delete compliance, RAG pipeline, architecture, and orchestrator images * Add images for problem and development lifecycle * Optimised images with calibre/image-actions --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Hamza Tahir <htahir111@gmail.com> Co-authored-by: Hamza Tahir <hamza@zenml.io> (cherry picked from commit c049143)
eee6818
to
e14dc69
Compare
The merge conflicts have been resolved! |
Describe changes
I implemented changes that were left-over as a part of #2370.
Pre-requisites
Please ensure you have done the following:
develop
and the open PR is targetingdevelop
. If your branch wasn't based on develop read Contribution guide on rebasing branch to develop.Types of changes