Skip to content
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

AIP-653: Sym-link python SDK components to avoid poetry lacking subdirectory install capabilities #1

Merged
merged 1 commit into from
Nov 4, 2020

Conversation

alexlatchford
Copy link
Collaborator

@alexlatchford alexlatchford commented Oct 29, 2020

We want to be able to install the ZG fork of kfp with poetry.

Unfortunately poetry has a limitation where it cannot install subdirectory packages for git paths, see python-poetry/poetry#755. So to get around this issue we symlink the appropriate Python SDK files to the root directory instead 😅


As an interesting aside another approach considered for this was to publish a package to Github Packages and have the aip-kfp-sdk library rely on it but as per github/roadmap#94 python packages are not support 😒 For the future open-source workstream it is likely we'll want to publish packages to the global PyPI.

@alexlatchford alexlatchford marked this pull request as ready for review October 29, 2020 22:35
setup.py Show resolved Hide resolved
@alexlatchford alexlatchford merged commit 7e93631 into feature/zg Nov 4, 2020
@alexlatchford alexlatchford deleted the alexla/AIP-653 branch November 4, 2020 17:09
alexlatchford pushed a commit that referenced this pull request Aug 30, 2021
* Changes:
* add kfp v2 hello world sample
* generate go proto code
* code to unmarshal via jsonpb
* generate container spec
* add root DAG task and Makefile to run in argo
* env setup to quickly build driver container
* generate mlmd grpc golang client
* mlmd grpc client usage example
* driver-dag: put execution and context
* driver dag: put context only when execution is a DAG
* driver task: resolve parameter from parent execution
* chain dag driver with task driver
* driver: output argo parameters
* driver: build driver docker image
* driver: push image to dev registry
* compiler: root dag driver compiler
* mlmd doc
* driver: use task spec and executor spec as driver input instead
* driver: pod spec patch output for type EXECUTOR
* compiler: passing - parameter placeholder with hello world example
* include generated argo workflow in source control
* driver no longer outputs parameters
* publisher: publish parameters to execution custom properties
* driver, publisher: execution state
* publisher: remove built binary from repo
* sample: add producer consumer sample
* sample: producer consumer sample with only parameters
* e2e output parameter support
* e2e: driver resolves input parameter from tasks in the same DAG
* compiler: convert KFP task dependency to argo task dependency
* feat: refactor publisher so it can run as an entrypoint
* build: entrypoint image
* feat: executor and publisher in one container via entrypoint rewriting

* fixed compile error (#1)

* add licenses

* update readme

Co-authored-by: capri-xiyue <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants