New website: review from [2/3] to [3/3]: New Pipelines#6
New website: review from [2/3] to [3/3]: New Pipelines#6ThomasDelteil wants to merge 101 commits intonew_website_delete_contentfrom
Conversation
Co-authored-by: Aaron Markham <markhama@amazon.com> Co-authored-by: Thomas Delteil <thomas.delteil1@gmail.com>
Co-authored-by: Thomas Delteil <thomas.delteil1@gmail.com> Co-authored-by: Aaron Markham <markhama@amazon.com>
pr validation functions for website
…rdocs R pdf creation
…rdocs activate r publishing; switch to archive for final site tar
update readmes for doc; remove unused files
|
I'm familiar with the md changes and have reviewed the Jenkins pipeline code. LGTM. |
|
|
||
|
|
||
| build_docs_setup() { | ||
| build_folder="docs/_build" |
|
|
||
| popd | ||
|
|
||
| GZIP=-9 tar zcvf $artifacts_path $docs_build_path |
There was a problem hiding this comment.
this zips up with high compression (GZIP=-9) the generated artifact in a specific path so that it can then be picked up by CI for building the entire website as the sum of its components
ci/docker/runtime_functions.sh
Outdated
|
|
||
| build_version_doc() { | ||
|
|
||
| # $1 is the list of branch or tag to build |
There was a problem hiding this comment.
This function can be deleted. It's only used for the current website.
| @@ -0,0 +1,218 @@ | |||
| # Creating custom operators with numpy | |||
There was a problem hiding this comment.
@ptrendx pointed out that AMP docs are missing
There was a problem hiding this comment.
that's correct, some tutorials are missing and will be part of a post-launch re-alignment
aaronmarkham
left a comment
There was a problem hiding this comment.
Some suggested changes and comments.
I'm ignoring the /docs/python_docs and other folders since I don't think they're intended to be in this PR.
| node(NODE_LINUX_CPU) { | ||
| ws('workspace/docs') { | ||
| timeout(time: max_time, unit: 'MINUTES') { | ||
| archiveArtifacts 'docs/_build/full_website.tgz' |
There was a problem hiding this comment.
This should be replaced with upload to S3 once we have a working CI-dev env for it.
| // utils.pack_lib('full_website', 'docs/_build/full_website.tgz', false) | ||
|
|
||
| // archive so the publish pipeline can access the artifact | ||
| archiveArtifacts 'docs/_build/full_website.tgz' |
There was a problem hiding this comment.
Should be switched to S3 when a test environment is made available on CI-dev.
ci/jenkins/Jenkins_steps.groovy
Outdated
| utils.docker_run('ubuntu_cpu_jekyll', 'build_docs_small', false) | ||
|
|
||
| // Publish preview to S3 | ||
| sh "cd docs/_build/html && ci/other/ci_deploy_doc.sh ${env.BRANCH_NAME} ${env.BUILD_NUMBER}" |
There was a problem hiding this comment.
Not useful until there's a website serving process attached to S3.
There was a problem hiding this comment.
will comment out until you / someone setups an alternative setup
Co-Authored-By: Aaron Markham <markhama@amazon.com>
…ax_error fix syntax error
…teil/incubator-mxnet into new_website_pipeline_2
* Set monitor callback basic support * Trigger CI * Add base.pyi and ndarray.pyx * Change not supported to experimental and check for both static_shape and static_alloc
* add aaron to codeowners * retrigger CI * ci
* add shard API * more balanced * fix a bug * Update test_gluon_data.py
Add test Fix Fix sanity Fix build failure Use array_function protocol only when np.version >= 1.17 Add unit tests for array ufunc protocol Fix pylint Reformat Fix build Fix Refactor test suite for numpy interoperability
* Fix README Build Status * Add labels to Jenkins Badges * Add more line breaks
* improve dataloader signals and messages * address comments * fix spawn tests on windows
* normal implemented * numpy behavior normal imlemented * retrigger CI * retrigger CI * regrigger ci * add normal parameter check * add raise for normal * remove dead code
* Fix mkldnn subgraph with float64 * Fix ci Change-Id: I0bb4e8d7a0a534aa661601887cc633cb9c4fcadf * Fix test Change-Id: I96c529abe7adb6def90a22f03b3432263ef12fda * Update dmlc-core Change-Id: I472fb7bbffc16ed8c36494ab49838b08c59b2f12 * pin to offical dmlc Change-Id: I5a27dc83b892bf8fcb34bb089449d1d3b6e9beed * Fix GPU CI Change-Id: I285947e01bdb0651c2c7830ed4eb76931a09b754 * Fix GPU CI Change-Id: I6f23b51d6bda44f6ae18766ebe390118740bb9c7
* Adding new website files Co-authored-by: Thomas Delteil <thomas.delteil1@gmail.com> Co-authored-by: Aaron Markham <markhama@amazon.com> * removing old content Co-authored-by: Thomas Delteil <thomas.delteil1@gmail.com> Co-authored-by: Aaron Markham <markhama@amazon.com> * add license header * fixing links and url + css * fixing tutorials * exclude python_docs folder when building prod site This allows things to coexist until a following PR deletes this file. * license header fixes * exclude static site from license header check * exclude static site and sphinx theme * exclude static_site and cpp_docs from sphinx * add licenses and excludes * fix headers * restore file for legacy builds * empty commit to trigger CI * update submodule to correct one * delete website check * add placeholder website check * add placeholder website check, for reals this time * trigger CI * trigger PR
Review the new pipeline in isolation here