Render code to load model when pyfunc flavor is unavailable #5006

harupy · 2021-11-04T12:45:05Z

Signed-off-by: harupy hkawamura0130@gmail.com

What changes are proposed in this pull request?

Currently, the model artifact viewer renders code to load the logged model as a pyfunc model and make predictions even when the logged model doesn't contain the pyfunc flavor. This PR fixes the issue by rendering code to load the logged model using the original flavor module.

Before

no-pyfunc-model.mov

The logged model doesn't contain the pyfunc flavor.
The model artifact viewer renders code that loads the model as a pyfunc model.

After

The model artifact viewer renders code that loads the model using mlflow.sklearn.

import numpy as np
from sklearn.decomposition import PCA

X = np.array([[-1, -1], [-2, -1], [-3, -2], [1, 1], [2, 1], [3, 2]])
pca = PCA(n_components=2)
pca.fit(X)

import mlflow

with mlflow.start_run():
    mlflow.sklearn.log_model(pca, "model")

How is this patch tested?

Unit tests

Release Notes

Is this a user-facing change?

No. You can skip the rest of this section.
Yes. Give a description of this change to be included in the release notes for MLflow users.

(Details in 1-2 sentences. You can just refer to another PR with a description if this PR is part of a larger change.)

What component(s), interfaces, languages, and integrations does this PR affect?

Components

Interface

area/uiux: Front-end, user experience, plotting, JavaScript, JavaScript dev server
area/docker: Docker use across MLflow's components, such as MLflow Projects and MLflow Models
area/sqlalchemy: Use of SQLAlchemy in the Tracking Service or Model Registry
area/windows: Windows support

Language

language/r: R APIs and clients
language/java: Java APIs and clients
language/new: Proposals for new client languages

Integrations

integrations/azure: Azure and Azure ML integrations
integrations/sagemaker: SageMaker integrations
integrations/databricks: Databricks integrations

How should the PR be classified in the release notes? Choose one:

rn/breaking-change - The PR will be mentioned in the "Breaking Changes" section
rn/none - No description will be included. The PR will be mentioned only by the PR number in the "Small Bugfixes and Documentation Updates" section
rn/feature - A new user-facing feature worth mentioning in the release notes
rn/bug-fix - A user-facing bug fix worth mentioning in the release notes
rn/documentation - A user-facing documentation change worth mentioning in the release notes

...s/src/experiment-tracking/components/artifact-view-components/ShowArtifactLoggedModelView.js

dbczumar · 2021-11-15T22:16:46Z

...s/src/experiment-tracking/components/artifact-view-components/ShowArtifactLoggedModelView.js

+  loadModelCodeText(modelPath, flavor) {
+    return (
+      `import mlflow\n` +
+      `logged_model = '${modelPath}'\n\n` +
+      `# Load model.\n` +
+      `loaded_model = mlflow.${flavor}.load_model(logged_model)\n`
+    );
+  }


@harupy This is perfect for everything except MLeap. MLeap models can't be reloaded in Python. Perhaps we can add a Java example for MLeap or show nothing for now?

Got it. If so, I'd prefer to show nothing for now.

Sounds good!

dbczumar · 2021-11-15T22:20:36Z

...s/src/experiment-tracking/components/artifact-view-components/ShowArtifactLoggedModelView.js

+            See <a href={CustomPyfuncModelsDocUrl}>Creating custom Pyfunc models</a> for how to log
+            this model as a PyFuncModel.


Suggested change

See <a href={CustomPyfuncModelsDocUrl}>Creating custom Pyfunc models</a> for how to log

this model as a PyFuncModel.

See <a href={CustomPyfuncModelsDocUrl}>creating custom Pyfunc models</a> to learn how

to customize this model and deploy it for batch or real-time scoring using the ``pyfunc`` model flavor.

It would be nice to link to the pyfunc docs too: https://www.mlflow.org/docs/latest/python_api/mlflow.pyfunc.html#module-mlflow.pyfunc.

Sure, I'll add the pyfunc link!

Signed-off-by: harupy <hkawamura0130@gmail.com>

…iew-components/ShowArtifactLoggedModelView.js Co-authored-by: dbczumar <39497902+dbczumar@users.noreply.github.com> Signed-off-by: harupy <hkawamura0130@gmail.com>

Signed-off-by: harupy <hkawamura0130@gmail.com>

harupy · 2021-11-17T01:40:50Z

@dbczumar Let me add a few screenshots to show how it looks like now.

harupy · 2021-11-17T03:28:33Z

mleap

sklearn without pyfunc flavor

pyfunc

dbczumar

LGTM! Thanks @harupy !

harupy · 2021-11-17T07:05:31Z

Thanks for the review!

github-actions bot added the rn/bug-fix Mention under Bug Fixes in Changelogs. label Nov 4, 2021

harupy changed the title ~~Avoid render conde snippets when mode doesn't have pyfunc flavor~~ Avoid rendering code snippets when model doesn't have pyfunc flavor Nov 4, 2021

harupy requested review from dbczumar and jinzhang21 November 4, 2021 13:15

dbczumar reviewed Nov 4, 2021

View reviewed changes

...s/src/experiment-tracking/components/artifact-view-components/ShowArtifactLoggedModelView.js Outdated Show resolved Hide resolved

harupy changed the title ~~Avoid rendering code snippets when model doesn't have pyfunc flavor~~ Render code to load model when pyfunc flavor is unavailable Nov 8, 2021

harupy changed the title ~~Render code to load model when pyfunc flavor is unavailable~~ [WIP] Render code to load model when pyfunc flavor is unavailable Nov 15, 2021

harupy marked this pull request as draft November 15, 2021 15:11

dbczumar reviewed Nov 15, 2021

View reviewed changes

harupy force-pushed the do-not-render-code branch from 165cd0a to 1c42092 Compare November 15, 2021 23:58

harupy changed the title ~~[WIP] Render code to load model when pyfunc flavor is unavailable~~ Render code to load model when pyfunc flavor is unavailable Nov 16, 2021

harupy and others added 8 commits November 17, 2021 03:32

Avoid render conde snippets when mode doesn't have pyfunc flavor

c276f30

Signed-off-by: harupy <hkawamura0130@gmail.com>

show code to load model

41a5321

Signed-off-by: harupy <hkawamura0130@gmail.com>

Update mlflow/server/js/src/experiment-tracking/components/artifact-v…

adf05d1

…iew-components/ShowArtifactLoggedModelView.js Co-authored-by: dbczumar <39497902+dbczumar@users.noreply.github.com> Signed-off-by: harupy <hkawamura0130@gmail.com>

lint

4116128

Signed-off-by: harupy <hkawamura0130@gmail.com>

address comments and update tests

2c98a9a

Signed-off-by: harupy <hkawamura0130@gmail.com>

refactor

f8609e3

Signed-off-by: harupy <hkawamura0130@gmail.com>

Fix FormattedMessage props

d4d237a

Signed-off-by: harupy <hkawamura0130@gmail.com>

18n

81e86f0

Signed-off-by: harupy <hkawamura0130@gmail.com>

harupy force-pushed the do-not-render-code branch from 653ba6a to 81e86f0 Compare November 16, 2021 18:33

harupy marked this pull request as ready for review November 17, 2021 00:14

dbczumar approved these changes Nov 17, 2021

View reviewed changes

harupy merged commit 422c1c9 into mlflow:master Nov 17, 2021

harupy deleted the do-not-render-code branch November 17, 2021 07:05

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Render code to load model when pyfunc flavor is unavailable #5006

Render code to load model when pyfunc flavor is unavailable #5006

harupy commented Nov 4, 2021 •

edited

Loading

dbczumar Nov 15, 2021 •

edited

Loading

harupy Nov 16, 2021

dbczumar Nov 16, 2021

dbczumar Nov 15, 2021 •

edited

Loading

harupy Nov 15, 2021

harupy commented Nov 17, 2021

harupy commented Nov 17, 2021 •

edited

Loading

dbczumar left a comment

harupy commented Nov 17, 2021

		See <a href={CustomPyfuncModelsDocUrl}>Creating custom Pyfunc models</a> for how to log
		this model as a PyFuncModel.

Render code to load model when pyfunc flavor is unavailable #5006

Render code to load model when pyfunc flavor is unavailable #5006

Conversation

harupy commented Nov 4, 2021 • edited Loading

What changes are proposed in this pull request?

Before

After

How is this patch tested?

Release Notes

Is this a user-facing change?

What component(s), interfaces, languages, and integrations does this PR affect?

How should the PR be classified in the release notes? Choose one:

dbczumar Nov 15, 2021 • edited Loading

Choose a reason for hiding this comment

harupy Nov 16, 2021

Choose a reason for hiding this comment

dbczumar Nov 16, 2021

Choose a reason for hiding this comment

dbczumar Nov 15, 2021 • edited Loading

Choose a reason for hiding this comment

harupy Nov 15, 2021

Choose a reason for hiding this comment

harupy commented Nov 17, 2021

harupy commented Nov 17, 2021 • edited Loading

mleap

sklearn without pyfunc flavor

pyfunc

dbczumar left a comment

Choose a reason for hiding this comment

harupy commented Nov 17, 2021

harupy commented Nov 4, 2021 •

edited

Loading

dbczumar Nov 15, 2021 •

edited

Loading

dbczumar Nov 15, 2021 •

edited

Loading

harupy commented Nov 17, 2021 •

edited

Loading