Add MlClientDocumentationIT tests for classification. #47569
Conversation
przemekwitek
force-pushed
the
classification_docs
branch
from
41c48de to
d39605a
Compare
przemekwitek
force-pushed
the
classification_docs
branch
5 times, most recently
from
65152a7 to
c8df7f8
Compare
przemekwitek
force-pushed
the
classification_docs
branch
from
c8df7f8 to
230020a
Compare
|
Pinging @elastic/ml-core (:ml) |
|
Pinging @elastic/es-docs (>docs) |
There was a problem hiding this comment.
Looks good. I left a few nits and questions. Maybe nothing needs to be changed if it's obvious which of the duplicate class names to use, not possible to link to bookmarks on external sites, and we don't care about full stops at the end of numbered items. But they are things to at least consider.
...high-level/src/test/java/org/elasticsearch/client/documentation/MlClientDocumentationIT.java
Show resolved
Hide resolved
| { | ||
| // tag::evaluate-data-frame-evaluation-regression | ||
| Evaluation evaluation = | ||
| new Regression( // <1> |
There was a problem hiding this comment.
This Regression class is not fully qualified. But I don't think the doc examples include the imports. So this doesn't make it clear which package to choose when typing Regression into an IDE and it suggests two possible classes that could be imported.
It might be best to rename one of the classes, or else fully qualify the name here as well as where the other one is used in the docs.
(Same for Classification on line 3341.)
There was a problem hiding this comment.
Done.
I've fully qualified the names Regression and Classification in this file for now.
LMK if you like the idea of renaming Regression to RegressionEvaluation and Classification to ClassificationEvaluation (or maybe have a different idea for naming). Then I could move on with renaming.
| <2> Name of the field in the index. Its value denotes the actual (i.e. ground truth) label for an example. Must be either true or false. | ||
| <3> Name of the field in the index. Its value denotes the probability (as per some ML algorithm) of the example being classified as positive. | ||
| <4> The remaining parameters are the metrics to be calculated based on the two fields described above. | ||
| <5> https://en.wikipedia.org/wiki/Precision_and_recall[Precision] calculated at thresholds: 0.4, 0.5 and 0.6 |
There was a problem hiding this comment.
Is it possible to link to the #Precision bookmark on this page?
There was a problem hiding this comment.
You mean, instead of wikipedia link, or in addition?
Such a section does not exist yet on our page. Should I add it?
There was a problem hiding this comment.
You can link to a specific bookmark on the Wikipedia page like this:
https://en.wikipedia.org/wiki/Precision_and_recall#Precision
I'm not sure it's possible in Asciidoc though. Maybe the # causes a problem. If not don't worry.
There was a problem hiding this comment.
Ah, that's what you meant.
Sure, done.
| <3> Name of the field in the index. Its value denotes the probability (as per some ML algorithm) of the example being classified as positive. | ||
| <4> The remaining parameters are the metrics to be calculated based on the two fields described above. | ||
| <5> https://en.wikipedia.org/wiki/Precision_and_recall[Precision] calculated at thresholds: 0.4, 0.5 and 0.6 | ||
| <6> https://en.wikipedia.org/wiki/Precision_and_recall[Recall] calculated at thresholds: 0.5 and 0.7 |
There was a problem hiding this comment.
Is it possible to link to the #Recall bookmark on this page?
There was a problem hiding this comment.
See my questions above.
przemekwitek
force-pushed
the
classification_docs
branch
from
e2e5d40 to
9a6a33c
Compare
przemekwitek
force-pushed
the
classification_docs
branch
from
9a6a33c to
36e6446
Compare
|
run elasticsearch-ci/packaging-sample-matrix |
|
run elasticsearch-ci/packaging-sample |
This PR enhances client documentation tests with the new
classificationanalysis type:testPutDataFrameAnalyticstestEvaluateDataFrame,testEvaluateDataFrame_Classification,testEvaluateDataFrame_RegressionAdditionally, it adds basic java rest high-level docs related to
classification.Relates #46735