Add parsing for single bucket aggregations #24564
Conversation
There was a problem hiding this comment.
This is weird and prevented the tests from working. We implement this method in InternalAggregation#getType() to return getWriteableName(), and InternalSampler and its subclass UnmappedSampler have different names there. If we used "sampler" for both classes in the XContent output as type prefix we wouldn't be able to distinguish the two when parsing. I wonder if this was added by mistake or for a reason, maybe you can double check this?
There was a problem hiding this comment.
I think the idea was that InternalSampler and UnmappedSampler should have the same type, so that we only parse into InternalSampler. Maybe we meant to call it "sampler" in both cases rather than mapped_sampler but I guess we don't need that.
There was a problem hiding this comment.
Why didn't we have this problem also in the ObjectParser version of this PR?
There was a problem hiding this comment.
I think we did, but I just found it. The fromXContent() tests are curerently skipped if the register the parser under "sampler" but then find "unmapped_sampler" or "mapped_sampler" in the xContent
There was a problem hiding this comment.
now I see what you meant yesterday saying that we have to parse meta here
cbuescher
force-pushed
the
addParsing-SingleBucket-v2
branch
2 times, most recently
from
e990186 to
be8e0f7
Compare
cbuescher
changed the title
javanna
left a comment
javanna
left a comment
There was a problem hiding this comment.
LGTM good to go when you resolved the last TODO
cbuescher
force-pushed
the
addParsing-SingleBucket-v2
branch
2 times, most recently
from
3b49c6d to
f4ad9b8
Compare
cbuescher
force-pushed
the
addParsing-SingleBucket-v2
branch
from
f4ad9b8 to
035f238
Compare
This adds parsing to all implementations of SingleBucketAggregations. They are mostly similar, so they share the common base class `ParsedSingleBucketAggregation` and the shared base test `InternalSingleBucketAggregationTestCase`.
3 participants
This is an alternative approach to #24386 that doesn't rely on extending ObjectParsers capabilities and uses manual parsing instead.