Evaluation with FlagTagsOperator to match any or all tags #408
Conversation
| @@ -1311,6 +1311,14 @@ definitions: | |||
| x-omitempty: true | |||
| items: | |||
| type: string | |||
| flagTagsOperator: | |||
There was a problem hiding this comment.
btw, this doc is generated, you can move the change to https://github.com/checkr/flagr/blob/master/swagger/index.yaml#L518, and then run make gen to get the same result
There was a problem hiding this comment.
Sure, I wasn't sure how this work
pkg/handler/eval.go
Outdated
|
|
||
| fs := cache.GetByTags(evalContext.FlagTags) | ||
| fs := cache.GetByTags(evalContext.FlagTags, operator) |
There was a problem hiding this comment.
for readability, I think you can just pass the evalContext.FlagTagsOperator as *string to cache.GetByTags, and then check if if operator == nil || *operator == models.EvaluationBatchRequestFlagTagsOperatorANY there.
There was a problem hiding this comment.
Thanks reads better
pkg/handler/eval_cache.go
Outdated
| if !ok { | ||
| // clear as no flags | ||
| results = map[uint]*entity.Flag{} | ||
| continue |
There was a problem hiding this comment.
if the tag is not there, I think should we just return nil here given the ALL operator.
There was a problem hiding this comment.
You're totally right
There was a problem hiding this comment.
But is it correct to return nil or should we return an empty list?
pkg/handler/eval_cache.go
Outdated
| missingFIDs := []uint{} | ||
| for fID := range results { | ||
| if fSet[fID] == nil { | ||
| missingFIDs = append(missingFIDs, fID) | ||
| } | ||
| } | ||
| for _, fID := range missingFIDs { | ||
| delete(results, fID) | ||
| } |
There was a problem hiding this comment.
How about one pass of the keys in results, and just delete the keys not the current fSet?
for fID := range results {
if _, ok := fSet[fID]; !ok {
delete(results, fID)
}
}
There was a problem hiding this comment.
Cool, I wasn't sure if in Go is possible to modify a slice while iterating it. Changing it for sure
There was a problem hiding this comment.
related topic https://stackoverflow.com/questions/23229975/is-it-safe-to-remove-selected-keys-from-map-within-a-range-loop, I also learned that delete marks the keys as deleted, thus safe to keep iterating.
pacoguzman
force-pushed
the
batch-eval-tags-operator
branch
from
af774ce to
756a071
Compare
|
I've applied all your feedback as that makes a lot of sense and simplify the solution. Thanks for the review |
Codecov Report
@@ Coverage Diff @@
## master #408 +/- ##
==========================================
- Coverage 82.99% 81.90% -1.09%
==========================================
Files 27 27
Lines 1752 1603 -149
==========================================
- Hits 1454 1313 -141
+ Misses 214 211 -3
+ Partials 84 79 -5
Continue to review full report at Codecov.
|
pkg/handler/eval_cache.go
Outdated
| results[fID] = f | ||
| } | ||
| } else { | ||
| if len(results) > 0 { |
There was a problem hiding this comment.
one more thing to optimize is that you can return early when len(results) == 0
There was a problem hiding this comment.
Done, after each delete loop we check if there is still any match to continue or return early.
pacoguzman
force-pushed
the
batch-eval-tags-operator
branch
from
756a071 to
8377d3a
Compare
|
👍 I'll try to remember thanks for the pointer
…On Tue, Oct 13, 2020, 5:51 PM Zhuojie Zhou ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In pkg/handler/eval_cache.go
<#408 (comment)>:
> + missingFIDs := []uint{}
+ for fID := range results {
+ if fSet[fID] == nil {
+ missingFIDs = append(missingFIDs, fID)
+ }
+ }
+ for _, fID := range missingFIDs {
+ delete(results, fID)
+ }
related topic
https://stackoverflow.com/questions/23229975/is-it-safe-to-remove-selected-keys-from-map-within-a-range-loop,
I also learned that delete marks the keys as deleted, thus safe to keep
iterating.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#408 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAACVLRWCEQSDP3XQJUMK53SKRZRFANCNFSM4SJYHM5Q>
.
|
Finding flags to be evaluated using tags allow to pass ANY/ALL tag matching as operator
Description
Just support a new enum value
ANYorALLto determine which flags matches the tags the default isANY.ALLjust filter out results applying tags one after the other.Motivation and Context
The main idea of the change is to allow to pass
flagsTagsOperatoron batch evaluation requests so we can handle scenarios where each tag used to describe the flags are independent so we could want to match flags which contains all the tags passed. Currently, tags are considered as any tag in the flags make the match.We have tags determining the ownership of the flags, and some teams group flags inside other tags and we have some generic rules to naming it.
owner:team-namebatch:batch-nameSo as could be collisions on the names between teams we think that adding a match all operator for tags solve this easily to just evaluate on those flags matching all the tags provided. This is opt in as the default behaviour is the same as before.
How Has This Been Tested?
I've added a couple of tests in the codebase to describe the new behaviour
Types of changes
Checklist: