(v18: feat) Volume-Split, setup gauges to split evenly #6085

stackman27 · 2023-08-16T16:48:57Z

Part of : #6057

What is the purpose of the change

this feature contains creation of GroupGauges using the gauge architecture as well as splitting incentives rewards evenly across a specified InternalGaugeIds. Once we have the logic for volume split, we will split based on that.

Testing and Verifying

added unit test for functions
added functional test for the whole flow

Documentation and Release Note

Does this pull request introduce a new feature or user-facing behavior changes?
Changelog entry added to Unreleased section of CHANGELOG.md?

Where is the change documented?

Specification (x/{module}/README.md)
Osmosis documentation site
Code comments?
N/A

stackman27 · 2023-08-16T19:24:52Z

x/incentives/keeper/distribute_test.go

+//   - we only distribute external incentive in epoch 1.
+//   - Check that incentive record has been correctly created and gauge has been correctly updated.
+//   - all perpetual gauges must finish distributing records
+//   - ClPool1 will recieve full 1Musdc, 1Meth in this epoch.
+//   - ClPool2 will recieve 500kusdc, 500keth in this epoch.
+//   - ClPool3 will recieve full 1Musdc, 1Meth in this epoch whereas
+//
 // 6. Remove distribution records for internal incentives using HandleReplacePoolIncentivesProposal
 // 7. let epoch 2 pass
-//		-  We distribute internal incentive in epoch 2.
-// 		- check only external non-perpetual gauges with 2 epochs distributed
-// 		- check gauge has been correctly updated
-// 		- ClPool1 will already have 1Musdc, 1Meth (from epoch1) as external incentive. Will recieve 750Kstake as internal incentive.
-// 		- ClPool2 will already have 500kusdc, 500keth (from epoch1) as external incentive. Will recieve 500kusdc, 500keth (from epoch 2) as external incentive and 750Kstake as internal incentive.
-// 	    - ClPool3 will already have 1M, 1M (from epoch1) as external incentive. This pool will not recieve any internal incentive.
+//   - We distribute internal incentive in epoch 2.
+//   - check only external non-perpetual gauges with 2 epochs distributed
+//   - check gauge has been correctly updated
+//   - ClPool1 will already have 1Musdc, 1Meth (from epoch1) as external incentive. Will recieve 750Kstake as internal incentive.
+//   - ClPool2 will already have 500kusdc, 500keth (from epoch1) as external incentive. Will recieve 500kusdc, 500keth (from epoch 2) as external incentive and 750Kstake as internal incentive.
+//   - ClPool3 will already have 1M, 1M (from epoch1) as external incentive. This pool will not recieve any internal incentive.
+//
 // 8. let epoch 3 pass
-// 		- nothing distributes as non-perpetual gauges with 2 epochs have ended and perpetual gauges have not been reloaded
-//		- nothing should change in terms of incentive records
+//   - nothing distributes as non-perpetual gauges with 2 epochs have ended and perpetual gauges have not been reloaded
+//   - nothing should change in terms of incentive records


linter change, can ignore

stackman27 · 2023-08-16T19:28:35Z

@devbot-wizard help

devbot-wizard · 2023-08-16T19:28:37Z

Hi! I'm DevBot, a bot that helps with common tasks in the development process. Commands:

devbot fix: Fix basic errors in the PR. (e.g. imports, changelog conflicts, go mod conflicts)
devbot merge base: Merge the base branch into the PR branch.
devbot add changelog [misc/feat/bug/api/sec] [message]: Add a changelog entry to the PR. (e.g. devbot add changelog feat Added a new feature)
- If message is blank, defaults to PR title.
devbot re-pr: Re-PR the PR, useful for external contributors where we have no edit perms.

stackman27 · 2023-08-16T19:29:14Z

devbot add changelog (v18: feat) Volume-Split, setup gauges to split evenly

AlpinYukseloglu

The core logic looks generally good, but I had a couple of concerns around checks & testing that will require a little bit more work before this is merge-ready

x/incentives/keeper/distribute.go

AlpinYukseloglu · 2023-08-17T02:17:50Z

proto/osmosis/incentives/gauge.proto

+// internalGauges.
+message GroupGauge {
+  uint64 group_gauge_id = 1;
+  repeated uint64 internal_ids = 2;


We should include a splitting policy field here as an enum, even if it currently just evenly splits

added: https://github.com/osmosis-labs/osmosis/pull/6085/files#diff-ba512501fa081ff95a2b23175b9bba1ef438ff455e357e71ce884da0add490d5R54

x/incentives/keeper/distribute.go

AlpinYukseloglu · 2023-08-17T02:22:21Z

x/incentives/keeper/distribute.go

@@ -441,7 +480,10 @@ func (k Keeper) Distribute(ctx sdk.Context, gauges []types.Gauge) (sdk.Coins, er
 			ctx.Logger().Debug("distributeSyntheticInternal, gauge id %d, %d", "module", types.ModuleName, "gaugeId", gauge.Id, "height", ctx.BlockHeight())
 			gaugeDistributedCoins, err = k.distributeSyntheticInternal(ctx, gauge, filteredLocks, &distrInfo)
 		} else {
-			gaugeDistributedCoins, err = k.distributeInternal(ctx, gauge, filteredLocks, &distrInfo)
+			// Donot distribue if LockQueryType = Group, because if we distribute here we will be double distributing.
+			if gauge.DistributeTo.LockQueryType != lockuptypes.ByGroup {


I believe skipping in this way might lead to totalDistributedCoins just double counting the previous loop's gaugeDistributedCoins (if this is the case, let's make sure that tests catch this in the future).

The more appropriate logic seems to be if [same condition], continue and then leave the distributeInternal unchanged

yup makes sense

added: https://github.com/osmosis-labs/osmosis/pull/6085/files#diff-620e494b71d3ef447092659ad8b99835a3088760bb0924956007e88b3745f5bcR505

x/incentives/keeper/distribute_test.go

AlpinYukseloglu · 2023-08-17T02:24:35Z

x/incentives/keeper/distribute_test.go

+	var (
+		epochInfo = s.App.IncentivesKeeper.GetEpochInfo(s.Ctx)
+
+		expectedCoinsDistributed       = sdk.NewCoins(sdk.NewCoin("uosmo", sdk.NewInt(11_111_111)))


In general, having new logic rely entirely on a single hard coded test vector is probably not a sufficient testing practice. Happy to discuss what might be a more robust approach offline if you'd like

Refactored to make it more robust and readable: https://github.com/osmosis-labs/osmosis/pull/6085/files#diff-d5967ddef40ae47ef4140b9c47026cac85b85d9d04173b271c6c23dd182abdbcR1200

x/incentives/keeper/gauge_test.go

AlpinYukseloglu · 2023-08-17T02:29:45Z

x/incentives/keeper/store_test.go

+			// create 3 internal Gauge
+			var internalGauges []uint64
+			for i := 0; i <= 2; i++ {
+				internalGauge := s.CreateNoLockExternalGauges(clPool.GetId(), sdk.NewCoins(), s.TestAccs[1], uint64(1)) // gauge id = 2,3,4


For the sake of more readable & robust tests, it would probably be better to abstract this setup into a test helper and add numberOfExistingGauges as a test parameter

created helper for creating internalGauges for GroupGauges
c99cb1d#diff-1000dbecb1333fe09226de57d089fe167e289bc9edae3aac2f040029bfa91b1eR51

AlpinYukseloglu · 2023-08-18T19:20:34Z

proto/osmosis/incentives/gauge.proto

+  Liquidity = 1;
+  Evenly = 2;


Suggested change

Liquidity = 1;

Evenly = 2;

Evenly = 1;

We're likely to simply remove liquidity based splitting for the reasons described in #6093

x/incentives/types/keys.go

AlpinYukseloglu · 2023-08-18T19:29:19Z

x/incentives/keeper/distribute_test.go

+		internalGaugeIds: []uint64{2, 3, 4, 5},
+	}
+
+	tests := []struct {


imo the map[string]struct structure is much cleaner (e.g. see how it's done here – no need for name as a field and you can pick up the test case name from the first for loop field on line 1289 below)

yea, i personally like writing name field, it just gives me more clarity while writing tests 😅

happy to change tho if we want to stay consistent with rest of the test

x/incentives/keeper/distribute.go

AlpinYukseloglu · 2023-08-18T19:33:12Z

x/incentives/keeper/distribute.go

+}
+
+// CalcSplitPolicyCoins calculates tokens to split given a policy and groupGauge.
+func (k Keeper) CalcSplitPolicyCoins(ctx sdk.Context, policy types.SplittingPolicy, groupGauge *types.Gauge, groupGaugeObj types.GroupGauge) (sdk.Coins, sdk.Coins, error) {


nit: this should probably be a private function

AlpinYukseloglu · 2023-08-18T19:37:02Z

x/incentives/keeper/distribute.go

+	}
+
+	for _, groupGauge := range groupGauges {
+		gauge, err := k.GetGaugeByID(ctx, groupGauge.GroupGaugeId)


Err what are we actually getting here? Don't we already have the group gauge?

GetAllGroupGauges returns the ids for ex: {GroupGaugeId, InternalIds and Splitting policy}, we still have to fetch the actual gauge to get its fields

AlpinYukseloglu · 2023-08-18T19:40:27Z

x/incentives/keeper/distribute.go

+		}
+
+		// only allow distribution if the GroupGauge is Active
+		if !currTime.Before(gauge.StartTime) && (gauge.IsPerpetual || gauge.FilledEpochs < gauge.NumEpochsPaidOver) {


nit: for general readability, I think doing something like "if currTime.After(startTime) && (!gaugeIsActive bool), then continue" would be cleaner instead of having half the function in an if branch.

Ofc feel free to ignore this if you think the current approach is more readable (hence the nit) – my thinking is just that you need to keep track of fewer things mentally when you're not in a branch

yup i think we can just replace this with
gauge.IsActiveGauge(currTime) as well. Looks cleaner

x/incentives/keeper/gauge.go

x/incentives/keeper/keeper_test.go

AlpinYukseloglu

LGTM, nice work

AlpinYukseloglu · 2023-08-23T03:03:00Z

x/incentives/keeper/gauge.go

+	k.SetGroupGauge(ctx, newGroupGauge)
+	k.SetLastGaugeID(ctx, gauge.Id)
+
+	// TODO: check if this is necessary, will investigate this in following PR.


Did you flag this in an issue?

* (v18: feat) Volume-Split, setup gauges to split evenly * update changelog * halfway done refactoring test * refactored test * alp comments and refactored tests * alp comments * nits --------- Co-authored-by: devbot-wizard <[email protected]>

github-actions bot added C:x/lockup C:x/incentives labels Aug 16, 2023

stackman27 force-pushed the stack/vol-split-inc branch from 57609c0 to 95a914b Compare August 16, 2023 19:19

stackman27 commented Aug 16, 2023

View reviewed changes

stackman27 mentioned this pull request Aug 16, 2023

Volume Splitting Incentives Tracking Issue #6057

Closed

(v18: feat) Volume-Split, setup gauges to split evenly

6812b45

stackman27 force-pushed the stack/vol-split-inc branch from 95a914b to 6812b45 Compare August 16, 2023 19:27

update changelog

57a4b1a

stackman27 added the V:state/breaking State machine breaking PR label Aug 16, 2023

stackman27 marked this pull request as ready for review August 16, 2023 19:29

stackman27 requested review from mattverse, czarcas7ic, AlpinYukseloglu and p0mvn as code owners August 16, 2023 19:29

AlpinYukseloglu reviewed Aug 17, 2023

View reviewed changes

stackman27 added 3 commits August 17, 2023 22:57

halfway done refactoring test

7b330ba

refactored test

58dde1a

alp comments and refactored tests

c99cb1d

stackman27 requested a review from AlpinYukseloglu August 18, 2023 07:22

AlpinYukseloglu reviewed Aug 18, 2023

View reviewed changes

x/incentives/keeper/keeper_test.go Show resolved Hide resolved

stackman27 added 2 commits August 18, 2023 20:26

alp comments

2fe6334

nits

cad538f

stackman27 added the F: volume-splitting feat: Volume-splitting incentives label Aug 19, 2023

stackman27 requested a review from AlpinYukseloglu August 22, 2023 16:31

AlpinYukseloglu approved these changes Aug 23, 2023

View reviewed changes

AlpinYukseloglu merged commit eca86f7 into main Aug 23, 2023

AlpinYukseloglu deleted the stack/vol-split-inc branch August 23, 2023 04:27

github-actions bot mentioned this pull request Mar 15, 2024

Broken Links Detected #7752

Closed

github-actions bot mentioned this pull request Apr 1, 2024

Broken Links Detected #7903

Closed

github-actions bot mentioned this pull request Apr 15, 2024

Broken Links Detected #8046

Closed

github-actions bot mentioned this pull request Nov 1, 2024

Broken Links Detected notional-labs/osmosis#76

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

(v18: feat) Volume-Split, setup gauges to split evenly #6085

(v18: feat) Volume-Split, setup gauges to split evenly #6085

stackman27 commented Aug 16, 2023 •

edited

Loading

stackman27 Aug 16, 2023 •

edited

Loading

stackman27 commented Aug 16, 2023

devbot-wizard commented Aug 16, 2023

stackman27 commented Aug 16, 2023

AlpinYukseloglu left a comment

AlpinYukseloglu Aug 17, 2023

stackman27 Aug 17, 2023

AlpinYukseloglu Aug 17, 2023

stackman27 Aug 17, 2023 •

edited

Loading

stackman27 Aug 17, 2023

AlpinYukseloglu Aug 17, 2023

stackman27 Aug 17, 2023

AlpinYukseloglu Aug 17, 2023

stackman27 Aug 18, 2023

AlpinYukseloglu Aug 18, 2023

AlpinYukseloglu Aug 18, 2023

AlpinYukseloglu Aug 18, 2023

stackman27 Aug 18, 2023

AlpinYukseloglu Aug 18, 2023

AlpinYukseloglu Aug 18, 2023

stackman27 Aug 18, 2023

AlpinYukseloglu Aug 18, 2023

stackman27 Aug 18, 2023

AlpinYukseloglu left a comment

AlpinYukseloglu Aug 23, 2023

(v18: feat) Volume-Split, setup gauges to split evenly #6085

(v18: feat) Volume-Split, setup gauges to split evenly #6085

Conversation

stackman27 commented Aug 16, 2023 • edited Loading

What is the purpose of the change

Testing and Verifying

Documentation and Release Note

stackman27 Aug 16, 2023 • edited Loading

Choose a reason for hiding this comment

stackman27 commented Aug 16, 2023

devbot-wizard commented Aug 16, 2023

stackman27 commented Aug 16, 2023

AlpinYukseloglu left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

stackman27 Aug 17, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

AlpinYukseloglu left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

stackman27 commented Aug 16, 2023 •

edited

Loading

stackman27 Aug 16, 2023 •

edited

Loading

stackman27 Aug 17, 2023 •

edited

Loading