Implement the TACo API module by piotr-roslaniec · Pull Request #263 · nucypher/taco-web

piotr-roslaniec · 2023-08-11T10:00:23Z

Type of PR:

Feature

Required reviews:

3

What this does:

Implements a simplified API version for TACo

Issues fixed/closed:

Fixes Alter PRE-adapted tDec API design for CBD-DKG #166

Why it's needed:

Explain how this PR fits in the greater context of the NuCypher Network.
E.g., if this PR address a nucypher/productdev issue, let reviewers know!

Notes for reviewers:

Rebased over Apply PR suggestions for #273 #286
Please see test/unit/taco.test.ts for usage examples

codecov-commenter · 2023-08-14T09:20:52Z

Codecov Report

Merging #263 (1dba09f) into alpha (c18f996) will decrease coverage by 0.01%.
The diff coverage is 72.09%.

@@            Coverage Diff             @@
##            alpha     #263      +/-   ##
==========================================
- Coverage   81.37%   81.36%   -0.01%     
==========================================
  Files          36       37       +1     
  Lines        1036     1068      +32     
  Branches      110      112       +2     
==========================================
+ Hits          843      869      +26     
- Misses        185      191       +6     
  Partials        8        8

Files Changed	Coverage Δ
src/conditions/context/context.ts	`96.82% <ø> (ø)`
src/agents/coordinator.ts	`24.48% <10.00%> (-2.79%)`	⬇️
src/sdk/strategy/cbd-strategy.ts	`93.33% <66.66%> (ø)`
src/dkg.ts	`33.33% <80.00%> (+5.12%)`	⬆️
src/taco.ts	`95.83% <95.83%> (ø)`
src/characters/cbd-recipient.ts	`92.72% <100.00%> (ø)`

src/taco.ts

theref

This looks great, exactly what adopters want to use

src/characters/cbd-recipient.ts

src/conditions/condition-expr.ts

src/taco.ts

piotr-roslaniec · 2023-09-04T11:11:08Z

src/taco.ts

+  // TODO: How do we get rid of these two fields? We need them for decrypting
+  // We ritualId in order to fetch the DKG participants and create DecryptionRequests for them
+  ritualId: number;
+  // We need to know the threshold in order to create DecryptionRequests
+  threshold: number;


Is there anything we can do to get rid of these parameters in the user-facing API?

Do we want to expose class TacoMessageKit.requireSigner so that decrypting users can detect when they need to pass signer in?

Given my comment above, I question whether this TacoMesageKit object is needed at all.

conditionExpr is in the ACP in the ThresholdMessageKit, and I don't believe ritualId and threshold are needed because they can be obtained from the Coordinator contract along with the ThresholdMessageKit.

My concern lies in returning objects that aren't subject to versioning, and the caller potentially serializing them. Ideally, encrypt would return a ThresholdMessageKit just like python.

Of course, if additional information is actually needed in the ThresholdMessageKit itself we can explore that.

derekpierre

See comment about TacoMessageKit.

src/agents/coordinator.ts

derekpierre

Needs another rebase over #286.

derekpierre · 2023-09-08T15:05:16Z

src/agents/coordinator.ts

+      word0: dkgPublicKeyBytes.slice(0, 16),
+      word1: dkgPublicKeyBytes.slice(16, 32),


A G1Point is 48 bytes and defined as follows.

Shouldn't word0 by (0,32) and word1 (32,48)?

I think we need to start thinking about "acceptance" tests in nucypher-ts that interact with actual contracts (cc @cygnusv , @KPrasch ).

IMHO acceptance tests with contracts are unnecessary. We're only supporting read operations (at this point), and setting those up is not different than existing mocks (so, just having simple pass-through facades for mock values). Yes, this is a bug that went unnoticed. But it would have been caught easily by QA which we do on every release.

I think the better solution to this particular problem is to enhance the BLS12381.G1PointStructOutput generated type with a wrapper that performs validation. The word0 and word fields are represented as strings by typechain in generated typings, which also prevented the TS compiler from catching this issue.

But it would have been caught easily by QA which we do on every release.

The goal of the "acceptance" tests is to catch these things before live QA. It is inefficient to catch these kind of bugs during live QA on a testnet. Doesn't mean we will catch everything in those tests, but it does reduce the need for subsequent fixes/releases after bugs are found during live testing.

My typescript philosophy knowledge is admittedly limited and I don't know what specifically the "acceptance" tests would entail and what it would require from the codebase to be set up, but having everything tested with mocks limits the breadth of testing and bugs that we can find in tests - who's testing the mocks is essentially the argument.

"Acceptance" tests is a broad term, and maybe there is more appropriate term here, but we should investigate how we can perform tests closer to what the actual testnet/production environment would be.

I understand the advantages of acceptance testing, I just don't think they are necessary at this point. We should probably get them before the release, but not until we're done with the monorepo or adapting viem: #271

That being said, I'd be happy to accept a PR that implements acceptance testing if anyone wants to chip in.

Definitely not something needed for this PR or this instant - but something we should nonetheless investigate. Filed #295 .

derekpierre · 2023-09-08T15:12:23Z

src/agents/coordinator.ts

+  accessController: string;
+  publicKey: BLS12381.G1PointStructOutput;
  aggregatedTranscript: string;
 }


Are participants not stored in this struct? I guess maybe not needed...?

The contents of this struct are taken from types generated from contracts by typechain. I can trim them down to a size.

derekpierre · 2023-09-08T15:13:42Z

src/sdk/strategy/cbd-strategy.ts

+    const decrypter = ThresholdDecrypter.create(
+      porterUri,
+      dkgRitual.id,
+      dkgRitual.dkgParams.threshold


Is threshold still needed, even though it is now part of the ritual struct?

The client needs to know how many shares to fetch from the nodes (/cbd_decrypt responses)

Yes, but now if you have the ritual id, you can get the value of the threshold from the Coordinator agent when doing the retrieve, or am I misunderstanding.

If you inspect the usage of DeployedCbdStrategy.create you can see that this is exactly what happens:

https://github.com/nucypher/nucypher-ts/blob/taco-api/src/sdk/strategy/cbd-strategy.ts#L47-L48

https://github.com/nucypher/nucypher-ts/blob/taco-api/src/sdk/strategy/cbd-strategy.ts#L95-L96

Right 👍 . I mixed up my layers.

Side question: any reason we need the DkgRitualParameters object? This line could look nice as:

const decrypter = ThresholdDecrypter.create( porterUri, dkgRitual.id, dkgRitual.threshold

? Anyways, just something to think about.

derekpierre · 2023-09-08T15:16:31Z

src/taco.ts

+  return await encryptLight(message, condition, dkgRitual.dkgPublicKey);
+};
+
+export const encryptLight = async (


Not sure how I feel about Light as the term 😅 . Can this just be an overloaded function with the same name?

There is no support function overloading in JS/TS. I didn't have better ideas for an appropriate name, maybe we could use the blockchainy naming from the original TACo API issue.

^ cc @cygnusv , @KPrasch, @jMyles (naming things).

I don't have a good name for it, but I would prefer something like

export const encryptWithDKG = async (

than the original one

test/unit/conditions/context.test.ts

TacoMessageKit was removed.

test/unit/conditions/context.test.ts

manumonti

LGTM. Great work! 👏

manumonti · 2023-09-08T17:22:46Z

src/taco.ts

+  return encrypter.encryptMessageCbd(toBytes(message), conditionExpr);
+};
+
+export const decrypt = async (


This simplicity is great!

derekpierre

🎸 - nice work

piotr-roslaniec force-pushed the taco-api branch from f340ff9 to e08cfc2 Compare August 11, 2023 14:12

nucypher deleted a comment from github-actions bot Aug 11, 2023

piotr-roslaniec force-pushed the set-default-variant branch from 752320c to 36fd43e Compare August 14, 2023 08:35

piotr-roslaniec force-pushed the taco-api branch from e08cfc2 to 60af38b Compare August 14, 2023 09:17

piotr-roslaniec force-pushed the taco-api branch from 60af38b to 2d6d92e Compare August 14, 2023 09:34

nucypher deleted a comment from github-actions bot Aug 14, 2023

Base automatically changed from set-default-variant to alpha August 14, 2023 14:04

nucypher deleted a comment from github-actions bot Aug 16, 2023

piotr-roslaniec force-pushed the taco-api branch from 2ccb4fa to b2cf2e4 Compare August 16, 2023 08:42

nucypher deleted a comment from github-actions bot Aug 16, 2023

piotr-roslaniec force-pushed the taco-api branch from 352ae8c to 7143d59 Compare August 24, 2023 12:01

nucypher deleted a comment from github-actions bot Aug 24, 2023

piotr-roslaniec force-pushed the taco-api branch from 7143d59 to db130b4 Compare August 28, 2023 08:44

piotr-roslaniec changed the base branch from alpha to nucypher-core-0.12.0 August 28, 2023 08:53

piotr-roslaniec commented Aug 28, 2023

View reviewed changes

src/taco.ts Outdated Show resolved Hide resolved

piotr-roslaniec marked this pull request as ready for review August 28, 2023 10:35

piotr-roslaniec requested review from KPrasch, cygnusv, derekpierre, jMyles, manumonti and theref August 28, 2023 10:35

theref approved these changes Aug 28, 2023

View reviewed changes

Base automatically changed from nucypher-core-0.12.0 to alpha August 28, 2023 15:25

nucypher deleted a comment from github-actions bot Aug 28, 2023

derekpierre requested changes Aug 29, 2023

View reviewed changes

src/characters/cbd-recipient.ts Outdated Show resolved Hide resolved

src/conditions/condition-expr.ts Outdated Show resolved Hide resolved

src/taco.ts Outdated Show resolved Hide resolved

piotr-roslaniec force-pushed the taco-api branch from 861f715 to 3046d55 Compare September 4, 2023 07:45

piotr-roslaniec requested a review from derekpierre September 4, 2023 07:45

piotr-roslaniec force-pushed the taco-api branch from 3046d55 to 5983ed0 Compare September 4, 2023 08:46

nucypher deleted a comment from github-actions bot Sep 4, 2023

piotr-roslaniec commented Sep 4, 2023

View reviewed changes

piotr-roslaniec force-pushed the taco-api branch 2 times, most recently from 8155c95 to 1921bc8 Compare September 4, 2023 11:15

derekpierre previously requested changes Sep 5, 2023

View reviewed changes

piotr-roslaniec commented Sep 5, 2023

View reviewed changes

src/agents/coordinator.ts Outdated Show resolved Hide resolved

piotr-roslaniec force-pushed the taco-api branch 2 times, most recently from 091a897 to 933416e Compare September 6, 2023 09:09

derekpierre reviewed Sep 8, 2023

View reviewed changes

derekpierre mentioned this pull request Sep 8, 2023

Implement "acceptance" type testing #295

Closed

2 tasks

piotr-roslaniec added 7 commits September 8, 2023 19:19

draft taco api

050bd60

draft light encryption method

3c6f0de

hide condition expression from the taco api

6ad6130

set a default porter uri

0319e0f

update after rebase

9f974c0

update coordinator contract

26a54e6

update after rebase

19a73b7

piotr-roslaniec force-pushed the taco-api branch from 933416e to f605e88 Compare September 8, 2023 17:33

derekpierre reviewed Sep 8, 2023

View reviewed changes

test/unit/conditions/context.test.ts Outdated Show resolved Hide resolved

test/unit/conditions/context.test.ts Outdated Show resolved Hide resolved

manumonti approved these changes Sep 8, 2023

View reviewed changes

apply pr suggestions

1dba09f

piotr-roslaniec force-pushed the taco-api branch from f605e88 to 1dba09f Compare September 9, 2023 07:26

piotr-roslaniec mentioned this pull request Sep 11, 2023

Migrate to a monorepo #294

Closed

9 tasks

derekpierre approved these changes Sep 11, 2023

View reviewed changes

remove redundant DkgRitualParameters struct

e1a0cd1

piotr-roslaniec merged commit ca9e163 into alpha Sep 12, 2023

piotr-roslaniec deleted the taco-api branch September 12, 2023 13:59

piotr-roslaniec mentioned this pull request Sep 18, 2023

Alter PRE-adapted tDec API design for CBD-DKG #166

Closed

		word0: dkgPublicKeyBytes.slice(0, 16),
		word1: dkgPublicKeyBytes.slice(16, 32),

Conversation

piotr-roslaniec commented Aug 11, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov-commenter commented Aug 14, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Uh oh!

theref left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

piotr-roslaniec Sep 4, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

derekpierre left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

derekpierre left a comment

Choose a reason for hiding this comment

Uh oh!

derekpierre Sep 8, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

piotr-roslaniec Sep 8, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

derekpierre Sep 8, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

piotr-roslaniec Sep 8, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

manumonti left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

piotr-roslaniec commented Aug 11, 2023 •

edited

Loading

codecov-commenter commented Aug 14, 2023 •

edited

Loading

piotr-roslaniec Sep 4, 2023 •

edited

Loading

derekpierre Sep 8, 2023 •

edited

Loading

piotr-roslaniec Sep 8, 2023 •

edited

Loading

derekpierre Sep 8, 2023 •

edited

Loading

piotr-roslaniec Sep 8, 2023 •

edited

Loading