goabstract / abstract-sdk Goto Github PK

I created a loop to save images of all my artboards on all pages. For some images I get a not found error. I looked inside Sketch but I couldn't identify why some artboards work and some won't.

(node:43291) UnhandledPromiseRejectionWarning: NotFoundError: Not found. at throwAPIError (/Users/Jonathan.Arnold/Desktop/_spikeweek_test/node_modules/abstract-sdk/lib/errors.js:143:13) at AbstractAPI.fetch (/Users/Jonathan.Arnold/Desktop/_spikeweek_test/node_modules/abstract-sdk/lib/AbstractAPI/index.js:509:39) at processTicksAndRejections (internal/process/next_tick.js:81:5) (node:43291) UnhandledPromiseRejectionWarning: Unhandled promise rejection. This error originated either by throwing inside of an async function without a catch block, or by rejecting a promise which was not handled with .catch(). (rejection id: 2) (node:43291) [DEP0018] DeprecationWarning: Unhandled promise rejections are deprecated. In the future, promise rejections that are not handled will terminate the Node.js process with a non-zero exit code.

@tommoor my team at Freeletics want to use that feature for a very quick project that has to be completed till friday. I would really appreciate if you can look into that and provide some insights.

Thank you :-)

previews.raw is great if you want a copy of the file to save in your own datastore, but it's an awkward interface for displaying in the web browser.

Lets add an additional method to grab a data-uri for the image using the bufferToUrl helper we have internally. This issue was triggered by confusion from a customer attempting to use the webUrl method…

The API and CLI transports currently return different data when querying collections. The CLI returns the collection along with its file data, whereas the API only returns the collection.

I'd suggest the following but am of course open to other things:

• First pass: Add a custom error that is thrown when the rate limit is hit
• Second pass: Internally handle the rate limit and wait for the appropriate amount of time based on the response headers before continuing.

We should consider exposing the CLI's asset generation functionality via the CLI transport. See #72 (comment).

The static SDK site includes reference guide documentation and API documentation, both of which are hand-curated. Manually updating API documentation in this manner can be tedious over time and lead to a divergence in documentation and implementation. We should investigate what solutions exist to generate API documentation.

JSDoc is a viable option, but it doesn't infer types from Flow type annotations. This means we'd have to document types in two places: in the actual JS code for Flow, and also in comments for JSDoc. The TypeScript community created versions of JSDoc that infer types from TS type annotations (see typedoc, but no similar tools yet exist for Flow.

When including [email protected] (failed to install a more recent version as per #101)

Seems like abstract has undeclared dependencies on core-js and regenerator-runtime packages. But even when manually installing these, the simplest-possible inclusion of the abstract sdk fails with a ReferenceError.

Abstract.app Version 77.0.1 (77.0.1.12639)
node v11.10.1
macOS 10.14.3 (18D109)

Steps to reproduyce:

index.js:

const Abstract = require("abstract-sdk");

> yarn add [email protected]
[…]
> node index.js
internal/modules/cjs/loader.js:615
    throw err;
    ^

Error: Cannot find module 'core-js/modules/es6.object.define-property'
[…]

> yarn add core-js
[…]
> node index.js

internal/modules/cjs/loader.js:615
    throw err;
    ^

Error: Cannot find module 'regenerator-runtime/runtime'
[…]

> yarn add regenerator-runtime
[…]
> node index.js
/Users/gerrit/spotify/abstracttest/node_modules/abstract-sdk/lib/AbstractCLI/index.js:576
      regeneratorRuntime.mark(function _callee14(args) {
      ^

ReferenceError: regeneratorRuntime is not defined
    at /Users/gerrit/spotify/abstracttest/node_modules/abstract-sdk/lib/AbstractCLI/index.js:576:7
    at /Users/gerrit/spotify/abstracttest/node_modules/abstract-sdk/lib/AbstractCLI/index.js:646:6
    at Object.<anonymous> (/Users/gerrit/spotify/abstracttest/node_modules/abstract-sdk/lib/AbstractCLI/index.js:698:2)
    at Module._compile (internal/modules/cjs/loader.js:738:30)
    at Object.Module._extensions..js (internal/modules/cjs/loader.js:749:10)
    at Module.load (internal/modules/cjs/loader.js:630:32)
    at tryModuleLoad (internal/modules/cjs/loader.js:570:12)
    at Function.Module._load (internal/modules/cjs/loader.js:562:3)
    at Module.require (internal/modules/cjs/loader.js:667:17)
    at require (internal/modules/cjs/helpers.js:20:18)

We are currently investigating using the Abstract SDK to extract values from the design files for use by developers, and we have reached a bottleneck with the current approach. Each value we want to extract is in a separate layer object, which requires a separate API call via abstract.data.info. We have hit rate limiting issues with only a few dozen items. Is there any way to batch data requests for multiple layers into one request?

We should modularize the SDK endpoints and remove the hardcoded AbstractSDK interface.

Loading local branches is a definite requirement.

https://sdk.goabstract.com/docs/reference/#branches

We should add new endpoints to the CLI transport to support listing all assets, getting one asset, and getting one asset's raw data. This will involve new CLI commands to support all operations.

We should add asynchronous iteration support for paginated list methods using an external function that leverages the cursor-based pagination we have in place.

I'd love the ability to filter data by user role, i.e Admin vs Contributor vs Viewer. I can make some educated guess by analyzing each users Activity by type, but since we are already able to sort by these fields in app via the People tab, I figured I would inquire about the feasibility of getting this added to the API first. 😎

Hey,
Is it possible to accomplish right now, or is this a feature request?

Hey, I'm currently getting UnhandledPromiseRejectionWarning: TypeError: Cannot read property 'list' of undefined on activities.list.

is this part of the API not implemented yet? or am I doing something wrong?

It would be cool to download a Java SDK of the Abstract SDK that can make these HTTPS calls to any Abstract information you're looking for. The CLI environment is nice to find layers and files locally, but i would like to accessing my Abstract information when my program is being executed on an AWS remote server.

For example, Amazon Web Services provide and CLI interface and a Java SDK library to access my AWS information.

Not sure if this is in the plan for the Abstract team, but I think this feature would be handy.

#40 (comment)

• comments.create
• comments.list
• comments.info

• projects.info is not documented
• projects.list does not include the options that may be passed

https://spectrum.chat/abstract/developers/is-there-a-way-to-get-a-list-of-just-active-projects~af00c66a-c415-4332-961e-f0c0ab660dfc

changesets.info is implemented, but undocumented.

We often have work sat on master that hasn't yet been published to NPM as we're working towards a release. As it stands right now documentation updates related to this work get auto-deployed with master.

We've discussed:

• Deploying docs when a release is tagged. This one is great in that it requires no thinking but prevents us from updating docs outside of the release cycle.
• Deploying docs by pushing to a branch. This requires manual deployment but we can do so at any time.

Currently swaying towards the branch approach.

Requires internal changes first

Many of the SDK use cases identified so far rely on the ability to take programmatic action in response to different events within the Abstract application.

Once web hook support lands (which is prioritized very highly), we should document using them with the SDK. It would be nice to both document usage and to show a tutorial that walks through a real-world web hook-powered SDK script, like one that extracts assets from new commits.

Hey,

I'm trying to get the preview to work - to get an png of a layer.

I'm using preview.info

Getting https://previews.goabstract.com/projects/eda68300-e1d0-11e8-8133-0b70a3eae7f0/commits/51bf03f17a6234d6f703b86d2309fe5ae4a28034/files/D06C3330-664C-4F89-B8E6-AC27B93AF4DC/layers/4DF3A97D-3F66-48A1-9459-B42B3154391C

then when i try to access it I get {"error":"forbidden"}

maybe the documentation is unclear or I got something wrong.

Thanks in advance!

The backend now supports loading branchName, layerName etc so they do not need to be passed to comment.create any more 👏

We should add documentation that details rate limits when using the API transport.

• activities.list (see project, org and more filters available in API controller)
• activities.info

Running yarn add abstract-sdk (as per https://sdk.goabstract.com/docs/installation/) fails with this error message: error Couldn't find package "@elasticprojects/abstract-cli@^1.0.0" required by "abstract-sdk@^1.0.0-beta.8" on the "npm" registry.

manually pinning to 0.7.4 downgrades this to a warning (`warning " > [email protected]" has unmet peer dependency "@elasticprojects/abstract-cli@^1.0.0".)

Seems like the public abstract-sdks packages are depending on a private NPM package?

We should add new endpoints to the API transport to support the creation of and updating of collections. The required API routes for this already exist, so it should just be a matter of adding these calls to the SDK.

I intermitently get InternalServerErrors when requesting layer data using the SDK. Retrying the same request will, most of the time, return correct data.

I realise this is an issue with the infrastructure behind it rather than the SDK itself, but the SDK is the entry point and I wasn't sure where else to raise this as an issue.

We should cache requests where possible to improve perceived performance and to reduce load on our servers.

Rate limiting is now in place – we're able to open this up more broadly 🎉

We should add new endpoints to the CLI transport to support the creation of and updating of collections. This will involve new CLI commands to support both operations.

Would be nice to remove pageId from LayerDescriptor

• Layer routes
• Share links
• comments.create
• LayerDescriptor

Our current documentation for authentication references API tokens but doesn't tell you how to go about getting one once you've been given access. 🙈

We should fix that somehow.

Ideally implemented at the server level. This request came directly from a customer using the SDK, that it would be helpful to limit the data returned to, for example just symbols, or just artboards.

It's missing right now

https://github.com/goabstract/abstract-sdk/blob/master/src/AbstractAPI/index.js#L196-L201

Nothing to see here!

We should add a new endpoint to both the CLI transport and the API transport to support retrieving the raw data for a single file. This will involve a new API and CLI command to support each respective operation.

We're currently missing endpoints for users.list and users.info – as we don't send this information back with branch data (#45) it's important that it can be loaded separately.

Add user.Avatar as part of branch.list/commit.list

• notifications.list (ability to filter by organizationId, read, unread)
• notifications.info

Hi Everyone,

Sadly the issue under #73 was closed without any notice or updates in the meantime. We are still waiting for some way to extract the raw sketch files from Abstract via the API. Is there any roadmap/plan to allow this in the future? Is there any way I could support in getting this feature done?

Since the API and CLI transports should be able to be used interchangeably for common methods, they should both return the same custom errors where applicable.

Related to #56

While the SDK will work in both a browser and a node context from a JavaScript perspective, the backing API that the SDK uses by default at https://api.goabstract.com still has CORS enabled and can't be accessed by external domains at this time.

We should enable CORS for SDK-supported endpoints and document any necessary configuration.

Currently list endpoints only return the first page of results

• assets.list
• assets.info
• assets.raw

We should add a new endpoint to the API transport to support the creation of share links. The required API route for this already exists, so it should just be a matter of adding this call to the SDK.

I want to start using the Abastract API but after using it in my project I get the following error:

Module not found: Error: Can't resolve 'core-js/modules/es7.symbol.async-iterator' in '/Users/developer/review-tracker-2019/node_modules/abstract-sdk/lib'

I googled the issue and solutions like installing core-js doesn't seem to work. Any idea on how tho fix this? Using the latest Abstract verison 1.0.0-beta.11.

There are some details and webpack config that are important, mostly contained within the blame sample project right now:

https://github.com/goabstract/abstract-blame

depends on #24