Add API: sliding and slidingSizeStep
还没有人认领这个 Issue。
评估
- 难度
- 5/5
- 预计耗时
- 一周以上
- 新手友好度
- 20/100
- Issue 类型
- 功能
- 描述清晰度
- 需要澄清
- 活跃度
- 停滞
- 领域
- api
调研方向
Read the linked context in issue #212 and the discussion here to understand the proposed API alternatives. The issue asks for a decision on return types and proposes four sliding functions, but does not specify implementation locations, tests, or settled behavior. Done would require agreeing on the API and defining its behavior before implementation.
由索引模型根据 Issue 内容生成。
描述
Backlinking to #212 for more context. The question here is whether to add sliding, a function that returns windows of the array, and what its type signature should be. A variation of sliding is slidingSizeStep which allows one to control the size of the window and how far to step before creating a new window. We'll cover this function later in this issue.
Starting with sliding, the core question is what the return type should be. Below, I'll refer to various ideas for implementations using _N where N refers to the possible way forward.
-- as proposed in #212
_1 :: forall a. Array a -> Array (Tuple a a)
_2 :: forall a. Array a -> Array { previous :: Maybe a, current :: a, next :: Maybe a }
_3 :: forall a b. Array a -> (Tuple a a -> b) -> Array b
_1 is what is proposed in #212. _2 is what Harry proposed in his comment, functioning more like a Zipper. _3 is something I am proposing for unifying the other two. For example, if we implemented _1 and _3 in this library, _2 could still easily be implemented via (Tuple l r) -> {previous: Nothing, current: l, next: Just r}.
Since _2's zipper-like type might want additional features (e.g. a Functor instance) that is largely outside the scope of this project, I propose not implementing it as end-users can implement it and define their own Window type. Moreover, _3 enables more potential use cases we cannot imagine right now (e.g. slidingF ((Tuple l r) -> l + r) [0, 1, 2, 3] would produce [1, 3, 5]), simulating a scanl like feature.
Let's move on to slidingSizeStep. If we ignore the size and step arguments in the function, the question arises again as to what the return type should be. Here's two factors that come to mind:
- whether the returned value is
Array _orMaybe (NonEmptyArray _) - whether the returned value is the complete subcollection or a zipper-like value
-- as proposed in #212
_11 :: forall a. Array a -> Array (NonEmptyArray a)
_12 :: forall a. Array a -> Maybe (NonEmptyArray (NonEmptyArray a))
_13 :: forall a. Array a -> Array { before :: List a, current :: a, after :: List a }
_14 :: forall a b. Array a -> (NonEmptyArray a -> b) -> Array b
I believe _11 is to be preferred over _12 because one can still get Foldable1 and friends without having to pay for the Just constructor unwrapping in _12 by doing something like this:
case NEA.fromArray $ slidingSizeStep [0, 1, 2, 3] of
Nothing -> -- returned array is empty
Just nea -> fold1 nea -- returned array is non-empty
_13 provides the same zipper-like idea. However, it cannot be implemented because arrays does not depend on lists. Working similar to sliding, _4 provides an escape hatch to this problem. One could use _4 in combination with my own arrays-zipper to get an Array-based zipper or use pointed-list to get a List-based zipper.
In conclusion, I propose we implement these four functions:
sliding :: forall a. Array a -> Array (Tuple a a)
sliding = slidingF identity
slidingF :: forall a b. (Tuple a a -> b) -> Array a -> Array b
slidingF = ...
slidingSizeStep :: forall a. Int -> Int -> Array a -> Array (NonEmptyArray a)
slidingSizeStep size step = slidingSizeStepF size step identity
slidingSizeStepF :: forall a b. Int -> Int -> (NonEmptyArray a -> b) -> Array a -> Array b
slidingSizeStepF size step f arr = ...
I'm not yet considering rangeWithStep that was also proposed in #212 because I want to first cover these two functions and their variations.
- 主要语言
- PureScript
- 星标
- 58
- 派生
- 70
- 平均合并
- 2 天 22 小时
- 30 天内合并 PR
- 1
环境准备
- 没有 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 没有贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
purescript/purescript-arrays 的其他 Issue
-
难度 1/5 1 小时以内 新手友好度 55/100
purescript/purescript-arrays#246 ·
-
难度 2/5 1-3 小时 新手友好度 35/100
purescript/purescript-arrays#244 · 2 条评论 ·
-
难度 4/5 3-5 天 新手友好度 35/100
purescript/purescript-arrays#242 · 2 条评论 ·
-
难度 4/5 3-5 天 新手友好度 35/100
purescript/purescript-arrays#241 ·
-
`groupBy` has unexpected `l` argument for `a -> a -> Boolean` predicate可能重新可做 关联的 PR 已关闭且未合并。 未关闭
难度 2/5 1-3 小时 新手友好度 25/100
purescript/purescript-arrays#229 · 3 条评论 ·
查看 purescript/purescript-arrays 的全部 Issue
相似的 Issue
-
难度 2/5 1-3 小时 新手友好度 62/100
-
ManageSieve proxy: a backend's NO/BYE greeting is skipped as a capability line and reported as EOF未关闭bug
难度 2/5 1-3 小时 新手友好度 72/100
维护者通常 1 天内回复
-
Mend: dependency security vulnerability
难度 2/5 1-3 小时 新手友好度 65/100
-
Mend: dependency security vulnerability
难度 2/5 1-3 小时 新手友好度 62/100
ManageIQ/manageiq-pods#1476 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 62/100
prime-radiant-inc/evener#6366 ·
维护者通常 1 天内回复