如何使用Ramda更新JSON中任何级别的键值？

Question

有没有一种方法可以找到键的值，它可以出现在任何对象下的任何级别，并使用Ramda更新它？

例如，

JSON-1

{
  "query": {
    "boundary": {
      "type": "polygon",
      "coordinates": "[[-85.33604,35.055749],[-85.33604,35.07499772909699],[-85.279134,35.07499772909699],[-85.279134,35.055749],[-85.33604,35.055749]]"
    }
}

JSON-2

{
  "query": {
    "nearby": {
      "radius": "30mi",
      "coordinates": "[-121.40019800,38.55378300]"
    }
  }
}

在这两个JSON中，我想做一些类似的事情：

query.nearby.coordinates = JSON.parse(query.nearby.coordinates)

和

query.boundary.coordinates = JSON.parse(query.boundary.coordinates)

只有一种功能。

Answer 1

一种方法是遍历对象树，并尝试将找到的任何字符串解析为JSON。

const parseNestedJSON = R.cond([
   [R.is(String), R.tryCatch(JSON.parse, R.nthArg(1))],
   [R.is(Object), obj => R.map(parseNestedJSON, obj)],
   [R.T, R.identity],
])

请注意，这可能会执行一些不需要的转换，例如将{foo: '1'}转换为{foo: 1} (string to number)。

只针对名为coordinates的嵌套键可能更安全

const parseNestedJSON = R.cond([
    [R.has('coordinates'), R.over(R.lensProp('coordinates'), JSON.parse)],
    [R.is(Object), obj => R.map(parseNestedJSON, obj)],
    [R.T, R.identity],
])

编辑:如果coordinates可能不是json，您也可以在这里使用tryCatch。

  [R.has('coordinates'), R.over(R.lensProp('coordinates'), R.tryCatch(JSON.parse, R.nthArg(1)))]

用法：

parseNestedJSON({"nearby": {"coordinates": "[-121.40019800,38.55378300]"}})
=> { nearby: { coordinates: [ -121.400198, 38.553783 ] } }

Answer 2

另一种选择是定义一个可以负责更新值的镜头。

第一种方法假设可以找到坐标的已知路径的数量是有限的。

​

// Creates a lens that accepts a list of paths and chooses the first
// matching path that exists in the target object

const somePathLens = paths => toFunctor => target => {
  const path = R.find(
    p => R.pathSatisfies(x => x != null, p, target),
    paths
  )
  return R.map(
    value => R.assocPath(path, value, target),
    toFunctor(R.path(path, target))
  )
}

// R.over can then be used with JSON.parse to parse the first
// matching path that is found.

const parseCoords = R.over(
  somePathLens([
    ['query', 'boundary', 'coordinates'],
    ['query', 'nearby', 'coordinates']
  ]),
  JSON.parse
)

console.log(parseCoords({
  "query": {
    "boundary": {
      "type": "polygon",
      "coordinates": "[[-85.33604,35.055749],[-85.33604,35.07499772909699],[-85.279134,35.07499772909699],[-85.279134,35.055749],[-85.33604,35.055749]]"
    }
  }
}))

console.log(parseCoords({
  "query": {
    "nearby": {
      "radius": "30mi",
      "coordinates": "[-121.40019800,38.55378300]"
    }
  }
}))

<script src="//cdnjs.cloudflare.com/ajax/libs/ramda/0.25.0/ramda.min.js"></script>

​

如果实际路径未知，并且您只需要通过给定的键找到第一个值，则可以使用第二种方法。

​

// Recursively search for the given key in an object, returning the
// first matching path if found.

const findKeyInPath = (keyToFind, obj) => {
  const findKeyInPath_ = o =>
    R.has(keyToFind, o)
      // if found, return this key as the path
      ? [keyToFind]
      // otherwise find all keys with objects and recursively
      // call this function.
      : R.reduceRight((k, acc) => {
          // find either the subpath of this key, or the subpath
          // found in the remaining keys
          const subPath = R.when(R.isEmpty, _ => acc, findKeyInPath_(o[k]))
          // if the subpath contains a key, prepend it with the
          // current key, otherwise return the empty list
          return R.unless(R.isEmpty, R.prepend(k), subPath)
        }, [], R.filter(k => R.propIs(Object, k, o), R.keys(o)))
  return findKeyInPath_(obj)
}

// Create a lens that recursively searches for the first matching
// key within a target object.

const someKeyLens = key => toFunctor => target => {
  // find the path using our new `findKeyInPath` function
  const path = findKeyInPath(key, target)
  return R.map(
    value => R.assocPath(path, value, target),
    toFunctor(R.path(path, target))
  )
}

const parseCoords = R.over(
  someKeyLens('coordinates'),
  JSON.parse
)

console.log(parseCoords({
  "query": {
    "boundary": {
      "type": "polygon",
      "coordinates": "[[-85.33604,35.055749],[-85.33604,35.07499772909699],[-85.279134,35.07499772909699],[-85.279134,35.055749],[-85.33604,35.055749]]"
    }
  }
}))

console.log(parseCoords({
  "query": {
    "nearby": {
      "radius": "30mi",
      "coordinates": "[-121.40019800,38.55378300]"
    }
  }
}))

<script src="//cdnjs.cloudflare.com/ajax/libs/ramda/0.25.0/ramda.min.js"></script>

​

值得一提的是，只有在保证在目标对象中找到路径的情况下，这些才是有效的镜头，否则行为是未定义的。