为 PyV8 将 Python 对象转换为 JavaScript

本文介绍了为 PyV8 将 Python 对象转换为 JavaScript的处理方法，对大家解决问题具有一定的参考价值，需要的朋友们下面随着小编来一起学习吧！

问题描述

我正在尝试将 Python 数据(列表、字典、字符串...，任意嵌套)传递给 PyV8.

I'm trying to pass Python data (lists, dicts, strings..., arbitrarily nested) to PyV8.

class Global(object):
    def __init__(self, data):
        self.data = data
ctx = PyV8.JSContext(Global([{'a':1}]))
ctx.enter()
res = ctx.eval('data.length')
js_len = PyV8.convert(res)
print js_len

上面的代码打印 None，大概是因为数据对象没有转换为 JSArray，因此 data.length 计算结果为 undefined.除了使用 JSON 之外，还有其他可靠的方法可以在 PyV8 中进行必要的转换吗?

The code above prints None, presumably because the data object is not transformed to a JSArray and thus data.length evaluates to undefined. Is there a reliable way to do the necessary conversion in PyV8 other than using JSON?

推荐答案

显然 PyV8 没有正确地将 python 列表转换为 Javascript 数组，导致 my_list.length 返回 undefined，它正在转换为 None.

Apparently PyV8 doesn't correctly convert python lists to Javascript arrays, which leads my_list.length to return undefined, which is getting converted to None.

ctx = PyV8.JSContext()
ctx.enter()
ctx.locals.a = [{'a':1}]
print ctx.locals.a
#> [{'a': 1}]
print ctx.eval("a.length")
#> None
print ctx.eval("a[0].a")
#> 1
ctx.locals.blub = {'a':1}
print ctx.eval("blub.a")
#> 1
print ctx.eval("Object.keys(blub)")
#> a
ctx.locals.blub = {'a':[1,2,3]}
print ctx.eval("Object.keys(blub)")
#> a
print ctx.eval("blub.a")
#> [1, 2, 3]
ctx.locals.blub2 = [{'a':[1,2,3]}]
print ctx.eval("blub2")
#> [{'a': [1, 2, 3]}]
print ctx.eval("blub2.length")
#> None
print ctx.eval("Array.isArray(blub2)")
#> False
print ctx.eval("typeof(blub2)")
#> object
print ctx.eval("blub2[0].a")
#> [1, 2, 3]
print ctx.eval("typeof(blub.a)")
#> object
print ctx.eval("Array.isArray(blub.a)")
#> False

答案是使用PyV8.JSArray(my_list).我为我的项目编写了以下帮助函数，用于处理各种小问题，并可以轻松地在 python 和 js 对象之间来回转换.然而，这些是针对特定版本的 PyV8(这是我唯一可以推荐的版本，请参阅链接问题中的讨论)，因此如果您按原样使用它们，结果可能会有所不同.示例用法:

The answer is to use PyV8.JSArray(my_list). I've written the following helper functions for my project that deal with various little problems and make it easy to convert back and forth between python and js objects. These are targeted at a specific version of PyV8 however (which is the only version I can recommend, see discussion in the linked issues), so your results may vary if you use them as-is. Example usage:

ctx.locals.blub3 = get_js_obj({'a':[1,2,3]})
ctx.locals.blub4 = get_js_obj([1,2,3])
ctx.eval("blub3.a.length")
#> 3
ctx.eval("blub4.length")
#> 3

这里是函数.

def access_with_js(ctx, route):
    if len(route) == 0:
        raise Exception("route must have at least one element")
    accessor_string = route[0]
    for elem in route[1:]:
        if type(elem) in [str, unicode]:
            accessor_string += "['" + elem + "']"
        elif type(elem) == int:
            accessor_string += "[" + str(elem) + "]"
        else:
            raise Exception("invalid element in route, must be text or number")
    return ctx.eval(accessor_string)

def get_py_obj(ctx, obj, route=[]):
    def dict_is_empty(dict):
        for key in dict:
            return False
        return True

    def access(obj, key):
        if key in obj:
            return obj[key]
        return None

    cloned = None
    if isinstance(obj, list) or isinstance(obj, PyV8.JSArray):
        cloned = []
        temp = str(access_with_js(ctx, route)) #working around a problem with PyV8 r429
        num_elements = len(obj)
        for index in range(num_elements):
            elem = obj[index]
            cloned.append(get_py_obj(ctx, elem, route + [index]))
    elif isinstance(obj, dict) or isinstance(obj, PyV8.JSObject):
        cloned = {}
        for key in obj.keys():
            cloned_val = None
            if type(key) == int:
                #workaround for a problem with PyV8 where it won't let me access
                #objects with integer accessors
                val = None
                try:
                    val = access(obj, str(key))
                except KeyError:
                    pass
                if val == None:
                    val = access(obj, key)
                cloned_val = get_py_obj(ctx, val, route + [key])
            else:
                cloned_val = get_py_obj(ctx, access(obj, key), route + [key])
            cloned[key] = cloned_val
    elif type(obj) == str:
        cloned = obj.decode('utf-8')
    else:
        cloned = obj
    return cloned

def get_js_obj(ctx,obj):
    #workaround for a problem with PyV8 where it will implicitely convert python lists to js objects
    #-> we need to explicitely do the conversion. see also the wrapper classes for JSContext above.
    if isinstance(obj, list):
        js_list = []
        for entry in obj:
            js_list.append(get_js_obj(ctx,entry))
        return PyV8.JSArray(js_list)
    elif isinstance(obj, dict):
        js_obj = ctx.eval("new Object();") # PyV8.JSObject cannot be instantiated from Python
        for key in obj.keys():

            try:
                js_obj[key] = get_js_obj(ctx,obj[key])
            except Exception, e:
                # unicode keys raise a Boost.Python.ArgumentError 
                # which can't be caught directly:
                # https://mail.python.org/pipermail/cplusplus-sig/2010-April/015470.html
                if (not str(e).startswith("Python argument types in")):
                    raise
                import unicodedata
                js_obj[unicodedata.normalize('NFKD', key).encode('ascii','ignore')] = get_js_obj(ctx,obj[key])
        return js_obj
    else:
        return obj

这篇关于为 PyV8 将 Python 对象转换为 JavaScript的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！

pyV8