需要一个Django Tastypie的递归CSV序列化器

Question

我为我的Django应用程序创建了一个简单的只支持GET请求的API，使用了tastypie这个工具。我需要提供扁平化的表格形式的CSV数据，但我的数据库结构是经过规范化的。根据文档，我实现了一个自定义的Serializer类，并添加了一个to_csv()方法，如下所示。

def to_csv(self, data, options=None):
    options = options or {}
    data = self.to_simple(data, options)

    raw_data = StringIO.StringIO()
    writer = csv.writer(raw_data, quotechar="'", quoting=csv.QUOTE_NONNUMERIC)

    if "meta" in data.keys():#if multiple objects are returned
        objects = data.get("objects")
        writer.writerow(objects[0].keys())

        for object in objects:
            test = object.values()
            writer.writerow(test)
    else:
        writer.writerow(data.values())

    CSVContent=raw_data.getvalue()
    return CSVContent

这个方法运行得很好，但默认情况下，任何资源都会以JSON格式呈现（当我在ModelResource的ForeignKey定义中包含full = True时），所以我得到的CSV数据中包含了嵌套的JSON数据，像这样。

foodID,foodName,related_details
1,"apricot","{'type':'fruit', 'cost':'medium'}"
2,"beef","{'type':'animal', 'cost':'high'}"
3,"celery","{'type':'vegetable', 'cost':'low'}"

我想要的输出是

foodID,foodName,type,cost
1,"apricot","fruit","medium"
2,"beef","animal","high"
3,"celery","vegetable","low"

我有一个想法，可能需要递归地应用我的序列化器，然后在写入CSV之前将结果合并，但到目前为止还没有成功。

Answer 1

你可以在每一行上使用下面这个函数 to_list：

def to_list(line):
    idx = -1
    for i, l in enumerate(line):
        if type(l) is str and '{' in l and '}' in l:
            idx = i
            break
    if idx != -1:
        result = line[:idx] + eval(line[idx]).values() + line[idx+1:]
    else:
        result = line
    return result

if __name__ == "__main__":
    lst = [[1,"apricot","{'type':'fruit', 'cost':'medium'}"],
           ["beef","{'type':'animal', 'cost':'high'}", 3],
           ["meat", "sugar"],
           ["{'type':'car', 'cost':'nothing'}", "something"]]
    for line in lst:
        print to_list(line)

那么对于以下这些列表：

1, "apricot", "{'type':'fruit', 'cost':'medium'}"
"beef", "{'type':'animal', 'cost':'high'}", 3
"meat", "sugar"
"{'type':'car', 'cost':'nothing'}", "something"

你会得到：

1, 'apricot', 'medium', 'fruit'
'beef', 'high', 'animal', 3
'meat', 'sugar'
'nothing', 'car', 'something'

正如你所看到的，这个方法不管元素的数量，也不管JSON字符串的位置。

Answer 2

def to_csv(self, data, options=None):
    options = options or {}
    data = self.to_simple(data, options)

    raw_data = StringIO.StringIO()
    first = True

    if "meta" in data.keys():#if multiple objects are returned
        objects = data.get("objects")

        for value in objects:
            test = {}
            self.flatten(value, test)
            if first:
                writer = csv.DictWriter(raw_data, test.keys(), quotechar="'", quoting=csv.QUOTE_NONNUMERIC)
                writer.writeheader()
                writer.writerow(test)
                first=False
            else:
                writer.writerow(test)
    else:
        test = {}
        self.flatten(data, test)
        if first:
            writer = csv.DictWriter(raw_data, test.keys(), quotechar="'", quoting=csv.QUOTE_NONNUMERIC)
            writer.writeheader()
            writer.writerow(test)
            first=False
        else:
            writer.writerow(test)
    CSVContent=raw_data.getvalue()
    return CSVContent

def flatten(self, data, odict = {}):
    if isinstance(data, list):
        for value in data:
            self.flatten(value, odict)
    elif isinstance(data, dict):
        for (key, value) in data.items():
            if not isinstance(value, (dict, list)):
                odict[key] = value
            else:
                self.flatten(value, odict)

当然可以！请把你想要翻译的内容发给我，我会帮你把它变得更简单易懂。