Python 请求：处理 JSON 响应、存储到列表或字典？

Question

我正在使用 python 请求库从 api 获取数据。数据以称为消息的大型 json 数组返回。其中包含许多单独的“消息”子级别 jsons（请参阅底部的 json 响应示例）。在这个 json 响应中，对于每条消息，我只关心几个（2 或 3）个数据点。我需要获取这几个数据点，并将它们存储到某些东西（列表列表、字典等）中，以便以后可以引用它，存储到一个对象并从另一个函数中使用。我需要存储的数据点是id,conversationId和body. The id is unique, while theconversationId` 在对话中的所有消息之间共享，因此不是唯一的。我首先想问一下您认为实现这一目标的最佳方法是什么。我猜是列表列表还是列表字典？我太新了，不知道哪个是更好的解决方案。此外，无论选择什么方法，我都需要知道如何引用它们并通过 id 调用它们以从另一个函数存储到对象。我还没有弄清楚如何成功地做到这一点。以下是我尝试过的一些事情，只是为了大致了解如何做到这一点：response=requests.get(url + id, headers=h, params=p)messages=json.loads(response.text)for message in messages:    print(message['body'])^^ 在这里我只是想看看我是否可以引用特定消息的正文，没有工作。r=requests.get(url + id, headers=h, params=p)inbound_dict = {}inbound=json.loads(r.text)for item in inbound['messages']:    inbound_dict[item['conversationId']] = item['body']print(inbound_dict)^^ 这个确实有点作用，但不允许我有效地组织数据以便稍后调用。当我打印字典时，它会显示最新的值，因为键不是唯一的......所以它是覆盖而不是附加。这就是让我觉得最好列出清单的原因。最后，我想要一个解决方案，将数据组织conversationId在字典或类似结构中的类似结构中，我可以通过conversationId、 或引用消息msgId，以及一种简洁易读的方式来存储所有数据......:)Messages     |___                  msgId:         |_conversationId-[         |                 body         |                 msgId:         |_conversationId-[         |                 body         |                 msgId:         |_conversationId-[                           body最后，这是一个 json 示例。请记住，我仍在学习和掌握 python 的速度。谢谢大家的时间！

Answer 1

您可以使用这样的组合列表和字典理解来做到这一点：

import json

from pprint import pprint

response = b'{"id":1005672,"messages":[{"id":4461048,"body":"Mnow test test","conversationId":1005672,"locationId":2045,"contactId":12792806,"assignedUserId":0,"status":"RECEIVED","error":null,"kind":"INCOMING","outgoing":false,"reviewRequest":false,"type":"SMS","readDate":0,"respondedDate":0,"sentDate":1576783232355,"attachments":[]},{"id":4461049,"body":"THIS NUMBER DOES NOT CURRENTLY ACCEPT TEXT MESSAGES PLEASE CALL [phone] TO WORK WITH ONE OF OUR INTAKE SPECIALISTS","conversationId":1005672,"locationId":2045,"contactId":12792806,"assignedUserId":0,"status":"RECEIVED","error":null,"kind":"AUTO_RESPONSE","outgoing":true,"reviewRequest":false,"type":"SMS","readDate":0,"respondedDate":0,"sentDate":1576783233546,"attachments":[]},{"id":4620511,"body":"test sms,test sms","conversationId":1005672,"locationId":2045,"contactId":12792806,"assignedUserId":17297,"status":"DELIVERED","error":null,"kind":"API","outgoing":true,"reviewRequest":false,"type":"SMS","readDate":0,"respondedDate":0,"sentDate":1577987093930,"attachments":[]}]}'

data = json.loads(response)

messages = [

    {'id': message['id'],

     'conversationId': message['conversationId'],

     'body': message['body']} for message in data['messages']

]

pprint(messages, sort_dicts=False)

输出：

[{'id': 4461048, 'conversationId': 1005672, 'body': 'Mnow test test'},

 {'id': 4461049,

  'conversationId': 1005672,

  'body': 'THIS NUMBER DOES NOT CURRENTLY ACCEPT TEXT MESSAGES PLEASE CALL '

          '[phone] TO WORK WITH ONE OF OUR INTAKE SPECIALISTS'},

 {'id': 4620511, 'conversationId': 1005672, 'body': 'test sms,test sms'}]

您可以使处理更加数据驱动，并消除理解中的大量重复编码，从而使其更加简洁，如下所示：

import json

from pprint import pprint

data_points = 'id', 'conversationId', 'body'

response = b'{"id":1005672,"messages":[{"id":4461048,"body":"Mnow test test","conversationId":1005672,"locationId":2045,"contactId":12792806,"assignedUserId":0,"status":"RECEIVED","error":null,"kind":"INCOMING","outgoing":false,"reviewRequest":false,"type":"SMS","readDate":0,"respondedDate":0,"sentDate":1576783232355,"attachments":[]},{"id":4461049,"body":"THIS NUMBER DOES NOT CURRENTLY ACCEPT TEXT MESSAGES PLEASE CALL [phone] TO WORK WITH ONE OF OUR INTAKE SPECIALISTS","conversationId":1005672,"locationId":2045,"contactId":12792806,"assignedUserId":0,"status":"RECEIVED","error":null,"kind":"AUTO_RESPONSE","outgoing":true,"reviewRequest":false,"type":"SMS","readDate":0,"respondedDate":0,"sentDate":1576783233546,"attachments":[]},{"id":4620511,"body":"test sms,test sms","conversationId":1005672,"locationId":2045,"contactId":12792806,"assignedUserId":17297,"status":"DELIVERED","error":null,"kind":"API","outgoing":true,"reviewRequest":false,"type":"SMS","readDate":0,"respondedDate":0,"sentDate":1577987093930,"attachments":[]}]}'

data = json.loads(response)

messages = [{dp: message.get(dp) for dp in data_points}

                for message in data['messages']]

pprint(messages, sort_dicts=False)

Answer 2

我的理解是，您希望您的消息采用可通过对话检索的数据结构。这是我要做的：

from pprint import pprint

#with this data structure whenever we refer to a conversation, if it

#doesn't exist, it gets created

from collections import defaultdict

inbound_dict = defaultdict(dict) 

inbound=json.loads(data)

for item in inbound['messages']:

    print (item)

    current_conversation=inbound_dict[item['conversationId']]

    #inbound_dict retrives the apropriate conversation,

    #   or creates a new one for us to fill

    current_conversation[item["id"]] = item['body'] #add our item to it.

    #or if there's a chance we might want *everything* else about the

    #message later even if just the date in order to preserve

    #conversation ordering or whatever:

    #currentconversation[item["id"]] = item

pprint(inbound_dict)

但这可能是矫枉过正，这取决于你以后要做什么处理，以及它是什么处理。如果您只是让他们选择一个对话，并显示最后 20 条消息，那么可切片的列表可能是内部数据结构的最佳选择，在这种情况下，我会这样做：

from pprint import pprint

#with this data structure whenever we refer to a conversation, if it

#doesn't exist, it gets created

from collections import defaultdict

inbound_dict = defaultdict(list) 

inbound=json.loads(data)

for item in inbound['messages']:

    print (item)

    current_conversation=inbound_dict[item['conversationId']]

    #inbound_dict retrives the apropriate conversation,

    #   or creates a new one for us to fill

    current_conversation.append(

        (item["id"], item['body'])

        ) # here we add our item to it, in this case a tuple of id and body

    #or if there's a chance we might want *everything* else about the

    #message later even if just the date in order to preserve

    #conversation ordering or whatever:

    #currentconversation.append(item)

pprint(inbound_dict)

基本相同的操作，但 defaultdict 为我们提供了不同类型的内部集合来填充。

Answer 3

如果您可以拥有多个具有相同会话 ID 的项目，那么您可以执行以下操作：

r=requests.get(url + id, headers=h, params=p).json()

inbound_dict = {}

for item in r['messages']:

    conv_id = item['conversationId']

    if conv_id not in inbound_dict:

        inbound_dict[conv_id]=[{'msg_id' : item['id'], 'body' : item['body']}]

    else:

        inbound_dict[conv_id].append({'msg_id' : item['id'], 'body' : item['body']})

print(inbound_dict)

生成的数据结构是一个以 conversation_id 作为键的字典，每个 conversation_id 映射到一个list项目dict。每个项目都存储特定消息的 message_id 和正文。然后，您可以通过检索为 conv_id 键存储的消息列表来迭代特定对话的消息。

或者，您可以选择以下数据结构进行映射： {conv_id -> { message_id : {message info ...}, ...}.

这可以像这样实现：

r=requests.get(url + id, headers=h, params=p).json()

inbound_dict = {}

for item in r['messages']:

    conv_id = item['conversationId']

    if conv_id not in inbound_dict:

        inbound_dict[conv_id]={item['id'] : {'msg_id' : item['id'], 'body' : item['body']}}

    else:

        inbound_dict[conv_id][item['id']] = {'msg_id' : item['id'], 'body' : item['body']}

print(inbound_dict)

在这种情况下，如果您知道 con_id 和 message_id，则可以直接从对话中访问消息。

所以它真的取决于这个数据结构的下游实用程序。

请注意，上述内容也可以通过列表推导来完成。