如何识别推文是回复还是新提及？Twitter Bot开发求助

问题说明 #

开发一款Twitter Bot，目标是回复所有提及它的用户，但需要添加判断逻辑：仅当提及是针对Bot原推文的回复时才回复。目前使用Tweepy实现，但尝试通过in_reply_to_status_id判断时无效。

原实现代码 #

import tweepy
import time
from datetime import datetime


# Connecting to Twitter API
client = tweepy.Client(bearer_token, api_key, api_secret, access_token, access_token_secret)
auth = tweepy.OAuth1UserHandler(api_key, api_secret, access_token, access_token_secret)
api = tweepy.API(auth)

# Message to reply with if someone mentions the bot
message = "Hi, Thank you so much for reaching TalkyTweety Out :) Please reply to continue this conversation. "

# Bot's unique ID
client_id = client.get_me().data.id

# This is so the bot only looks for the most recent tweets.
start_id = 1
initialisation_resp = client.get_users_mentions(client_id)

if initialisation_resp.data != None:
    start_id = initialisation_resp.data[0].id

# Looking for mentions tweets in an endless loop
while True:
    response = client.get_users_mentions(client_id, since_id=start_id)
    # Reply Code
    if response.data != None:
        for tweet in response.data:
            try:
                print(tweet.text)
            
                client.create_tweet(in_reply_to_tweet_id=tweet.id, text=message)
                start_id = tweet.id
            except Exception as error:
                print(error)

    # Delay (so the bot doesn't search for new tweets a bunch of times each second)
    time.sleep(5)

解决方案 #

问题出在Tweepy v2 API的get_users_mentions方法默认不会返回in_reply_to_status_id和in_reply_to_user_id字段，需要显式指定tweet_fields参数获取这些信息。判断逻辑可通过以下步骤实现：

1. 在调用get_users_mentions时，添加tweet_fields参数，请求回复相关字段
2. 检查提及推文的in_reply_to_user_id是否等于Bot的ID（说明该提及是直接回复Bot的推文）
3. 仅符合条件的提及才触发回复

修改后的代码 #

import tweepy
import time
from datetime import datetime


# Connecting to Twitter API
client = tweepy.Client(bearer_token, api_key, api_secret, access_token, access_token_secret)
auth = tweepy.OAuth1UserHandler(api_key, api_secret, access_token, access_token_secret)
api = tweepy.API(auth)

# Message to reply with if someone mentions the bot
message = "Hi, Thank you so much for reaching TalkyTweety Out :) Please reply to continue this conversation. "

# Bot's unique ID
client_id = client.get_me().data.id

# This is so the bot only looks for the most recent tweets.
start_id = 1
initialisation_resp = client.get_users_mentions(client_id, tweet_fields=["in_reply_to_status_id", "in_reply_to_user_id"])

if initialisation_resp.data != None:
    start_id = initialisation_resp.data[0].id

# Looking for mentions tweets in an endless loop
while True:
    # 请求包含回复相关字段的提及数据
    response = client.get_users_mentions(client_id, since_id=start_id, tweet_fields=["in_reply_to_status_id", "in_reply_to_user_id"])
    # Reply Code
    if response.data != None:
        for tweet in response.data:
            try:
                print(tweet.text)
                # 判断该提及是否是回复Bot的推文
                if tweet.in_reply_to_user_id == client_id:
                    client.create_tweet(in_reply_to_tweet_id=tweet.id, text=message)
                    print(f"已回复推文ID: {tweet.id}")
                else:
                    print(f"推文ID {tweet.id} 不是回复Bot的推文，跳过")
                start_id = tweet.id
            except Exception as error:
                print(f"错误: {error}")

    # Delay (so the bot doesn't search for new tweets a bunch of times each second)
    time.sleep(5)

补充说明 #

• 如果需要更严格的判断（比如确认被回复的推文确实是Bot发布的），可以通过client.get_tweet(tweet.in_reply_to_status_id)获取被回复的推文，检查其作者ID是否等于Bot的ID，但这会增加API调用次数，需要注意限流。
• 确保你的Twitter开发者账号已开通v2 API的相关权限，避免因权限不足导致字段无法获取。

内容的提问来源于stack exchange，提问作者Ankit Sharma