Question

上下文：我正在读取如下所示的日志文件数据：

{"action": "tweet", id: 1234, user: "user1", timestamp: 3120}
{"action": "retweet", target_id: 1234, user: "user2", timestamp: 4020}
{"action": "tweet", id: 1235, user: "user3", timestamp: 5320}
{"action": "retweet", target_id: 1235, user: "user4", timestamp: 5820}
{"action": "retweet", target_id: 1235, user: "user2", timestamp: 6540}

日志文件中有更多数据，但这些是最重要的字段。

我创建了以下函数来读取日志文件的每一行，并检测用户是否是垃圾邮件发送者。垃圾邮件发送者的标准是：

必须没有“推特”动作
必须至少有一次“转推”行动
所有转发必须在目标推文的1000毫秒内发生（显示为target_id与id的关联）。

该函数应该返回一个垃圾邮件发送者阵列，但不要重复已经被标记过至少一次的任何人。

我相信我可以清理它很多：

identifySpammers = (file_path) => {
  const spammers = [];
  const notSpammers = []; // I"m only using this to track when someone comes up as having a retweet that was after 1000 milliseconds
  let previousId;
  let previousTimeStamp;
  const data = fs.readFileSync(`/${file_path}`, 'utf8').trim().split("\n")
  for(index in data) {
    const parsedData = JSON.parse(data[index]);
    if(parsedData.timestamp - previousTimeStamp >= 1001 && spammers.includes(parsedData.user)) {
      notSpammers.push(parsedData.user)
    } else if (!spammers.includes(parsedData.user) && parsedData.target_id === previousID && parsedData.timestamp - previousTimeStamp < 1001) {
      spammers.push(parsedData.user)
    } else if (parsedData.action === "tweet") {
      previousId = parsedData.id
      previousTimeStamp = parsedData.timestamp
      notSpammers.push(parsedData.user)
    }
  }

  for(let i = 0; i < spammers.length; i++) {
    if(notSpammers.includes(spammers[i])) {
      spammers.splice(i, 1)
    }
  }

  return spammers;
}

我真正应该做的是利用类似的条件，但写入一个对象，其中每个键是用户的名字，如果值是否是垃圾邮件发送者，那么类似于：

  const spammers = {
    user1: null, // Initialize as null
    user2: false, // Once a user is marked as false, it is impossible for them to ever be considered a spammer.
    user3: true // This could end up being false later on
  }

我尝试使用与我的双阵列解决方案类似的逻辑来做到这一点，但没有运气。

基于上面的代码，如果有人可以提供任何指导，我可以提出一个更有效的对象驱动解决方案，我会很感激。不需要完整的答案，只需要一些关于如何处理的建议以及除了你可能有的任何想法。

谢谢！

编辑1：清除了一些错误并明确了用户是垃圾邮件发送者的条件。编辑2：扩展日志文件，以演示如何更早地将用户标记为垃圾邮件，但随后将必要条件传递给后者。

Answer 1

我使用更简洁的代码方式重构并使用建议的数据结构：

var data = [
 {"action": "tweet", id: 1234, user: "user1", timestamp: 1341414},
 {"action": "retweet", target_id: 1234, user: "user2", timestamp: 1341415},
 {"action": "tweet", id: 1235, user: "user3", timestamp: 2341414},
 {"action": "retweet", target_id: 1235, user: "user1", timestamp: 2341415},
]

const isSpammerByUser = {};
let previousId;
let previousTimeStamp;
for(index in data) {
  const parsedData = data[index];
  const user = parsedData.user;
  const isUserASpammer = (isSpammerByUser[user] === true);
  const hasUserProvenToNotBeASpammer = (isSpammerByUser[user] === false);
  const isFastAction = (parsedData.timestamp - previousTimeStamp < 1001);
  const isRetweetFromPrevious = (parsedData.target_id === previousId);
  
  if (parsedData.action === "tweet") {
    isSpammerByUser[user] = false;
    previousId = parsedData.id
    previousTimeStamp = parsedData.timestamp

  } else if(isUserASpammer && !isFastAction) {
    isSpammerByUser[user] = false;

  } else if (!hasUserProvenToNotBeASpammer && isRetweetFromPrevious && isFastAction) {
    isSpammerByUser[user] = true;
  }
}

const spammers = [];
for (user in isSpammerByUser)
  if (isSpammerByUser[user])
    spammers.push(user);

console.log("spammers: " + spammers.join(","));

回答：简化

1 个答案: