Invalid use of 'since_id' or 'until_id' in conjunction with 'start_time' or 'end_time'.

[highku]

Get Recent Tweets iterator throws an error when the request contains a date filter.

        try {
            let resp = await roClient.v2.search(
                text,
                {
                    'max_results': 100,
                    'tweet.fields': 'created_at,in_reply_to_user_id,geo,author_id,conversation_id,context_annotations,lang,text,source,public_metrics,withheld,possibly_sensitive',
                    'place.fields': 'name,country_code',
                    'user.fields': 'created_at,username,name,verified,public_metrics,profile_image_url',
                    'expansions': 'author_id,geo.place_id,in_reply_to_user_id',
                    'start_time': startTimeStr
                })
            for await (const tweet of resp) {
                console.log("Tweet: " + JSON.stringify(tweet))
                tweets.push(tweet)
                totalRetweets += tweet.public_metrics.retweet_count
                totalLikes += tweet.public_metrics.like_count
    
                const authorUser = resp.includes.author(tweet);
    
                if (!authorsMap.has(tweet.author_id)) {
                    authorsMap.set(tweet.author_id, { author: authorUser, tweets: 1 })
                } else {
                    authorsMap.get(tweet.author_id).tweets += 1
                }
            }
    
            tweets.map((tweet) => {
                allTweetsText = allTweetsText.concat(tweet.text)
                allTweetsText = allTweetsText.concat("\n")
            })
        } catch (error) {
            console.error("Error while fetching tweets. ")
            console.error(JSON.stringify(error))
        }

Error thrown:

{
  "type": "response",
  "code": 400,
  "error": {
    "errors": [
      {
        "parameters": {
          "until_id": [
            "1479477826428358657"
          ],
          "start_time": [
            "2022-01-06T20:27Z"
          ]
        },
        "message": "Invalid use of 'since_id' or 'until_id' in conjunction with 'start_time' or 'end_time'."
      }
    ],
    "title": "Invalid Request",
    "detail": "One or more parameters to your request was invalid.",
    "type": "https://api.twitter.com/2/problems/invalid-request"
  },
  ...
}

[alkihis]

Hi,

This is quite a strange choice from Twitter to deny use of until_id and start_time together, as they're not exclusive filters 😅.

I'll look how to paginate without use of since_id/until_id, but the docs seems to not mention clearly pagination tokens for this endpoint. next_token is mentionned in response field list, but is marked as optional.

As a mitigation, you can approximate a start_time using Twitter snowflake IDs (as they are encoded with a timestamp inside):

const TWITTER_START_EPOCH = 1288834974657n;

function dateToTwitterSnowflake(date) {
  const dateTimestamp = BigInt(date.valueOf());
  return ((dateTimestamp - TWITTER_START_EPOCH) << 22n).toString();
}

then use start_time as since_id:

const resp = await roClient.v2.search(text, { since_id: dateToTwitterSnowflake(new Date(startTimeStr)) })

[alkihis]

Fix now available in 1.9.1 :)