Tweepy-仅捕获发布了特定关键字的关注者

Question

I've been looking all over but cannot find an answer, or even a discussion, on this topic. I'm trying to sample ALL the followers of a twitter user that have tweeted something with a specific keyword. I've found a script that will allow me to pull all the followers of a specific user, but how would one edit this to sort these users by content keyword? Is this possible?

It seems that being able to sort followers by these kind of content interests would be beneficial, but I haven't seen anyone discuss it elsewhere. Thank you for any insight you can provide!

"""
http://stackoverflow.com/questions/31000178/how-to-get-large-list-of-followers-tweepy
ask user for account name to harvest follower names from.
print follower names to screen
pause  users to screen
"""
import tweepy
import time
import csv
import sys
import logging
logging.basicConfig()


accountvar = "nytimes"
#todo: upgrade this to read usernames from a file.
print "searching for followers of "+accountvar

consumer_key = "xxxxx"
consumer_secret = "xxxxx"
access_token = "xxxxx"
access_token_secret = "xxxxx"

auth = tweepy.OAuthHandler(consumer_key, consumer_secret)
auth.set_access_token(access_token, access_token_secret)
api = tweepy.API(auth, wait_on_rate_limit=True, wait_on_rate_limit_notify=True)
#refer http://docs.tweepy.org/en/v3.2.0/api.html#API
#tells tweepy.API to automatically wait for rate limits to replenish


users = tweepy.Cursor(api.followers, screen_name=accountvar).items()
count = 0
errorCount=0


outputfilecsv = accountvar+"followers.csv"
fc = csv.writer(open(outputfilecsv, 'wb'))
fc.writerow(["screen_name", "ID", "followers_count","statuses_count","location","geo_enabled"])

while True:
    try:
        user = next(users)
        count += 1
        #use count-break during dev to avoid twitter restrictions
        #if (count>10):
        #    break
    except tweepy.TweepError:
        #catches TweepError when rate limiting occurs, sleeps, then restarts.
        #nominally 15 minnutes, make a bit longer to avoid attention.
        print "sleeping...."
        time.sleep(60*16)
        user = next(users)
    except StopIteration:
        break
    try:
        print "@" + user.screen_name + " has " + str(user.followers_count) +\
              " followers, has made "+str(user.statuses_count)+" tweets and location=" +\
              user.location+" geo_enabled="+str(user.geo_enabled)+" count="+str(count)

        fc.writerow([user.screen_name, user.id_str, str(user.followers_count), str(user.statuses_count), user.location, str(user.geo_enabled)])
    except UnicodeEncodeError:
        errorCount += 1
        print "UnicodeEncodeError,errorCount ="+str(errorCount)


#apparently don't need to close csv.writer.
print "completed, errorCount ="+str(errorCount)+" total users="+str(count)
    #print "@" + user.screen_name
    #todo: write users to file, search users for interests, locations etc.

Answer 1

that's an interesting question. Let's say that you want to search for the word "banana".

Your search query will be q=banana - see https://twitter.com/search?q=banana

If you want to see if specific accounts have Tweeted that word, the query is

q=banana from:bbcnews OR from:cnn OR from:itv

See https://twitter.com/search?q=banana%20from:bbcnews%20OR%20from:cnn%20OR%20from:itv

I don't know if there's a limit to how many from: addresses you can have.