SimRecall/filter_video.py

#coding utf-8
import sys
from utils import  filter_video_status

if __name__=="__main__":
    f = open(sys.argv[1])
    video_set = set('')
    video_rec = []
    while True:
        line = f.readline()
        if not line:
           break
        items = line.strip().split("\t")
        if len(items)<3:
            continue
        vid = -1
        try:
            vid = int(items[2])
        except:
            continue
        if vid ==-1:
            continue
        video_set.add(vid)
        video_rec.append(line)
    f.close()
    print(len(video_set))
    video_list = list(video_set)
    left_video_list = filter_video_status(video_list)
    left_video_set = set(left_video_list)
    print(left_video_list)
    f2 = open(sys.argv[2], 'w')
    for line in video_rec:
        items  = line.strip().split("\t")
        if len(items)<3:
           continue
        vid = -1
        try:
            vid  = int(items[2])
        except:
            continue
        if vid not in left_video_set:
            continue
        f2.write(line)  
    f2.close()
    #print(len(left_video_list))