Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
0 parents
commit 30d9425
Showing
1 changed file
with
78 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,78 @@ | ||
# -*- coding: utf-8 -*- | ||
import tweepy, re, time, json, os | ||
from keys import * #import ACCESS_TOKENs and API_KEYs | ||
from collections import Counter | ||
|
||
keywords = ['BlackLivesMatter','#BlackLivesMatter'] | ||
counter_dict = Counter() | ||
# In case of some error(like no network connection ๐ ) Counter dict can be set from last line on data.txt | ||
#counter_dict = Counter({"โ๐ฟ": 813, "๐คข": 33, "๐๐ฟ": 24, "๐คท๐ฝโโ๏ธ": 6, "โ": 5, "โผ": 307, "๐ข": 13, "๐คฏ": 23, "๐ณ": 38, "๐": 14, "๐น": 26, "๐": 31, "๐ค": 30, "๐": 229, "๐": 24, "๐": 169, "โค": 1063, "โ": 203, "โ๐ป": 292, "โ๐ผ": 399, "โ๐ฝ": 812, "โ๐พ": 989, "๐ฉ": 46, "๐คธ๐ฟโโ๏ธ": 1, "๐": 3, "๐ค": 852, "๐ฌ": 17, "๐ช๐ฝ": 20, "๐คฆ๐ฝโโ๏ธ": 15, "๐": 6, "๐ณ๐ฑ": 4, "๐": 5, "๐ฅบ": 243, "๐๐ฝ": 3, "๐": 85, "๐": 16, "๐จ": 15, "๐จ": 5, "๐": 230, "๐งก": 53, "๐": 80, "๐": 117, "๐": 92, "๐": 74, "๐": 47, "โผ": 4, "โก": 21, "๐": 4, "๐ข": 124, "๐ต๐ธ": 4, "๐ฏ": 129, "๐": 46, "๐": 25, "๐": 10, "๐ผ๐ฝ": 3, "๐๐ฝ": 179, "๐คท๐ฝโโ๏ธ": 12, "๐บ๐ธ": 105, "๐ฆ ": 5, "๐๐ผ": 23, "๐ช๐ป": 4, "๐ ": 41, "๐ก": 8, "๐๐ฟ": 15, "๐ช๐ฟ": 14, "๐๐ฟ": 34, "๐ธ๐ฟ": 2, "๐๐ป": 49, "๐": 17, "๐": 90, "๐ฅ": 35, "๐": 108, "๐ค๐ฝ": 1, "๐๐ฝ": 20, "๐": 406, "๐": 24, "๐ค": 151, "๐คฌ": 146, "๐ท": 13, "โจ": 119, "๐ง": 19, "๐ธ": 19, "๐": 18, "๐๐พ": 37, "๐ฆ ": 6, "๐": 73, "๐": 114, "๐ฆ": 13, "๐ธ": 17, "๐คก": 76, "๐ ๐ฝโโ๏ธ": 1, "๐คง": 13, "๐": 184, "๐ญ": 522, "๐ฌ": 1, "๐ฅ": 192, "๐": 2, "โฅ": 123, "๐": 191, "๐": 3, "๐๐พ": 118, "๐ ": 41, "๐๐ผ": 23, "โ": 72, "๐": 4, "๐ต": 9, "๐ถ": 24, "๐": 4, "๐๐พ": 87, "๐๐ฝ": 73, "๐ข": 2, "๐ถ๐พโโ๏ธ": 4, "๐ฅ": 10, "๐": 4, "๐คฒ๐ฝ": 6, "๐คฒ๐พ": 3, "๐คฒ๐ฟ": 5, "๐ค": 21, "๐ฝ": 4, "๐พ": 4, "๐ฟ": 5, "๐คฆ๐พโโ๏ธ": 18, "๐๐ผ": 3, "๐๐ฝ": 5, "๐๐พ": 5, "๐๐ฟ": 4, "โ๐ป": 19, "๐ค๐ป": 6, "๐ค๐ฟ": 11, "๐๐ฟ": 5, "๐คด๐ฟ": 1, "๐คด๐พ": 8, "๐ณ๏ธโ๐": 44, "๐ค": 29, "๐ฏ": 7, "๐ฅฐ": 91, "๐ฎ": 23, "โ": 70, "๐โโ๏ธ": 8, "๐น": 11, "๐คฒ๐ป": 20, "๐๐ผ": 46, "๐": 55, "๐": 20, "๐": 24, "๐น": 26, "๐ค๐ป": 2, "๐ค๐ป": 5, "๐": 17, "๐คฆโโ๏ธ": 20, "๐": 81, "๐": 26, "๐ช๐พ": 23, "๐๐ผ": 72, "๐ค๐พ": 8, "๐": 9, "๐คณ๐พ": 1, "๐": 9, "๐ช๐ผ": 10, "๐๐ฝ": 2, "๐ท": 27, "๐ฎ": 4, "๐๐ป": 49, "๐ณ": 4, "๐": 26, "๐ค": 1, "๐": 2, "๐คญ": 13, "๐คท": 3, "๐": 37, "๐": 2, "๐คทโโ๏ธ": 11, "โข": 12, "โ": 5, "๐ฅ": 1, "๐๐ป": 10, "๐๐ฝ": 26, "๐บ": 5, "๐": 3, "๐คฆ๐ฝโโ๏ธ": 13, "๐": 13, "โบ": 14, "๐": 51, "๐": 16, "๐": 21, "โฎ": 15, "๐": 2, "๐๐ป": 13, "๐": 4, "๐ผ": 3, "๐ฅ": 1, "๐จ": 109, "โฃ": 23, "๐": 27, "๐ค๐ฝ": 5, "๐๐ป": 16, "๐น๐ด": 1, "๐ช": 22, "๐ค": 3, "๐ต๐ท": 26, "๐": 41, "๐๐พ": 11, "๐คฆ๐ปโโ๏ธ": 9, "๐คท๐ปโโ๏ธ": 16, "๐": 1, "๐คฎ": 43, "โซ": 10, "๐ช": 32, "๐ฃ": 123, "๐": 6, "๐ซ": 12, "๐": 49, "๐ฃ": 11, "๐": 2, "๐": 15, "๐๐ฝโโ๏ธ": 1, "๐": 8, "๐ฐ": 7, "๐ต": 10, "โ": 13, "๐๐ป": 14, "๐ค": 3, "๐": 13, "๐ค": 3, "โ ": 23, "๐ธ": 2, "๐จ": 4, "๐": 10, "๐๐ฝ": 1, "๐๐พ": 1, "๐๐ฝ": 20, "๐จ": 5, "๐ฆ": 3, "๐": 21, "๐": 15, "๐": 14, "๐": 13, "โก": 21, "๐ฌ๐ญ": 8, "๐": 2, "๐ฟ๐ผ": 3, "โช": 14, "๐ฎโโ๏ธ": 2, "๐": 9, "๐": 28, "๐ต๐ญ": 9, "๐": 3, "๐ก": 92, "๐๐ฝ": 10, "๐": 7, "๐คฉ": 14, "๐ฃ": 17, "๐ซ": 17, "๐ง๐ฝ": 2, "๐ก": 5, "๐ณ๐ฟ": 17, "๐ด๓ ง๓ ข๓ ณ๓ ฃ๓ ด๓ ฟ": 5, "๐ค": 3, "๐๐ผโโ๏ธ": 1, "๐ง ": 3, "๐ป": 2, "๐คท๐พโโ๏ธ": 18, "๐คฃ": 168, "๐ค๐ป": 1, "๐๐ผ": 15, "๐๐พ": 24, "๐๐ฟ": 30, "โ๐ผ": 14, "๐ญ๐ฐ": 2, "๐คท๐ฟโโ๏ธ": 3, "๐": 10, "๐ฆ๐บ": 3, "๐": 1, "๐ท": 2, "๐ด": 19, "๐ต": 1, "๐ฅ": 1, "๐ท": 7, "๐": 12, "๐ซ": 4, "๐ฝ": 7, "๐งต": 4, "๐": 15, "๐ธ๐พ": 1, "๐": 16, "๐": 5, "๐ฝ": 15, "ยฉ": 7, "๐คจ": 18, "๐คฆ๐ปโโ๏ธ": 13, "๐๐ฟโโ๏ธ": 3, "๐งข": 2, "๐ฏ": 3, "๐ฌ๐ง": 12, "๐จ๐ณ": 2, "๐ฐ": 5, "โฌ": 3, "๐ฟ": 1, "โฌ": 55, "๐ฑ": 20, "๐ช๐ธ": 2, "๐ฅ": 24, "๐ฟ๐ฆ": 5, "๐๐ฝโโ๏ธ": 2, "๐ท": 4, "๐ฅด": 25, "๐": 1, "๐ค": 1, "๐ฒ": 10, "โฌ": 6, "๐ธ๐ฆ": 1, "โณ": 1, "๐ถ": 6, "๐ง": 5, "๐ฉ": 15, "๐คฆโโ๏ธ": 19, "๐งโโ๏ธ": 3, "๐๐ผ": 6, "๐๐ผ": 10, "๐๐ผ": 4, "๐ค": 9, "๐ฌ": 1, "๐พ": 6, "๐๐ป": 20, "๐ฐ๐ช": 2, "๐": 41, "๐ฎ๐ณ": 3, "๐คท๐ปโโ๏ธ": 12, "๐": 23, "๐๐ผ": 1, "๐ค": 2, "๐ค": 1, "๐ต": 4, "๐": 5, "๐งฟ": 4, "โ ": 42, "ยฎ": 2, "๐๐ผ": 3, "๐คฆ๐ฝ": 1, "๐": 19, "๐๐พ": 27, "๐๐ฟ": 22, "๐ฅต": 15, "๐๐ฝ": 15, "๐๐ผ": 1, "๐": 6, "๐": 5, "๐": 7, "๐คช": 21, "๐ซ": 21, "๐ฆ": 1, "๐๐ป": 6, "๐": 4, "๐": 8, "โ": 1, "โ": 5, "๐": 36, "๐ ๐ฝ": 2, "โ": 6, "๐ฝ": 1, "๐": 15, "โน": 3, "๐ด": 14, "๐๐ฝ": 8, "๐ช": 2, "๐": 1, "๐ง": 1, "๐จ๐ฆ": 12, "๐บ": 6, "โ๐ฝ": 18, "๐ค๐พ": 4, "๐ค๐พ": 3, "๐ถ": 2, "๐ฅณ": 6, "๐ฉ": 2, "๐ฉโ๐ป": 2, "๐จ๐ต": 1, "๐": 2, "๐": 8, "๐คฆ๐พโโ๏ธ": 10, "๐": 6, "๐ฆ": 3, "๐ฒ๐ฝ": 7, "๐ด": 8, "๐": 5, "๐๐ป": 1, "โ๐ฝ": 3, "๐ฐ๐ท": 1, "๐ฏ๐ต": 1, "๐ซ๐ท": 6, "๐ฉ๐ช": 9, "๐ฎ๐น": 5, "๐คซ": 4, "๐คฆ๐ผโโ๏ธ": 5, "๐งผ": 1, "๐๐ฝ": 7, "๐๐ป": 5, "๐": 13, "๐": 5, "๐คฆ": 5, "โ": 1, "๐ค๐พ": 3, "๐": 17, "๐ค๐ผ": 2, "๐ฆ": 16, "๐ป": 3, "โ": 3, "๐ฏ": 4, "๐๐ป": 2, "โ": 6, "๐๐ป": 1, "๐คด๐ฝ": 2, "๐๐พ": 9, "๐": 3, "๐ฎ": 1, "โคต": 3, "๐๐ฝโโ๏ธ": 3, "๐๐พ": 14, "๐ง๐ฝโโ๏ธ": 1, "๐": 2, "โ": 2, "๐": 1, "๐ฒ๐ป": 3, "๐ฎ": 2, "๐": 4, "๐ต": 2, "๐คท๐พโโ๏ธ": 9, "๐": 9, "๐ต๐ช": 3, "๐": 3, "๐": 1, "๐": 4, "๐": 2, "๐ง๐พโโ๏ธ": 1, "๐": 21, "๐ณ": 8, "โ๐พ": 12, "โ๐ฟ": 10, "โช": 2, "๐ฟ": 1, "๐บ๐ฌ": 1, "๐ง": 2, "๐": 3, "๐ค๐ฝ": 3, "๐ฑ๐ปโโ๏ธ": 2, "๐": 1, "๐โโ๏ธ": 4, "๐ธ": 2, "๐": 7, "๐ฑ": 5, "๐ฉ๐ฟ": 2, "๐คฆ๐ผโโ๏ธ": 3, "๐ฆข": 1, "๐": 1, "๐บ": 1, "๐คฆ๐ฟโโ๏ธ": 2, "๐ฏ": 5, "โ๐ฝ": 3, "๐ค๐ฝ": 3, "๐ค๐ฝ": 1, "๐งโโ๏ธ": 4, "๐ค๐ฟ": 3, "๐ฌ": 6, "โ": 5, "๐น": 18, "๐ป": 1, "๐ฆ": 5, "๐": 2, "๐": 2, "๐ฆ": 1, "๐": 6, "๐": 1, "๐บ๐ฒ": 13, "๐ธ": 5, "โถ": 5, "๐ฒ": 4, "๐": 2, "๐ฝ": 4, "๐ ๐ผ": 3, "๐คฅ": 3, "๐ญ": 3, "โฌ": 6, "๐๐ผ": 3, "โ": 12, "๐๐พ": 3, "๐ฅ": 1, "โ": 10, "๐ฅ": 2, "๐": 1, "๐ถ": 1, "๐ด": 1, "๐ค": 7, "๐ฏโโ๏ธ": 1, "๐ธ๐ฎ": 3, "๐น๐ณ": 1, "๐ฃ": 4, "๐ธ": 1, "๐": 6, "๐ฆ": 1, "๐": 5, "๐ผ": 1, "โ": 2, "๐คฐ๐ฝ": 1, "โ": 20, "๐": 5, "๐ค ": 5, "๐คฒ": 5, "๐": 1, "๐ฒ๐ญ": 1, "๐คทโโ๏ธ": 9, "๐ท": 4, "๐๐ปโโ๏ธ": 1, "๐ค๐ผ": 3, "โ": 5, "๐ฎ๐ช": 2, "๐ ๐ฟโโ๏ธ": 1, "๐ ๐ฟโโ๏ธ": 1, "๐ค๐ฝ": 5, "๐ด๓ ง๓ ข๓ ฅ๓ ฎ๓ ง๓ ฟ": 3, "โ": 2, "โ๐ฝ": 1, "๐๐ป": 1, "๐คค": 4, "๐": 5, "โญ": 1, "๐": 6, "๐ฅ": 3, "๐ฒ๐ฒ": 1, "โ": 1, "๐ป": 4, "โ": 2, "๐": 1, "โ": 4, "๐": 1, "๐": 1, "โ": 1, "๐ค": 8, "โ": 4, "๐๐ป": 1, "๐ฆ๐ฟ": 1, "๐": 2, "๐ฉ": 6, "1๏ธโฃ": 2, "2๏ธโฃ": 2, "3๏ธโฃ": 1, "4๏ธโฃ": 1, "5๏ธโฃ": 1, "๐ก": 1, "๐๐ปโโ๏ธ": 1, "๐งฉ": 1, "๐": 4, "๐ฆ๐ฟ": 1, "๐ฆ๐ป": 1, "๐ค": 3, "โป": 2, "๐ฌ": 2, "๐ญ": 1, "๐ด๐ป": 3, "๐ธ๐ป": 1, "๐": 3, "๐จ๐ฒ": 1, "๐ฐ": 3, "โ๐ป": 3, "๐ง๐ปโโ๏ธ": 2, "๐": 5, "๐คฐ๐พ": 3, "๐": 1, "๐๐พ": 1, "๐ถ": 1, "๐ป": 2, "๐ฆ๐ธ": 1, "๐ด": 2, "๐ซ": 1, "๐ง๐ผโโ๏ธ": 1, "๐๐ป": 2, "โ ": 8, "๐": 9, "๐ง๐พ": 1, "๐ฎ๐ท": 1, "๐๐ฟ": 23, "๐๐ฟ": 10, "๐๐พ": 10, "๐๐พ": 8, "๐๐ฝ": 4, "๐๐ฝ": 4, "๐๐ผ": 3, "๐๐ผ": 2, "๐๐ป": 1, "๐๐ผ": 3, "๐๐ฟ": 9, "โพ": 1, "๐ธ": 2, "๐ต": 1, "๐": 1, "๐ธ๐ผ": 1, "๐": 1, "๐ฅ": 1, "๐พ": 1, "โป": 1, "๐โโ๏ธ": 1, "๐๐ผโโ๏ธ": 1, "๐": 3, "๐ฝ": 1, "๐คท๐ผโโ๏ธ": 1, "๐คท๐ฟโโ๏ธ": 2, "๐คท๐ผโโ๏ธ": 6, "๐ธ๐ฝ": 1, "๐๐พโโ๏ธ": 2, "๐ค๐ฟ": 5, "๐": 1, "๐ซ": 1, "๐ณ๐ฌ": 2, "โซ": 2, "๐ฎ๐ผ": 1, "๐ท๐บ": 1, "๐บ": 3, "โฌ ": 1, "๐ฟ": 2, "๐ฎ": 1, "๐": 1, "๐ท": 5, "๐ต๐ฆ": 1, "โ": 1, "โ": 9, "๐ค": 2, "๐": 1, "๐": 1, "โญ": 3, "๐": 5, "๐ค๐ป": 1, "๐": 2, "๐ฉ๐ด": 5, "๐จ๐ฟโ๐ฆฒ": 1, "๐ง๐ง": 1, "๐ฉ": 2, "๐๐พ": 4, "๐น": 1, "โ๐พ": 1, "๐ฆ": 1, "๐ ฟ": 3, "๐": 4, "๐ง๐ฏ": 1, "๐": 4, "๐โโ๏ธ": 2, "๐ต๐น": 1, "๐งฅ": 1, "๐ฉ๐ฟโ๐ฆฑ": 1, "โ": 1, "โ": 2, "๐ฆ": 8, "๐๐พ": 2, "๐น": 7, "๐": 1, "๐ญ": 2, "โ ": 1, "โ": 24, "๐น": 2, "๐๐พโโ๏ธ": 2, "๐": 1, "๐น": 1, "๐": 1, "๐ฑ": 3, "๐๐ปโโ๏ธ": 1, "๐ญ๐น": 1, "๐ ": 1, "๐": 1, "๐ท๐ฝโโ๏ธ": 1, "๐งฑ": 1, "๐ค๐ป": 2, "โ": 18, "๐ฌ": 3, "๐": 3, "๐": 1, "๐ง": 1, "๐๐ฟ": 1, "๐คด": 1, "๐ง๐ฟ": 1, "๐ฆ": 3, "๐": 1, "๐ ": 1, "๐ค๐ฟ": 2, "๐": 3, "๐": 11, "๐ฌ๐พ": 4, "๐": 2, "โฉ": 3, "๐": 10, "๐ฒ๐จ": 1, "๐บ": 2, "๐บ": 1, "๐ง": 3, "๐": 1, "๐๐ปโโ๏ธ": 1, "๐๐ผโโ๏ธ": 1, "๐๐ฝโโ๏ธ": 1, "๐๐พโโ๏ธ": 1, "๐๐ฟโโ๏ธ": 1, "๐จ๐ฉ": 3, "๐คด๐ผ": 1, "๐บ๐ฟ": 1, "๐ฎโโ๏ธ": 1, "๐": 1, "๐": 1, "โ": 1, "๐งป": 1, "๐": 1, "๐บ": 1, "โฝ": 2, "๐ฃ": 1, "๐ง๐ป": 2, "๐ผ": 1, "๐ซ": 3, "๐ง": 1, "๐๐ฟ": 1, "๐ค": 1, "๐ถ๐พ": 1, "๐ง๐พ": 1, "๐ฆฑ": 1, "๐จ๐ฝโ๐ฆฑ": 1, "๐จ๐พ": 2, "๐จ๐พโ๐ฆฒ": 1, "๐ด๐พ": 1, "๐ค๐ผ": 2, "๐ธ": 2, "๐๐พโโ๏ธ": 4, "๐ข": 1, "๐": 2, "๐ฑ๐ท": 1, "๐ฆธ๐ผโโ๏ธ": 2, "๐ฆ": 2, "๐น๐ท": 1, "๐ต๐ฐ": 1, "๐ฟ": 1, "๐ค๐ฟ": 2, "๐ซ": 1, "๐ ๐พโโ๏ธ": 1, "โฝ": 3, "๐คณ": 1, "๐ธ๐พ": 1, "๐": 1, "๐": 1, "๐ฉ๐ป": 1, "๐ป": 1, "โ๐ผ": 1, "โ๐พ": 1, "โ๐ฟ": 1, "๐ค๐พ": 1, "๐": 1, "๐": 3, "๐ฅฆ": 1, "๐งฐ": 1, "๐ ๐พโโ๏ธ": 1, "โฆ": 1, "๐": 2, "๐๐พโโ๏ธ": 1, "๐ป": 1, "๐ซ๐ฎ": 5, "๐ฟ": 1, "๐ฅ": 1, "๐โโ๏ธ": 2, "๐ค": 1, "๐": 2, "โ": 10, "๐ ": 1, "โธ": 1, "โ": 2, "๐ค๐ผ": 1, "๐๐ปโโ๏ธ": 1, "โ๐พ": 1, "๐ค๐ผ": 1, "๐๐ปโโ๏ธ": 1, "๐": 1, "๐พ": 1, "โฎ": 1, "๐ฎ": 1, "๐๐ผโโ๏ธ": 1, "๐๐พ": 1, "๐๐ผโโ๏ธ": 1, "๐": 1, "๐ณ": 1, "๐ฑ": 1, "โฃ": 1, "๐งโโ๏ธ": 1, "๐๐ฟ": 1, "๐ง๐ผโโ๏ธ": 2}) | ||
|
||
dirname = os.path.dirname(os.path.abspath(__file__)) | ||
datatxt = os.path.join(dirname, 'data.txt') | ||
|
||
# Create a file data.txt file - remove/comment if restarting the script in case of some errors with stream connection | ||
with open(datatxt, mode='a', encoding='utf-8') as f: | ||
json.dump([], f) | ||
|
||
emoji_regex = re.compile('[#*0-9]๏ธโฃ|[ยฉยฎโผโโขโนโ-โโฉโชโโโจโโฉ-โณโธ-โบโโชโซโถโโป-โพโ-โโโโโโ]|โ[๐ป-๐ฟ]?|[โ โขโฃโฆโชโฎโฏโธ-โบโโโ-โโโ โฃโฅโฆโจโปโพโฟโ-โโโโโ โกโชโซโฐโฑโฝโพโโ โโโโโโโฉโชโฐ-โตโทโธ]|โน(?:๏ธโ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[โบโฝโโ โโ]|[โ-โ][๐ป-๐ฟ]?|[โโโโโโกโจโณโดโโโโโ-โโโฃโคโ-โโกโฐโฟโคดโคตโฌ -โฌโฌโฌโญโญใฐใฝใใ๐๐๐ ฐ๐ ฑ๐ พ๐ ฟ๐๐-๐]|๐ฆ[๐จ-๐ฌ๐ฎ๐ฑ๐ฒ๐ด๐ถ-๐บ๐ผ๐ฝ๐ฟ]|๐ง[๐ฆ๐ง๐ฉ-๐ฏ๐ฑ-๐ด๐ถ-๐น๐ป๐ผ๐พ๐ฟ]|๐จ[๐ฆ๐จ๐ฉ๐ซ-๐ฎ๐ฐ-๐ต๐ท๐บ-๐ฟ]|๐ฉ[๐ช๐ฌ๐ฏ๐ฐ๐ฒ๐ด๐ฟ]|๐ช[๐ฆ๐จ๐ช๐ฌ๐ญ๐ท-๐บ]|๐ซ[๐ฎ-๐ฐ๐ฒ๐ด๐ท]|๐ฌ[๐ฆ๐ง๐ฉ-๐ฎ๐ฑ-๐ณ๐ต-๐บ๐ผ๐พ]|๐ญ[๐ฐ๐ฒ๐ณ๐ท๐น๐บ]|๐ฎ[๐จ-๐ช๐ฑ-๐ด๐ถ-๐น]|๐ฏ[๐ช๐ฒ๐ด๐ต]|๐ฐ[๐ช๐ฌ-๐ฎ๐ฒ๐ณ๐ต๐ท๐ผ๐พ๐ฟ]|๐ฑ[๐ฆ-๐จ๐ฎ๐ฐ๐ท-๐ป๐พ]|๐ฒ[๐ฆ๐จ-๐ญ๐ฐ-๐ฟ]|๐ณ[๐ฆ๐จ๐ช-๐ฌ๐ฎ๐ฑ๐ด๐ต๐ท๐บ๐ฟ]|๐ด๐ฒ|๐ต[๐ฆ๐ช-๐ญ๐ฐ-๐ณ๐ท-๐น๐ผ๐พ]|๐ถ๐ฆ|๐ท[๐ช๐ด๐ธ๐บ๐ผ]|๐ธ[๐ฆ-๐ช๐ฌ-๐ด๐ท-๐น๐ป๐ฝ-๐ฟ]|๐น[๐ฆ๐จ๐ฉ๐ซ-๐ญ๐ฏ-๐ด๐ท๐น๐ป๐ผ๐ฟ]|๐บ[๐ฆ๐ฌ๐ฒ๐ณ๐ธ๐พ๐ฟ]|๐ป[๐ฆ๐จ๐ช๐ฌ๐ฎ๐ณ๐บ]|๐ผ[๐ซ๐ธ]|๐ฝ๐ฐ|๐พ[๐ช๐น]|๐ฟ[๐ฆ๐ฒ๐ผ]|[๐๐๐๐ฏ๐ฒ-๐บ๐๐๐-๐ก๐ค-๐]|๐ [๐ป-๐ฟ]?|[๐-๐๐๐๐-๐๐-๐]|๐[๐ป-๐ฟ]?|[๐๐](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐ ๐]|๐[๐ป-๐ฟ]?|[๐๐]|๐(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐๐](?:๏ธโ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐-๐ฐ]|๐ณ(?:๏ธโ๐)?|๐ด(?:โโ ๏ธ|๓ ง๓ ข(?:๓ ฅ๓ ฎ๓ ง|๓ ณ๓ ฃ๓ ด|๓ ท๓ ฌ๓ ณ)๓ ฟ)?|[๐ต๐ท-๐]|๐(?:๏ธโ๐จ๏ธ)?|[๐๐][๐ป-๐ฟ]?|[๐๐ ]|[๐-๐][๐ป-๐ฟ]?|[๐-๐ฅ]|[๐ฆ๐ง][๐ป-๐ฟ]?|๐จ(?:โ(?:[โโโ]๏ธ|โค๏ธโ(?:๐โ)?๐จ|[๐พ๐ณ๐๐ค๐จ๐ซ๐ญ]|๐ฆ(?:โ๐ฆ)?|๐ง(?:โ[๐ฆ๐ง])?|[๐จ๐ฉ]โ(?:๐ฆ(?:โ๐ฆ)?|๐ง(?:โ[๐ฆ๐ง])?)|[๐ป๐ผ๐ง๐ฌ๐๐๐ฆฐ-๐ฆณ])|[๐ป-๐ฟ](?:โ(?:[โโโ]๏ธ|[๐พ๐ณ๐๐ค๐จ๐ซ๐ญ๐ป๐ผ๐ง๐ฌ๐๐๐ฆฐ-๐ฆณ]))?)?|๐ฉ(?:โ(?:[โโโ]๏ธ|โค๏ธโ(?:๐โ)?[๐จ๐ฉ]|[๐พ๐ณ๐๐ค๐จ๐ซ๐ญ]|๐ฆ(?:โ๐ฆ)?|๐ง(?:โ[๐ฆ๐ง])?|๐ฉโ(?:๐ฆ(?:โ๐ฆ)?|๐ง(?:โ[๐ฆ๐ง])?)|[๐ป๐ผ๐ง๐ฌ๐๐๐ฆฐ-๐ฆณ])|[๐ป-๐ฟ](?:โ(?:[โโโ]๏ธ|[๐พ๐ณ๐๐ค๐จ๐ซ๐ญ๐ป๐ผ๐ง๐ฌ๐๐๐ฆฐ-๐ฆณ]))?)?|[๐ช-๐ญ]|๐ฎ(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|๐ฏ(?:โ[โโ]๏ธ)?|๐ฐ[๐ป-๐ฟ]?|๐ฑ(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|๐ฒ[๐ป-๐ฟ]?|๐ณ(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐ด-๐ถ][๐ป-๐ฟ]?|๐ท(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|๐ธ[๐ป-๐ฟ]?|[๐น-๐ป]|๐ผ[๐ป-๐ฟ]?|[๐ฝ-๐]|[๐๐](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|๐[๐ป-๐ฟ]?|๐|๐ [๐ป-๐ฟ]?|[๐๐](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐-๐ฉ]|๐ช[๐ป-๐ฟ]?|[๐ซ-๐ฝ๐ฟ-๐ฝ๐-๐๐-๐ง๐ฏ๐ฐ๐ณ]|๐ด[๐ป-๐ฟ]?|๐ต(?:๏ธโ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐ถ-๐น]|๐บ[๐ป-๐ฟ]?|[๐๐-๐]|[๐๐๐][๐ป-๐ฟ]?|[๐ค๐ฅ๐จ๐ฑ๐ฒ๐ผ๐-๐๐-๐๐-๐๐ก๐ฃ๐จ๐ฏ๐ณ๐บ-๐]|[๐ -๐](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐-๐]|๐(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|๐[๐ป-๐ฟ]?|[๐๐](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|๐[๐ป-๐ฟ]?|[๐-๐ข]|๐ฃ(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐ค-๐ณ]|[๐ด-๐ถ](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐ท-๐ฟ]|๐[๐ป-๐ฟ]?|[๐-๐ ๐]|๐[๐ป-๐ฟ]?|[๐-๐๐ -๐ฅ๐ฉ๐ซ๐ฌ๐ฐ๐ณ-๐น๐ค-๐ค]|[๐ค-๐ค][๐ป-๐ฟ]?|๐ค|[๐ค๐ค][๐ป-๐ฟ]?|[๐ค -๐คฅ]|๐คฆ(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐คง-๐คฏ]|[๐คฐ-๐คถ][๐ป-๐ฟ]?|๐คท(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐คธ๐คน](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|๐คบ|๐คผ(?:โ[โโ]๏ธ)?|[๐คฝ๐คพ](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐ฅ-๐ฅ ๐ฅ-๐ฅฐ๐ฅณ-๐ฅถ๐ฅบ๐ฅผ-๐ฆข๐ฆฐ-๐ฆด]|[๐ฆต๐ฆถ][๐ป-๐ฟ]?|๐ฆท|[๐ฆธ๐ฆน](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐ง-๐ง๐ง]|[๐ง-๐ง][๐ป-๐ฟ]?|๐ง(?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐ง-๐ง](?:โ[โโ]๏ธ|[๐ป-๐ฟ](?:โ[โโ]๏ธ)?)?|[๐ง๐ง](?:โ[โโ]๏ธ)?|[๐ง -๐งฟ]') | ||
|
||
#Create a Stream listener | ||
class EmojiStreamListener(tweepy.StreamListener): | ||
|
||
def __init__(self): | ||
# inherit class attributes | ||
super(EmojiStreamListener, self).__init__() | ||
# add tweets counter and time logger variables | ||
self.started = time.time() | ||
self.tweet_counter = 42586 | ||
|
||
def on_status(self, status): | ||
|
||
if hasattr(status, "retweeted_status"): | ||
pass | ||
# avoid retweets without comments | ||
else: | ||
# check if text has been truncated | ||
if hasattr(status,"extended_tweet"): | ||
text = status.extended_tweet["full_text"] | ||
else: | ||
text = status.text | ||
emoji_array = emoji_regex.findall(text) | ||
# print(emoji_array) | ||
self.tweet_counter += 1 | ||
# print(self.tweet_counter) | ||
counter_dict.update(emoji_array) | ||
elapsed = time.time() - self.started | ||
#write counter state to file every minute | ||
if elapsed > 60 : | ||
self.started = time.time() | ||
with open(datatxt, mode='a') as fd: | ||
line = '#:'+ str(self.tweet_counter) + '|time:'+ str(self.started) + '|dict:' + json.dumps(counter_dict, ensure_ascii=False) + '\n' | ||
fd.write(line) | ||
|
||
def on_error(self, status_code): | ||
print(status_code) | ||
if status_code == 420: | ||
print("420 error: sleeping for 15 min...") | ||
time.sleep(15 * 60) | ||
return True | ||
|
||
#create a Stream | ||
class EmojiStream(): | ||
|
||
def __init__(self, auth, listener): | ||
self.stream = tweepy.Stream(auth=auth, listener=listener) | ||
|
||
def start(self, keywords): | ||
self.stream.filter(track=keywords, languages=['en']) | ||
|
||
#start the Stream | ||
if __name__ == "__main__": | ||
listener = EmojiStreamListener() | ||
|
||
auth = tweepy.OAuthHandler(API_KEY, API_SECRET_KEY) | ||
auth.set_access_token(ACCESS_TOKEN, ACCESS_TOKEN_SECRET) | ||
|
||
stream = EmojiStream(auth, listener) | ||
stream.start(keywords) |