twitter-thoughts-server/main.py at main · Siddgh/twitter-thoughts-server · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
from fastapi import FastAPI, Response

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
import re
import twint
import nest_asyncio
import numpy as np
from fastapi.middleware.cors import CORSMiddleware
from langdetect import detect


nest_asyncio.apply()


app = FastAPI()

origins = ["*"]
app.add_middleware(
    CORSMiddleware,
    allow_origins=origins,
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

tokenizer = AutoTokenizer.from_pretrained(
    'nlptown/bert-base-multilingual-uncased-sentiment')
model = AutoModelForSequenceClassification.from_pretrained(
    'nlptown/bert-base-multilingual-uncased-sentiment')


def get_tweets_by_keyword(keyword, fetch_rate=50):
    k = twint.Config()
    k.Pandas = True
    k.Lang = 'en'
    k.Search = keyword
    k.Limit = fetch_rate

    twint.run.Search(k)
    df = twint.storage.panda.Tweets_df
    return df


def get_users_tweets(username, fetch_rate=50):
    u = twint.Config()
    u.Pandas = True
    u.Lang = 'en'
    u.Search = 'from:@'+str(username)
    u.Limit = fetch_rate

    twint.run.Search(u)
    df = twint.storage.panda.Tweets_df
    return df


def clean_tweets(tweet):
    tweet = tweet.lower()
    tweet = re.sub("@[A-Za-z0-9_]+", "", tweet)  # Remove Mentions
    tweet = re.sub("#[A-Za-z0-9_]+", "", tweet)  # Remove Hashtags
    tweet = re.sub(r"http\S+", "", tweet)  # Remove Hyper Links
    tweet = re.sub(r"www.\S+", "", tweet)  # Remove Hyper Links
    tweet = re.sub('[()!?]', '', tweet)  # Remove Punctuations
    tweet = re.sub('\[.*?\]', '', tweet)  # Remove Punctuations
    tweet = re.sub("[^a-z0-9]", " ", tweet)  # Remove Non-Alphanumeric Values
    tweet = tweet.lstrip().rstrip()
    tweet = ' '.join(tweet.split())
    return tweet


def get_sentiment_scores(tweet):
    tokens = tokenizer.encode(tweet, return_tensors='pt')
    result = model(tokens)
    return int(torch.argmax(result.logits)) + 1


def detect_lang(tweet):
    try:
        val = detect(tweet)
    except:
        val = 'nan'
    return val


@app.get('/keyword/{keyword}/{fetchrate}')
def trend_sentiments(keyword, fetchrate):
    keyword_tweets_df = get_tweets_by_keyword(keyword, fetchrate)
    if len(keyword_tweets_df) > 0:
        keyword_tweets_df['cleaned_tweets'] = keyword_tweets_df['tweet'].apply(
            clean_tweets)
        keyword_tweets_df['cleaned_tweets'] = keyword_tweets_df['cleaned_tweets'].replace(
            '', np.nan)
        keyword_tweets_df = keyword_tweets_df.dropna(
            axis=0, subset=['cleaned_tweets'])

        keyword_tweets_df['lang'] = keyword_tweets_df['cleaned_tweets'].apply(
            detect_lang)

        keyword_input_df = keyword_tweets_df.loc[keyword_tweets_df['lang'] == 'en']
        keyword_input_df['sentiment_scores'] = keyword_input_df['cleaned_tweets'].apply(
            get_sentiment_scores)

        res_json = keyword_input_df[[
            'cleaned_tweets', 'sentiment_scores']].to_json()

        print(len(keyword_tweets_df), len(keyword_input_df))
        return Response(content=res_json, media_type="application/json")
    else:
        return f'nan'


@app.get('/user/{keyword}/{fetchrate}')
def trend_sentiments(keyword, fetchrate):
    keyword_tweets_df = get_users_tweets(keyword, fetchrate)
    if len(keyword_tweets_df) > 0:
        keyword_tweets_df['cleaned_tweets'] = keyword_tweets_df['tweet'].apply(
            clean_tweets)
        keyword_tweets_df['cleaned_tweets'] = keyword_tweets_df['cleaned_tweets'].replace(
            '', np.nan)
        keyword_tweets_df = keyword_tweets_df.dropna(
            axis=0, subset=['cleaned_tweets'])

        keyword_tweets_df['lang'] = keyword_tweets_df['cleaned_tweets'].apply(
            detect_lang)

        keyword_input_df = keyword_tweets_df.loc[keyword_tweets_df['lang'] == 'en']
        keyword_input_df['sentiment_scores'] = keyword_input_df['cleaned_tweets'].apply(
            get_sentiment_scores)

        res_json = keyword_input_df[[
            'cleaned_tweets', 'sentiment_scores']].to_json()
        return Response(content=res_json, media_type="application/json")
    else:
        return f'nan'