Update app.py
Browse files
app.py
CHANGED
@@ -21,6 +21,7 @@ demo = gr.Interface(fn=classify_sentiment,
|
|
21 |
|
22 |
demo.launch()
|
23 |
'''
|
|
|
24 |
import gradio as gr
|
25 |
from transformers import TFBertForSequenceClassification, BertTokenizer
|
26 |
import tensorflow as tf
|
@@ -53,5 +54,87 @@ demo = gr.Interface(
|
|
53 |
)
|
54 |
|
55 |
demo.launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
56 |
|
57 |
|
|
|
21 |
|
22 |
demo.launch()
|
23 |
'''
|
24 |
+
'''
|
25 |
import gradio as gr
|
26 |
from transformers import TFBertForSequenceClassification, BertTokenizer
|
27 |
import tensorflow as tf
|
|
|
54 |
)
|
55 |
|
56 |
demo.launch()
|
57 |
+
'''
|
58 |
+
import gradio as gr
|
59 |
+
from transformers import TFBertForSequenceClassification, BertTokenizer
|
60 |
+
import tensorflow as tf
|
61 |
+
import snscrape.modules.twitter as sntwitter
|
62 |
+
import praw
|
63 |
+
import os
|
64 |
+
|
65 |
+
# Load model and tokenizer
|
66 |
+
model = TFBertForSequenceClassification.from_pretrained("shrish191/sentiment-bert")
|
67 |
+
tokenizer = BertTokenizer.from_pretrained("shrish191/sentiment-bert")
|
68 |
+
|
69 |
+
# Label Mapping
|
70 |
+
LABELS = {
|
71 |
+
0: "Neutral",
|
72 |
+
1: "Positive",
|
73 |
+
2: "Negative"
|
74 |
+
}
|
75 |
+
|
76 |
+
# Reddit API setup with environment variables
|
77 |
+
reddit = praw.Reddit(
|
78 |
+
client_id=os.getenv("REDDIT_CLIENT_ID"),
|
79 |
+
client_secret=os.getenv("REDDIT_CLIENT_SECRET"),
|
80 |
+
user_agent=os.getenv("REDDIT_USER_AGENT", "sentiment-classifier-script")
|
81 |
+
)
|
82 |
+
|
83 |
+
# Tweet text extractor
|
84 |
+
def fetch_tweet_text(tweet_url):
|
85 |
+
try:
|
86 |
+
tweet_id = tweet_url.split("/")[-1]
|
87 |
+
for tweet in sntwitter.TwitterTweetScraper(tweet_id).get_items():
|
88 |
+
return tweet.content
|
89 |
+
return "Unable to extract tweet content."
|
90 |
+
except Exception as e:
|
91 |
+
return f"Error fetching tweet: {str(e)}"
|
92 |
+
|
93 |
+
# Reddit post extractor
|
94 |
+
def fetch_reddit_text(reddit_url):
|
95 |
+
try:
|
96 |
+
submission = reddit.submission(url=reddit_url)
|
97 |
+
return f"{submission.title}\n\n{submission.selftext}"
|
98 |
+
except Exception as e:
|
99 |
+
return f"Error fetching Reddit post: {str(e)}"
|
100 |
+
|
101 |
+
# Sentiment classification logic
|
102 |
+
def classify_sentiment(text_input, tweet_url, reddit_url):
|
103 |
+
if reddit_url.strip():
|
104 |
+
text = fetch_reddit_text(reddit_url)
|
105 |
+
elif tweet_url.strip():
|
106 |
+
text = fetch_tweet_text(tweet_url)
|
107 |
+
elif text_input.strip():
|
108 |
+
text = text_input
|
109 |
+
else:
|
110 |
+
return "[!] Please enter text or a post URL."
|
111 |
+
|
112 |
+
if text.lower().startswith("error") or "Unable to extract" in text:
|
113 |
+
return f"[!] Error: {text}"
|
114 |
+
|
115 |
+
try:
|
116 |
+
inputs = tokenizer(text, return_tensors="tf", truncation=True, padding=True)
|
117 |
+
outputs = model(inputs)
|
118 |
+
probs = tf.nn.softmax(outputs.logits, axis=1)
|
119 |
+
pred_label = tf.argmax(probs, axis=1).numpy()[0]
|
120 |
+
confidence = float(tf.reduce_max(probs).numpy())
|
121 |
+
return f"Prediction: {LABELS[pred_label]} (Confidence: {confidence:.2f})"
|
122 |
+
except Exception as e:
|
123 |
+
return f"[!] Prediction error: {str(e)}"
|
124 |
+
|
125 |
+
# Gradio Interface
|
126 |
+
demo = gr.Interface(
|
127 |
+
fn=classify_sentiment,
|
128 |
+
inputs=[
|
129 |
+
gr.Textbox(label="Custom Text Input", placeholder="Type your tweet or message here..."),
|
130 |
+
gr.Textbox(label="Tweet URL", placeholder="Paste a tweet URL here (optional)"),
|
131 |
+
gr.Textbox(label="Reddit Post URL", placeholder="Paste a Reddit post URL here (optional)")
|
132 |
+
],
|
133 |
+
outputs="text",
|
134 |
+
title="Multilingual Sentiment Analysis",
|
135 |
+
description="Analyze sentiment of text, tweets, or Reddit posts. Supports multiple languages using BERT!"
|
136 |
+
)
|
137 |
+
|
138 |
+
demo.launch()
|
139 |
|
140 |
|