Update app.py
Browse files
app.py
CHANGED
@@ -1,11 +1,9 @@
|
|
1 |
import streamlit as st
|
2 |
-
import
|
3 |
-
|
4 |
-
from transformers import pipeline
|
5 |
-
from transformers import AutoTokenizer, AutoModelForSequenceClassification, AutoModelForTokenClassification
|
6 |
import time
|
7 |
|
8 |
-
#
|
9 |
st.set_page_config(page_title="Stock News Sentiment Analysis", layout="centered")
|
10 |
|
11 |
st.markdown("""
|
@@ -28,80 +26,7 @@ st.markdown("""
|
|
28 |
</style>
|
29 |
""", unsafe_allow_html=True)
|
30 |
|
31 |
-
#
|
32 |
-
sentiment_model_id = "LinkLinkWu/Stock_Analysis_Test_Ahamed"
|
33 |
-
sentiment_tokenizer = AutoTokenizer.from_pretrained(sentiment_model_id)
|
34 |
-
sentiment_model = AutoModelForSequenceClassification.from_pretrained(sentiment_model_id)
|
35 |
-
sentiment_pipeline = pipeline("sentiment-analysis", model=sentiment_model, tokenizer=sentiment_tokenizer)
|
36 |
-
|
37 |
-
ner_tokenizer = AutoTokenizer.from_pretrained("dslim/bert-base-NER")
|
38 |
-
ner_model = AutoModelForTokenClassification.from_pretrained("dslim/bert-base-NER")
|
39 |
-
ner_pipeline = pipeline("ner", model=ner_model, tokenizer=ner_tokenizer, grouped_entities=True)
|
40 |
-
|
41 |
-
# ----------- Functions -----------
|
42 |
-
def fetch_news(ticker):
|
43 |
-
try:
|
44 |
-
url = f"https://finviz.com/quote.ashx?t={ticker}"
|
45 |
-
headers = {
|
46 |
-
'User-Agent': 'Mozilla/5.0',
|
47 |
-
'Accept': 'text/html',
|
48 |
-
'Accept-Language': 'en-US,en;q=0.5',
|
49 |
-
'Referer': 'https://finviz.com/',
|
50 |
-
'Connection': 'keep-alive',
|
51 |
-
}
|
52 |
-
response = requests.get(url, headers=headers)
|
53 |
-
if response.status_code != 200:
|
54 |
-
st.error(f"Failed to fetch page for {ticker}: Status code {response.status_code}")
|
55 |
-
return []
|
56 |
-
|
57 |
-
soup = BeautifulSoup(response.text, 'html.parser')
|
58 |
-
title = soup.title.text if soup.title else ""
|
59 |
-
if ticker not in title:
|
60 |
-
st.error(f"Page for {ticker} not found or access denied.")
|
61 |
-
return []
|
62 |
-
|
63 |
-
news_table = soup.find(id='news-table')
|
64 |
-
if news_table is None:
|
65 |
-
st.error(f"News table not found for {ticker}. The website structure might have changed.")
|
66 |
-
return []
|
67 |
-
|
68 |
-
news = []
|
69 |
-
for row in news_table.findAll('tr')[:50]:
|
70 |
-
a_tag = row.find('a')
|
71 |
-
if a_tag:
|
72 |
-
title = a_tag.get_text()
|
73 |
-
link = a_tag['href']
|
74 |
-
news.append({'title': title, 'link': link})
|
75 |
-
return news
|
76 |
-
except Exception as e:
|
77 |
-
st.error(f"Failed to fetch news for {ticker}: {e}")
|
78 |
-
return []
|
79 |
-
|
80 |
-
def analyze_sentiment(text):
|
81 |
-
try:
|
82 |
-
result = sentiment_pipeline(text)[0]
|
83 |
-
return "Positive" if result['label'] == 'POSITIVE' else "Negative"
|
84 |
-
except Exception as e:
|
85 |
-
st.error(f"Sentiment analysis failed: {e}")
|
86 |
-
return "Unknown"
|
87 |
-
|
88 |
-
def extract_org_entities(text):
|
89 |
-
try:
|
90 |
-
entities = ner_pipeline(text)
|
91 |
-
org_entities = []
|
92 |
-
for ent in entities:
|
93 |
-
if ent["entity_group"] == "ORG":
|
94 |
-
clean_word = ent["word"].replace("##", "").strip()
|
95 |
-
if clean_word.upper() not in org_entities:
|
96 |
-
org_entities.append(clean_word.upper())
|
97 |
-
if len(org_entities) >= 5:
|
98 |
-
break
|
99 |
-
return org_entities
|
100 |
-
except Exception as e:
|
101 |
-
st.error(f"NER entity extraction failed: {e}")
|
102 |
-
return []
|
103 |
-
|
104 |
-
# ----------- UI -----------
|
105 |
st.title("\U0001F4CA Stock News Sentiment Analysis")
|
106 |
st.markdown("""
|
107 |
This tool analyzes the sentiment of news articles related to companies you mention in text.
|
@@ -118,6 +43,7 @@ if tickers:
|
|
118 |
else:
|
119 |
tickers = []
|
120 |
|
|
|
121 |
if st.button("Get News and Sentiment"):
|
122 |
if not tickers:
|
123 |
st.warning("Please mention at least one recognizable company.")
|
@@ -152,10 +78,7 @@ if st.button("Get News and Sentiment"):
|
|
152 |
sentiment = sentiments[i-1]
|
153 |
st.markdown(f"{i}. [{news['title']}]({news['link']}) - **{sentiment}**")
|
154 |
|
155 |
-
|
156 |
-
st.write(f"**Overall Sentiment for {ticker}: {overall_sentiment}**")
|
157 |
-
else:
|
158 |
-
st.write("**No clear sentiment (does not meet threshold conditions).**")
|
159 |
else:
|
160 |
st.write(f"No news available for {ticker}.")
|
161 |
|
|
|
1 |
import streamlit as st
|
2 |
+
from func import fetch_news, analyze_sentiment, extract_org_entities
|
3 |
+
|
|
|
|
|
4 |
import time
|
5 |
|
6 |
+
# ---------------- Page Setup ----------------
|
7 |
st.set_page_config(page_title="Stock News Sentiment Analysis", layout="centered")
|
8 |
|
9 |
st.markdown("""
|
|
|
26 |
</style>
|
27 |
""", unsafe_allow_html=True)
|
28 |
|
29 |
+
# ---------------- User Interface ----------------
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
st.title("\U0001F4CA Stock News Sentiment Analysis")
|
31 |
st.markdown("""
|
32 |
This tool analyzes the sentiment of news articles related to companies you mention in text.
|
|
|
43 |
else:
|
44 |
tickers = []
|
45 |
|
46 |
+
# ---------------- Button Trigger ----------------
|
47 |
if st.button("Get News and Sentiment"):
|
48 |
if not tickers:
|
49 |
st.warning("Please mention at least one recognizable company.")
|
|
|
78 |
sentiment = sentiments[i-1]
|
79 |
st.markdown(f"{i}. [{news['title']}]({news['link']}) - **{sentiment}**")
|
80 |
|
81 |
+
st.write(f"**Overall Sentiment for {ticker}: {overall_sentiment}**")
|
|
|
|
|
|
|
82 |
else:
|
83 |
st.write(f"No news available for {ticker}.")
|
84 |
|