Update app.py

app.py

@@ -84,4 +84,118 @@ else:
 
 
 
+Begindatestring = start_date
+
+
+#Begindatestring = datetime.strptime(Begindatestring, "%Y-%m-%d").date()
+
+
+val = 39448 + int(delta.days)
+url = 'https://economictimes.indiatimes.com/archivelist/year-'+str(Begindatestring.year)+',month-'+str(Begindatestring.month)+',starttime-'+str(val)+'.cms'  # Replace with your URL
+
+response = requests.get(url)
+
+if response.status_code == 200:
+    html_text = response.text
+    soup = BeautifulSoup(html_text, "lxml")
+else:
+    st.write(f"Failed to fetch the page. Status code: {response.status_code}")
+jobs = soup.find_all("li")
+headlines = []
+for job in jobs:
+        try:
+            target_element = job.find("a")
+            target_element.text
+            headlines.append(target_element.text)
+        except:
+            continue
+
+
+
+
+index = [idx for idx, s in enumerate(headlines) if s=='Most Read' ][0]
+del headlines[index:]
+news = pd.DataFrame({"News": headlines})
+news.insert(0, 'Date', Begindatestring)
+#st.dataframe(df[0:1])
+
+
+news = news.drop_duplicates()
+news = news.dropna(how='any')
+news = news.reset_index(drop=True)
+
+import pandas as pd
+import numpy as np
+
+
+from transformers import pipeline
+import torch
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+
+tokenizer = AutoTokenizer.from_pretrained("nickmuchi/sec-bert-finetuned-finance-classification")
+model = AutoModelForSequenceClassification.from_pretrained("nickmuchi/sec-bert-finetuned-finance-classification")
+
+
+
+
+
+
+
+
+
+
+nlp = pipeline("text-classification", model=model, tokenizer=tokenizer, device=device)
+
+length = len(news[ 'News'].to_list())
+news_list = news[ 'News'].to_list() 
+
+df = pd.DataFrame()
+for i in range (0, length):
+
+
+    results = nlp(news_list[i])
+    df.loc[i, "News"] = news_list[i]
+    df.loc[i , 'label'] = results[0]["label"]
+    df.loc[i , 'score'] = results[0]["score"]
+
+
+
+#st.dataframe(df)
+
+# Filter the DataFrame to get rows with "neutral" sentiment
+bullish_rows = df[df['label'] == 'bullish']
+
+# Calculate the sum of the 'Score' column for "neutral" rows
+bullish_score_sum = bullish_rows['score'].sum()
+
+num_bullish_rows = len(bullish_rows)
+# Calculate the average score for "neutral" sentiment
+average_score_for_bullish = bullish_score_sum / num_bullish_rows
+
+
+# Filter the DataFrame to get rows with "neutral" sentiment
+bearish_rows = df[df['label'] == 'bearish']
+
+# Calculate the sum of the 'Score' column for "neutral" rows
+bearish_score_sum = bearish_rows['score'].sum()
+
+# Cabearishlculate the number of "neutral" rows
+num_bearish_rows = len(bearish_rows)
+
+# Calculate the average score for "neutral" sentiment
+average_score_for_bearish = bearish_score_sum / num_bearish_rows
+
+
+if(average_score_for_bearish > average_score_for_bullish):
+     st.write("Stock will go down")
+if(average_score_for_bearish < average_score_for_bullish):
+     st.write("Stock will go up")
+
+
+
+
+