Create app.py

app.py

@@ -0,0 +1,111 @@
+import requests
+import pandas as pd
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+import torch
+import torch.nn.functional as F
+from datetime import datetime
+
+# ==========================
+# 1️⃣ FINBERT LOAD
+# ==========================
+MODEL_NAME = "ProsusAI/finbert"
+
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = AutoModelForSequenceClassification.from_pretrained(MODEL_NAME)
+model.eval()
+
+# ==========================
+# 2️⃣ REUTERS SCRAPER
+# ==========================
+def fetch_reuters_articles(section="/business/", size=20):
+    """
+    Reuters unofficial JSON endpoint
+    """
+    url = "https://www.reuters.com/pf/api/v3/content/fetch/articles-by-section-alias-or-id-v1"
+    
+    query = {
+        "query": {
+            "section_id": section,
+            "size": size
+        }
+    }
+
+    params = {
+        "query": str(query).replace("'", '"')
+    }
+
+    headers = {
+        "User-Agent": "Mozilla/5.0"
+    }
+
+    response = requests.get(url, params=params, headers=headers)
+    data = response.json()
+
+    articles = []
+    try:
+        contents = data["result"]["articles"]
+        for item in contents:
+            headline = item.get("title", "")
+            summary = item.get("description", "")
+            published = item.get("published_time", "")
+            articles.append({
+                "headline": headline,
+                "summary": summary,
+                "published": published
+            })
+    except Exception as e:
+        print("Parsing error:", e)
+
+    return pd.DataFrame(articles)
+
+
+# ==========================
+# 3️⃣ FINBERT SENTIMENT
+# ==========================
+def finbert_score(text):
+    inputs = tokenizer(
+        text,
+        return_tensors="pt",
+        truncation=True,
+        padding=True,
+        max_length=512
+    )
+
+    with torch.no_grad():
+        outputs = model(**inputs)
+        probs = F.softmax(outputs.logits, dim=1)
+
+    labels = ["negative", "neutral", "positive"]
+    scores = probs[0].tolist()
+
+    return dict(zip(labels, scores))
+
+
+# ==========================
+# 4️⃣ PIPELINE
+# ==========================
+def run_pipeline():
+    df = fetch_reuters_articles(section="/business/", size=25)
+
+    sentiments = []
+    for _, row in df.iterrows():
+        text = row["headline"] + ". " + row["summary"]
+        score = finbert_score(text)
+        sentiments.append(score)
+
+    sentiment_df = pd.DataFrame(sentiments)
+    final_df = pd.concat([df, sentiment_df], axis=1)
+
+    # Hawkish score (neg - pos)
+    final_df["hawkish_score"] = final_df["negative"] - final_df["positive"]
+
+    filename = f"reuters_finbert_{datetime.today().date()}.csv"
+    final_df.to_csv(filename, index=False)
+
+    print("Saved:", filename)
+    print(final_df.head())
+
+
+# ==========================
+if __name__ == "__main__":
+    run_pipeline()