Create utils.py

utils.py

@@ -0,0 +1,35 @@
+from transformers import pipeline
+from gtts import gTTS
+
+# Loading models
+summarizer = pipeline("summarization", model="facebook/bart-large-cnn") # Load summarizer
+sentiment_analyzer = pipeline("sentiment-analysis") # Load sentiment analyzer
+# classifier = pipeline("zero-shot-classification", model="facebook/bart-large-mnli") # Load classifier
+
+def analyze_sentiment(text):
+    result = sentiment_analyzer(text[:500])[0]
+    return result['label']
+
+def summarize_text(text):
+    cleaned_text = text.strip().replace("\n", " ")
+    cleaned_text = cleaned_text[:3000]  # Limit to avoid token overflow
+
+    result = summarizer(
+        cleaned_text,
+        max_length=130,
+        min_length=30,
+        do_sample=False
+    )
+
+    summary_text = result[0]['summary_text']
+    return summary_text
+
+def generate_hindi_tts(text, filename="output.mp3"):
+    try:
+        tts = gTTS(text=text, lang='hi')
+        tts.save(filename)
+        print(f"Hindi audio saved to {filename}")
+        return filename
+    except Exception as e:
+        print(f"Error in generating the TTS: {e}")
+        return None