Update main.py

main.py

@@ -10,28 +10,32 @@ import matplotlib.pyplot as plt
 import base64
 from io import BytesIO
 
-# Set writable cache directories
-os.environ["HF_HOME"] = "/root/.cache/huggingface"
-os.environ["MPLCONFIGDIR"] = "/tmp/matplotlib"  # Use /tmp instead of /root
+#  Set writable cache directories for Hugging Face and Matplotlib
+os.environ["TRANSFORMERS_CACHE"] = "/tmp/huggingface_cache"
+os.environ["HF_HOME"] = "/tmp/huggingface_cache"
+os.environ["MPLCONFIGDIR"] = "/tmp/matplotlib"
 
-#  Use Hugging Face token from Secrets (instead of the protected /root path)
+# Ensure the directories exist
+os.makedirs("/tmp/huggingface_cache", exist_ok=True)
+os.makedirs("/tmp/matplotlib", exist_ok=True)
+
+# Retrieve Hugging Face Token securely from environment variables
 HF_TOKEN = os.getenv("HUGGINGFACE_TOKEN", None)
 
 app = Flask(__name__)
 
-# Load Model from Hugging Face with Explicit Token
-MODEL_HF_REPO = "philipobiorah/bert-imdb-model"  # Replace with your Hugging Face model repo
-
-
-
-
+#  Fix the Hugging Face Model Repository Name
+MODEL_HF_REPO = "philipobiorah/bert-imdb-model"  # Ensure this exists on Hugging Face
 
 print("🚀 Loading model from Hugging Face Hub with authentication...")
-
-model = BertForSequenceClassification.from_pretrained(
-    MODEL_HF_REPO,
-    token=os.getenv("HUGGINGFACE_TOKEN", None)
-)
+try:
+    model = BertForSequenceClassification.from_pretrained(
+        MODEL_HF_REPO,
+        token=HF_TOKEN,  # Ensure correct authentication
+    )
+except Exception as e:
+    print(f"❌ Error loading model: {e}")
+    exit(1)
 
 model.eval()
 tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
@@ -73,19 +77,19 @@ def upload_file_post():
         # Predict sentiment for each review
         data['sentiment'] = data['review'].apply(predict_sentiment)
 
-        #  Sentiment Analysis Summary
+        # Sentiment Analysis Summary
         sentiment_counts = data['sentiment'].value_counts().to_dict()
         summary = f"Total Reviews: {len(data)}<br>" \
                   f"Positive: {sentiment_counts.get('Positive', 0)}<br>" \
                   f"Negative: {sentiment_counts.get('Negative', 0)}<br>"
 
-        #  Generate bar chart
+        # Generate bar chart
         fig, ax = plt.subplots()
         ax.bar(sentiment_counts.keys(), sentiment_counts.values(), color=['red', 'blue'])
         ax.set_ylabel('Counts')
         ax.set_title('Sentiment Analysis Summary')
 
-        #  Convert plot to base64 for embedding
+        # Convert plot to base64 for embedding
         img = BytesIO()
         plt.savefig(img, format='png', bbox_inches='tight')
         img.seek(0)