Create pipeline.py

pipeline.py

@@ -0,0 +1,51 @@
+import numpy as np
+import tensorflow as tf
+import keras
+import json
+from keras.preprocessing.sequence import pad_sequences
+from huggingface_hub import hf_hub_download
+
+# Download model
+model_path = hf_hub_download(repo_id="NightPrince/Toxic_Classification", filename="toxic_classifier.keras")
+
+# Download tokenizer
+tokenizer_path = hf_hub_download(repo_id="NightPrince/Toxic_Classification", filename="tokenizer.json")
+
+# Load model
+model = keras.saving.load_model(model_path)
+
+# Load tokenizer
+from keras.preprocessing.text import tokenizer_from_json
+with open(tokenizer_path, "r", encoding="utf-8") as f:
+    tokenizer_json = f.read()
+    tokenizer = tokenizer_from_json(tokenizer_json)
+
+# Label map (same as config)
+label_map = {
+    0: "Child Sexual Exploitation",
+    1: "Elections",
+    2: "Non-Violent Crimes",
+    3: "Safe",
+    4: "Sex-Related Crimes",
+    5: "Suicide & Self-Harm",
+    6: "Unknown S-Type",
+    7: "Violent Crimes",
+    8: "Unsafe"
+}
+
+# Pipeline function
+def classify_toxic(query, image_description):
+    max_len = 150  # Keep it same as training
+    text = query + " " + image_description
+    seq = tokenizer.texts_to_sequences([text])
+    pad = pad_sequences(seq, maxlen=max_len, padding='post', truncating='post')
+    pred = model.predict(pad, verbose=0)
+    pred_label = np.argmax(pred, axis=1)[0]
+    return label_map.get(pred_label, "Unknown")
+
+# Example usage
+if __name__ == "__main__":
+    query = "This is a dangerous post"
+    image_desc = "Knife shown in the image"
+    result = classify_toxic(query, image_desc)
+    print("Predicted Category:", result)