Update app.py

app.py

@@ -1,68 +1,79 @@
-import os
+# import os
+
+# # Set Java paths manually
+# os.environ["JAVA_HOME"] = "/usr/local/lib/jvm/java-17-openjdk-amd64"
+# os.environ["JVM_PATH"] = "/usr/local/lib/jvm/java-17-openjdk-amd64/lib/server/libjvm.so"
+# os.environ["PATH"] = os.environ["JAVA_HOME"] + "/bin:" + os.environ["PATH"]
 
-# Set Java paths manually
-os.environ["JAVA_HOME"] = "/usr/local/lib/jvm/java-17-openjdk-amd64"
-os.environ["JVM_PATH"] = "/usr/local/lib/jvm/java-17-openjdk-amd64/lib/server/libjvm.so"
-os.environ["PATH"] = os.environ["JAVA_HOME"] + "/bin:" + os.environ["PATH"]
+# from flask import Flask, request, jsonify
+# from sentence_transformers import CrossEncoder
+# import py_vncorenlp
 
-from flask import Flask, request, jsonify
-from sentence_transformers import CrossEncoder
-import py_vncorenlp
 
+# app = Flask(__name__)
+# save_dir_vncore = "/home/user/app/vncorenlp"
+# rdrsegmenter = py_vncorenlp.VnCoreNLP(annotators=["wseg"], save_dir=save_dir_vncore)
 
-app = Flask(__name__)
-save_dir_vncore = "/home/user/app/vncorenlp"
-rdrsegmenter = py_vncorenlp.VnCoreNLP(annotators=["wseg"], save_dir=save_dir_vncore)
 
+# # Load your cross-encoder model
+# model_name = "truong1301/reranker_pho_BLAI"  # Replace with your actual model if different
+# cross_encoder = CrossEncoder(model_name, max_length=256, num_labels=1)
 
-# Load your cross-encoder model
-model_name = "truong1301/reranker_pho_BLAI"  # Replace with your actual model if different
-cross_encoder = CrossEncoder(model_name, max_length=256, num_labels=1)
+# # Function to preprocess text with Vietnamese word segmentation
+# def preprocess_text(text):
+#     if not text:
+#         return text
+#     segmented_text = rdrsegmenter.word_segment(text)
+#     # Join tokenized sentences into a single string
+#     return " ".join([" ".join(sentence) for sentence in segmented_text])
 
-# Function to preprocess text with Vietnamese word segmentation
-def preprocess_text(text):
-    if not text:
-        return text
-    segmented_text = rdrsegmenter.word_segment(text)
-    # Join tokenized sentences into a single string
-    return " ".join([" ".join(sentence) for sentence in segmented_text])
+# @app.route("/rerank", methods=["POST"])
+# def rerank():
+#     try:
+#         # Get JSON data from the request (query and list of documents)
+#         data = request.get_json()
+#         query = data.get("query", "")
+#         documents = data.get("documents", [])
 
-@app.route("/rerank", methods=["POST"])
-def rerank():
-    try:
-        # Get JSON data from the request (query and list of documents)
-        data = request.get_json()
-        query = data.get("query", "")
-        documents = data.get("documents", [])
+#         if not query or not documents:
+#             return jsonify({"error": "Missing query or documents"}), 400
 
-        if not query or not documents:
-            return jsonify({"error": "Missing query or documents"}), 400
+#         # Preprocess query and documents with vncorenlp
+#         processed_query = preprocess_text(query)
+#         processed_docs = [preprocess_text(doc) for doc in documents]
 
-        # Preprocess query and documents with vncorenlp
-        processed_query = preprocess_text(query)
-        processed_docs = [preprocess_text(doc) for doc in documents]
+#         # Create pairs of query and documents for reranking
+#         query_doc_pairs = [(processed_query, doc) for doc in processed_docs]
 
-        # Create pairs of query and documents for reranking
-        query_doc_pairs = [(processed_query, doc) for doc in processed_docs]
+#         # Get reranking scores from the cross-encoder
+#         scores = cross_encoder.predict(query_doc_pairs).tolist()
 
-        # Get reranking scores from the cross-encoder
-        scores = cross_encoder.predict(query_doc_pairs).tolist()
+#         # Combine documents with their scores and sort
+#         ranked_results = sorted(
+#             [{"document": doc, "score": score} for doc, score in zip(documents, scores)],
+#             key=lambda x: x["score"],
+#             reverse=True
+#         )
 
-        # Combine documents with their scores and sort
-        ranked_results = sorted(
-            [{"document": doc, "score": score} for doc, score in zip(documents, scores)],
-            key=lambda x: x["score"],
-            reverse=True
-        )
+#         return jsonify({"results": ranked_results})
 
-        return jsonify({"results": ranked_results})
+#     except Exception as e:
+#         return jsonify({"error": str(e)}), 500
 
-    except Exception as e:
-        return jsonify({"error": str(e)}), 500
+# @app.route("/", methods=["GET"])
+# def health_check():
+#     return jsonify({"status": "Server is running"}), 200
+
+# if __name__ == "__main__":
+#     app.run(host="0.0.0.0", port=7860)  # Default port for Hugging Face Spaces
+import os
+import subprocess
 
-@app.route("/", methods=["GET"])
-def health_check():
-    return jsonify({"status": "Server is running"}), 200
+# Find libjvm.so
+output = subprocess.run("find /usr/lib/jvm -name libjvm.so", shell=True, capture_output=True, text=True)
+print("🔍 Searching for libjvm.so...")
+print(output.stdout or "❌ libjvm.so not found!")
 
-if __name__ == "__main__":
-    app.run(host="0.0.0.0", port=7860)  # Default port for Hugging Face Spaces
+# Print JAVA_HOME and PATH
+print(f"JAVA_HOME: {os.environ.get('JAVA_HOME', 'Not Set')}")
+print(f"PATH: {os.environ.get('PATH', 'Not Set')}")