Update app.py

app.py

@@ -1,6 +1,5 @@
 import gradio as gr
 import gspread
-import re
 from oauth2client.service_account import ServiceAccountCredentials
 from llama_cpp import Llama
 from llama_index.core import VectorStoreIndex, Settings
@@ -15,72 +14,60 @@ from llama_index.core.schema import Document
 # ===================================
 # 1️⃣ Fungsi Membaca Data Google Spreadsheet
 # ===================================
-def read_google_sheets(sheet_name, query=None):
+def read_google_sheets():
     try:
         scope = ["https://www.googleapis.com/auth/spreadsheets", "https://www.googleapis.com/auth/drive"]
         creds = ServiceAccountCredentials.from_json_keyfile_name("credentials.json", scope)
         client = gspread.authorize(creds)
-
+        
         SPREADSHEET_ID = "1e_cNMhwF-QYpyYUpqQh-XCw-OdhWS6EuYsoBUsVtdNg"
+        sheet_names = ["datatarget", "datacuti", "dataabsen", "datalembur","pkb"]
+
+        all_data = []
         spreadsheet = client.open_by_key(SPREADSHEET_ID)
-        sheet = spreadsheet.worksheet(sheet_name)
-        data = sheet.get_all_values()
-        
-        if sheet_name == "pkb" and query:
-            filtered_data = []
-            for row in data[1:]:
-                bab, pasal, judul, isi = row
-                if query.lower() in judul.lower() or query.lower() in isi.lower():
-                    filtered_data.append(f"Bab {bab}, Pasal {pasal}: {judul}\n{isi}\n")
-            return "\n".join(filtered_data) if filtered_data else "Maaf, tidak ditemukan pasal terkait."
         
-        formatted_data = [" | ".join(row) for row in data]
-        return "\n".join(formatted_data).strip()
+        for sheet_name in sheet_names:
+            try:
+                sheet = spreadsheet.worksheet(sheet_name)
+                data = sheet.get_all_values()
+                all_data.append(f"=== Data dari {sheet_name.upper()} ===")
+                all_data.extend([" | ".join(row) for row in data])
+                all_data.append("\n")
+            except gspread.exceptions.WorksheetNotFound:
+                all_data.append(f"❌ ERROR: Worksheet {sheet_name} tidak ditemukan.")
+
+        return "\n".join(all_data).strip()
+    
+    except gspread.exceptions.SpreadsheetNotFound:
+        return "❌ ERROR: Spreadsheet tidak ditemukan!"
     
     except Exception as e:
         return f"❌ ERROR: {str(e)}"
 
 # ===================================
-# 2️⃣ Fungsi Menentukan Sheet yang Relevan
-# ===================================
-def get_relevant_sheet(question):
-    question = question.lower()
-    if re.search(r"(jam kerja|aturan perusahaan|kerja lembur|hak|kewajiban|sanksi|upah|phk|jaminan sosial|disiplin kerja|larangan|upah|sanksi|pesangon|peraturan|sistem kerja)", question):
-        return "pkb"
-    if re.search(r"(target|aktual|selisih).*?(mesin|bulan)", question):
-        return "datatarget"
-    if re.search(r"(sisa cuti|cuti terpakai|cuti.*(nama karyawan)?)", question):
-        return "datacuti"
-    if re.search(r"(lembur|lembur bulan ini.*(nama karyawan)?)", question):
-        return "datalembur"
-    if re.search(r"(kehadiran|hadir|sakit|izin|absen.*(nama karyawan)?)", question):
-        return "dataabsen"
-    return None
-
-# ===================================
-# 3️⃣ Inisialisasi Model Llama
+# 2️⃣ Inisialisasi Model Llama
 # ===================================
 def initialize_llama_model():
     model_path = hf_hub_download(
-        repo_id="TheBloke/zephyr-7b-beta-GGUF",
+        repo_id="TheBLoke/zephyr-7b-beta-GGUF",
         filename="zephyr-7b-beta.Q4_K_M.gguf",
         cache_dir="./models"
     )
     return model_path
 
 # ===================================
-# 4️⃣ Inisialisasi Pengaturan Model
+# 3️⃣ Inisialisasi Pengaturan Model
 # ===================================
 def initialize_settings(model_path):
     Settings.llm = LlamaCPP(model_path=model_path, temperature=0.7)
 
 # ===================================
-# 5️⃣ Inisialisasi Index & Chat Engine
+# 4️⃣ Inisialisasi Index & Chat Engine
 # ===================================
 def initialize_index():
-    text_data = read_google_sheets("pkb")
+    text_data = read_google_sheets()
     document = Document(text=text_data)
-    parser = SentenceSplitter(chunk_size=200, chunk_overlap=20)
+    parser = SentenceSplitter(chunk_size=100, chunk_overlap=30)
     nodes = parser.get_nodes_from_documents([document])
     
     embedding = HuggingFaceEmbedding("sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2")
@@ -93,21 +80,21 @@ def initialize_chat_engine(index):
     retriever = index.as_retriever(similarity_top_k=1)
     chat_engine = CondensePlusContextChatEngine.from_defaults(
         retriever=retriever,
-        verbose=False
+        verbose=False  # ❌ Hapus verbose agar tidak ada referensi dokumen
     )
     return chat_engine
 
 # ===================================
-# 6️⃣ Fungsi Merapikan Jawaban Chatbot
+# 5️⃣ Fungsi untuk Merapikan Jawaban Chatbot
 # ===================================
 def clean_response(response):
     text = "".join(response.response_gen)  # Gabungkan teks yang dihasilkan
     text = text.replace("\n\n", "\n").strip()  # Hilangkan newline berlebihan
-    text = text.split("\n")[0]  # Ambil hanya kalimat pertama agar tidak menjawab berlebihan
+    text = text.replace("user:", "").replace("jawaban:", "").replace("assistant:", "").strip()
     return text
 
 # ===================================
-# 7️⃣ Fungsi Utama Chatbot
+# 6️⃣ Fungsi untuk Menghasilkan Respons Chatbot
 # ===================================
 def generate_response(message, history, chat_engine):
     if history is None:
@@ -132,18 +119,18 @@ def generate_response(message, history, chat_engine):
     return cleaned_text
 
 # ===================================
-# 8️⃣ Menjalankan Aplikasi Chatbot
+# 7️⃣ Fungsi Utama untuk Menjalankan Aplikasi
 # ===================================
 def main():
     model_path = initialize_llama_model()
     initialize_settings(model_path)
-    
+
     index = initialize_index()
     chat_engine = initialize_chat_engine(index)
-    
+
     def chatbot_response(message, history):
         return generate_response(message, history, chat_engine)
-    
+
     gr.Interface(
         fn=chatbot_response,
         inputs=["text"],