Update app.py

app.py

@@ -1,101 +1,261 @@
-# Import Library yang Diperlukan
-import gradio as gr
-import shutil
-import os
-import subprocess
+# -*- coding: utf-8 -*-
+"""Chatbot
+
+Automatically generated by Colab.
+
+Original file is located at
+    https://colab.research.google.com/drive/1mbREVflL7r6y0hCUPE-MD3VzJJUs5OfR
+"""
+
+!huggingface-cli download TheBloke/zephyr-7B-beta-GGUF zephyr-7b-beta.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
+
+!CMAKE_ARGS="-DGGML_CUDA=on" pip install llama-cpp-python
+
 from llama_cpp import Llama
-from llama_index.core import SimpleDirectoryReader, VectorStoreIndex, Settings
-from llama_index.core.llms import ChatMessage
+
+llm = Llama(
+      model_path="/content/zephyr-7b-beta.Q4_K_M.gguf",
+      n_gpu_layers=-1, # Uncomment to use GPU acceleration
+      # seed=1337, # Uncomment to set a specific seed
+      # n_ctx=2048, # Uncomment to increase the context window
+)
+
+llm("What is Large Language Model?", max_tokens=128)
+
+!pip install llama-index gradio
+
+pip install llama-index-llms-llama-cpp
+
+from typing import Sequence, Optional
+
+from llama_index.core import SimpleDirectoryReader, VectorStoreIndex
 from llama_index.llms.llama_cpp import LlamaCPP
-from llama_index.embeddings.huggingface import HuggingFaceEmbedding
-from huggingface_hub import hf_hub_download
-from llama_index.core.node_parser import SentenceSplitter
+from llama_index.llms.llama_cpp.llama_utils import (
+    messages_to_prompt,
+    completion_to_prompt,
+)
+from llama_index.core.llms import ChatMessage
 
-# Fungsi untuk mengunduh model Llama
-def initialize_llama_model():
-    # Unduh model jika belum ada di direktori kerja
-    model_path = hf_hub_download(
-        repo_id="TheBLoke/zephyr-7b-beta-GGUF",  # Nama repo model
-        filename="zephyr-7b-beta.Q4_K_M.gguf",  # Nama file model
-        cache_dir="./models"  # Lokasi direktori untuk menyimpan model
-    )
-    return model_path
+def zephyr_messages_to_prompt(
+    messages: Sequence[ChatMessage],
+    system_prompt: Optional[str]=None
+) -> str:
+    prompt = ""
+    for message in messages:
+        prompt += f"<|{message.role}|>\n"
+        prompt += f"{message.content}</s>\n"
 
-# Fungsi untuk mengatur konfigurasi Settings
-def initialize_settings(model_path):  
-    Settings.llm = LlamaCPP(
-        model_path=model_path,
-        temperature=0.7,
-    )
+    return prompt + "<|assistant|>\n"
+
+llm = LlamaCPP(
+    # You can pass in the URL to a GGML model to download it automatically
+    model_url=None,
+    # optionally, you can set the path to a pre-downloaded model instead of model_url
+    model_path="/content/zephyr-7b-beta.Q4_K_M.gguf",
+    temperature=0.1,
+    max_new_tokens=1000,
+    # llama2 has a context window of 4096 tokens, but we set it lower to allow for some wiggle room
+    context_window=3900,
+    # kwargs to pass to __call__()
+    generate_kwargs={},
+    # kwargs to pass to __init__()
+    # set to at least 1 to use GPU
+    model_kwargs={"n_gpu_layers": 1},
+    # transform inputs into Llama2 format
+    messages_to_prompt=messages_to_prompt,
+    completion_to_prompt=completion_to_prompt,
+    verbose=True,
+)
 
-# Fungsi untuk Menginisialisasi Index
-def initialize_index():
-    # Tentukan dokumen input untuk pembacaan data
-    documents = SimpleDirectoryReader(input_files=["bahandokumen/K3.txt",
-                                                   "bahandokumen/bonus.txt",
-                                                   "bahandokumen/cuti.txt",
-                                                   "bahandokumen/disiplinkerja.txt",
-                                                   "bahandokumen/fasilitas&bantuan.txt",
-                                                   "bahandokumen/upahlembur.txt",
-                                                   "bahandokumen/waktukerja.txt"]).load_data()
-
-    parser = SentenceSplitter(chunk_size=150, chunk_overlap=10)
-    nodes = parser.get_nodes_from_documents(documents)
-    embedding = HuggingFaceEmbedding("BAAI/bge-base-en-v1.5")
-    Settings.embed_model = embedding
-    index = VectorStoreIndex(nodes)
-    return index
-
-# Inisialisasi Mesin Chat
-def initialize_chat_engine(index):
-    from llama_index.core.prompts import PromptTemplate
-    from llama_index.core.chat_engine.condense_plus_context import CondensePlusContextChatEngine
-    retriever = index.as_retriever(similarity_top_k=3)
-    chat_engine = CondensePlusContextChatEngine.from_defaults(
-        retriever=retriever,
-        verbose=True,
+messages = [
+    ChatMessage(
+        role="system",
+        content="Anda adalah chatbot yang selalu menjawab pertanyaan secara singkat, ramah, dan jelas dalam bahasa Indonesia."
+    ),
+    ChatMessage(
+        role="user",
+        content="There's a llama on my lawn, how can I get rid of him?"
     )
-    return chat_engine
+]
+
+# zephyr_messages_to_prompt(messages)
+
+response = llm.stream_chat(messages)
+for r in response:
+  print(r.delta, end="", flush=True)
+
+!pip install flask-ngrok
+!pip install pymongo
+
+!ngrok http 27017
+
+import gradio as gr
 
-# Fungsi untuk menghasilkan respons chatbot
-def generate_response(message, history, chat_engine):
+
+def generate_response(message, history):
     chat_messages = [
         ChatMessage(
             role="system",
             content="Anda adalah chatbot yang selalu menjawab pertanyaan secara singkat, ramah, dan jelas dalam bahasa Indonesia."
         ),
     ]
+    for human, ai in history:
+        chat_messages.append(ChatMessage(
+            role="user",
+            content=human
+        ))
+        chat_messages.append(ChatMessage(
+            role="assistant",
+            content=ai
+        ))
+
+    chat_messages.append(ChatMessage(
+        role="user",
+        content=message
+    ))
+
+    response = llm.stream_chat(chat_messages)
+    text = ""
+    for r in response:
+      text += r.delta
+      yield text
+
+gr.ChatInterface(generate_response).launch()
+
+from google.colab import files
+uploaded = files.upload()
+for filename in uploaded.keys():
+  print(uploaded[filename])
+
+!pip install pypdf -q
+
+from llama_index.core import SimpleDirectoryReader
+documents = SimpleDirectoryReader(input_files=["/content/bahandokumen/K3.txt",
+                                                   "/content/bahandokumen/bonus.txt",
+                                                   "/content/bahandokumen/jadwallembur.txt",
+                                                   "/content/bahandokumen/datalembur.txt",
+                                                   "/content/bahandokumen/absensi.txt",
+                                                   "/content/bahandokumen/sisacuti.txt",
+                                                   "/content/bahandokumen/target.txt",
+                                                   "/content/bahandokumen/cuti.txt",
+                                                   "/content/bahandokumen/disiplinkerja.txt",
+                                                   "/content/bahandokumen/fasilitas&bantuan.txt",
+                                                   "/content/bahandokumen/fasilitaskerja.txt",
+                                                   "/content/bahandokumen/hak.txt",
+                                                   "/content/bahandokumen/hubunganpengusaha&serikat.txt",
+                                                   "/content/bahandokumen/istilah.txt",
+                                                   "/content/bahandokumen/jaminanserikat.txt",
+                                                   "/content/bahandokumen/jamkes.txt",
+                                                   "/content/bahandokumen/jamsos.txt",
+                                                   "/content/bahandokumen/keluhkesah.txt",
+                                                   "/content/bahandokumen/kenaikanupah.txt",
+                                                   "/content/bahandokumen/kewajiban.txt",
+                                                   "/content/bahandokumen/kompensasi.txt",
+                                                   "/content/bahandokumen/larangan.txt",
+                                                   "/content/bahandokumen/lembur.txt",
+                                                   "/content/bahandokumen/luaskesepakatan.txt",
+                                                   "/content/bahandokumen/mogok.txt",
+                                                   "/content/bahandokumen/pelanggaran&sanksi.txt",
+                                                   "/content/bahandokumen/pendidikan.txt",
+                                                   "/content/bahandokumen/pengangkatan.txt",
+                                                   "/content/bahandokumen/penilaian&promosi.txt",
+                                                   "/content/bahandokumen/pensiun.txt",
+                                                   "/content/bahandokumen/perjadin.txt",
+                                                   "/content/bahandokumen/pesangon.txt",
+                                                   "/content/bahandokumen/phk.txt",
+                                                   "/content/bahandokumen/pihak.txt",
+                                                   "/content/bahandokumen/pkb.txt",
+                                                   "/content/bahandokumen/resign.txt",
+                                                   "/content/bahandokumen/sanksi.txt",
+                                                   "/content/bahandokumen/shift.txt",
+                                                   "/content/bahandokumen/syaratkerja.txt",
+                                                   "/content/bahandokumen/tatacara.txt",
+                                                   "/content/bahandokumen/tka.txt",
+                                                   "/content/bahandokumen/tunjangan.txt",
+                                                   "/content/bahandokumen/uangpisah.txt",
+                                                   "/content/bahandokumen/upah.txt",
+                                                   "/content/bahandokumen/upahlembur.txt",
+                                                   "/content/bahandokumen/waktukerja.txt"]).load_data()
+
+from llama_index.core.node_parser import SentenceSplitter
+parser = SentenceSplitter(chunk_size=300, chunk_overlap=20)
+nodes = parser.get_nodes_from_documents(documents)
+
+nodes[1].get_content()
+
+!pip install llama-index-embeddings-huggingface
+
+from llama_index.embeddings.huggingface import HuggingFaceEmbedding
+from llama_index.core import Settings
+
+embedding = HuggingFaceEmbedding("BAAI/bge-base-en-v1.5")
+Settings.llm=llm
+Settings.embed_model=embedding
+
+from llama_index.core import VectorStoreIndex
+
+index = VectorStoreIndex(nodes)
+
+retriever = index.as_retriever(similarity_top_k=3)
+
+results = retriever.retrieve("Sebutkan Jadwal Lembur Tahun 2024?")
+
+len(results)
+
+results[0].get_content()
+
+query_engine = index.as_query_engine(similarity_top_k=3)
+response = query_engine.query("Sebutkan Jadwal Lembur Tahun 2024?")
+response.response
+
+from llama_index.core.prompts import PromptTemplate
+from llama_index.core.llms import ChatMessage, MessageRole
+from llama_index.core.chat_engine.condense_plus_context import (
+    CondensePlusContextChatEngine,
+)
+
+chat_engine = CondensePlusContextChatEngine.from_defaults(
+    retriever=retriever,
+    verbose=True,
+)
+
+response = chat_engine.chat("Sebutkan Jadwal Lembur Tahun 2024?")
+
+response.response
+
+response = chat_engine.chat("Apakah anda bisa menjelaskan target Bulan Januari?")
+
+response.response
+
+import gradio as gr
+
+chat_engine.reset()
+
+def generate_response(message, history):
+    chat_messages = [
+        ChatMessage(
+            role="system",
+            content="Anda adalah chatbot yang selalu pertanyaan dalam bahasa Indonesia dengan singkat dan ramah."
+        ),
+    ]
     response = chat_engine.stream_chat(message)
-    text = "".join(response.response_gen)  # Gabungkan semua token menjadi string
-    history.append((message, text))  # Tambahkan ke riwayat
-    return history
-
-def clear_history(chat_engine):
-    chat_engine.clear()
-    
-# Inisialisasi Komponen Gradio untuk UI
-def launch_gradio(chat_engine):
-    with gr.Blocks() as demo:
-        # Mengatur tombol untuk menghapus riwayat chat
-        clear_btn = gr.Button("Clear")
-        clear_btn.click(lambda: clear_history(chat_engine))
-
-        # Membuat antarmuka chat
-        chat_interface = gr.ChatInterface(
-            lambda message, history: generate_response(message, history, chat_engine)
-        )
-    demo.launch()
-
-# Fungsi Utama untuk Menjalankan Aplikasi
-def main():
-    # Unduh model dan inisialisasi pengaturan
-    model_path = initialize_llama_model()
-    initialize_settings(model_path)  # Mengirimkan model_path ke fungsi initialize_settings
-    # Inisialisasi index dan engine
-    index = initialize_index()
-    chat_engine = initialize_chat_engine(index)
-    # Luncurkan antarmuka
-    launch_gradio(chat_engine)
-    
-if __name__ == "__main__":
-    main()
+    text = ""
+    for token in response.response_gen:
+        text += token
+        yield text
+
+def clear_history():
+    chat_engine.reset()
+
+# Membuat tombol 'Clear'
+clear_btn = gr.Button("Clear")
+
+# Mengatur layout dengan gr.Blocks
+with gr.Blocks() as demo:
+    # Mengatur tombol untuk menghapus riwayat chat
+    clear_btn.click(clear_history)
+
+    # Membuat antarmuka chat tanpa argumen clear_btn
+    chat_interface = gr.ChatInterface(generate_response)
+
+demo.launch()