finish

Dockerfile

@@ -1,6 +1,5 @@
 FROM python:3.12-slim
 
-# Set direktori kerja di dalam container
 WORKDIR /app
 
 RUN apt-get update && apt-get install -y \

chroma_data/cdcb1c1c-f374-4f62-9cc7-7e62dcdaccd0/length.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:acd769936f1d22a7253e4ff279f1366aac68ebb667c50946415ea98c3dbcc1bd
+oid sha256:4a2b2eaaf35eded0d090310dd7b61d5b8881b439105becc13f6387989c3fb27a
 size 400

chroma_data/chroma.sqlite3

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:26051dcbbbc7ed704c9c79a0e8770dfe0cf2634b5e8a1649558c284cbff6b780
+oid sha256:db126d77883a3f3f04dd9e92788769e76e4a6bff24270eb1150dd94cd95b54c1
 size 1761280

db_setup.py

@@ -40,11 +40,9 @@ def main():
         )
         documents.append(doc)
 
-    # 4. Setup Embedding Model
     print("Memuat Embedding Model")
     embeddings = HuggingFaceEmbeddings(model_name="Qwen/Qwen3-Embedding-0.6B")
 
-    # 5. Simpan ke ChromaDB via LangChain
     print("Menyimpan data ke ChromaDB")
     Chroma.from_documents(
         documents=documents,

docker-compose.yml

@@ -26,7 +26,7 @@ services:
   frontend_web:
     image: nginx:alpine
     container_name: chilicare_frontend
-    # Nginx untuk menjalankan web HTML 
+    # Nginx untuk web HTML 
     ports:
       - "3000:80"
     volumes:

frontend/index.html

@@ -74,7 +74,6 @@
         background: #94a3b8;
       }
 
-      /* Sembunyikan scrollbar di menu mobile agar lebih rapi */
       .hide-scrollbar::-webkit-scrollbar {
         display: none;
       }

frontend/script.js

@@ -23,10 +23,8 @@ function parseMarkdown(text) {
           tableHtml += '<tr class="border-b border-slate-200 last:border-0 hover:bg-slate-50 transition-colors">';
           cols.forEach(col => {
               if (index === 0) {
-                  // Baris pertama jadi Header (TH)
                   tableHtml += `<th class="px-4 py-3 bg-slate-100 font-semibold text-slate-700 whitespace-nowrap">${col}</th>`;
               } else {
-                  // Baris selanjutnya jadi Data (TD)
                   tableHtml += `<td class="px-4 py-3 align-top">${col}</td>`;
               }
           });
@@ -36,35 +34,27 @@ function parseMarkdown(text) {
       return tableHtml;
   });
 
-  // 2. Heading 3 (### Teks)
   html = html.replace(/^###\s+(.*$)/gim, '<h3 class="text-lg font-bold text-slate-800 mt-5 mb-2">$1</h3>');
   
-  // 3. Heading 2 (## Teks)
   html = html.replace(/^##\s+(.*$)/gim, '<h2 class="text-xl font-bold text-slate-800 mt-5 mb-2">$1</h2>');
 
-  // 4. Bold (**Teks**)
   html = html.replace(/\*\*(.*?)\*\*/g, '<strong class="font-bold text-slate-800">$1</strong>');
   
-  // 5. Bullet points (* Teks atau - Teks di awal baris)
   html = html.replace(/^[\*\-]\s+(.*$)/gim, '<div class="flex gap-2 mt-1.5"><span class="text-rose-500 font-bold shrink-0">•</span><span>$1</span></div>');
 
-  // 6. Italic (*Teks*)
   html = html.replace(/(?<!^)\*(.*?)\*/g, '<em class="italic text-slate-700">$1</em>');
 
-  // 7. Ubah Enter (\n) menjadi <br>
   html = html.replace(/\n/g, '<br/>');
   
-  // 8. Bersihkan sisa <br> berlebih agar spasi tidak terlalu jauh
   html = html.replace(/<\/h3><br\/>/g, '</h3>');
   html = html.replace(/<\/h2><br\/>/g, '</h2>');
   html = html.replace(/<\/div><br\/>/g, '</div>');
-  html = html.replace(/<\/div><br\/><br\/>/g, '</div>'); // Jaga jarak bawah tabel
+  html = html.replace(/<\/div><br\/><br\/>/g, '</div>'); 
   html = html.replace(/(<br\/>){3,}/g, '<br/><br/>');
 
   return html;
 }
 
-// --- 1. UPDATE FUNGSI NAVIGASI ---
 function switchMenu(menu) {
   const secDetect = document.getElementById("section-detect");
   const secChat = document.getElementById("section-chat");
@@ -91,7 +81,6 @@ function switchMenu(menu) {
   }
 }
 
-// --- FUNGSI DETEKSI PENYAKIT ---
 function handleFileChange(event) {
   const file = event.target.files[0];
   if (file) {
@@ -145,7 +134,6 @@ async function handleUpload() {
   }
 }
 
-// --- 2. UPDATE KOTAK HASIL DETEKSI ---
 function tampilkanHasilDeteksi(data) {
   const resultContainer = document.getElementById("detect-result");
   const resultImage = document.getElementById("result-image");
@@ -190,7 +178,6 @@ function tampilkanHasilDeteksi(data) {
   resultContainer.classList.remove("hidden");
 }
 
-// --- FUNGSI CHATBOT ---
 async function handleSendChat(event) {
   event.preventDefault();
 
@@ -231,7 +218,6 @@ async function handleSendChat(event) {
   }
 }
 
-// --- 3. UPDATE CHAT BUBBLE ---
 function appendMessage(role, content) {
   const chatBox = document.getElementById("chat-messages");
   const wrapper = document.createElement("div");

ingest.py

@@ -5,7 +5,6 @@ sys.path.append(root_dir)
 
 from src.ingestion.loader import load_data
 from src.ingestion.chunker import split_documents
-# Hapus import get_embedding_model karena sudah di-handle di dalam vector_store.py
 from src.retrieval.vector_store import get_vector_store
 
 SOURCES = [
@@ -31,10 +30,8 @@ def run_ingestion_pipeline():
     if all_chunks:
         print(f"\nMenyiapkan model embedding dan menyimpan {len(all_chunks)} chunks ke ChromaDB...")
         
-        # 1. Panggil database yang sudah ada (termasuk embedder-nya)
         db = get_vector_store()
         
-        # 2. Tambahkan chunks baru ke dalam koleksi 'chilicare_kb'
         db.add_documents(all_chunks)
         
         print("\nSelesai! Semua data web telah masuk ke database dan siap digunakan oleh API/Streamlit.")

src/chains/chain.py

@@ -41,21 +41,14 @@ def create_rag_chain(disease_label=None):
     prompt = get_rag_prompt()
     
     def format_docs(docs):
-        # ==========================================
-        # INTERCEPTOR: Print metadata ke Terminal
-        # ==========================================
-        print("\n" + "▼"*50)
         print("🔍 [DEBUG] DOKUMEN YANG DITARIK MULTI-QUERY RETRIEVER:")
         for i, doc in enumerate(docs):
             sumber = doc.metadata.get('label', 'Sumber tidak diketahui')
             print(f"  [{i+1}] Topik/Label: {sumber}")
-        print("▲"*50 + "\n")
-        # ==========================================
         
         # Gabungkan teks dokumen yang berhasil dikumpulkan dari semua query alternatif
         return "\n\n".join(doc.page_content for doc in docs)
         
-    # 6. Rangkai menjadi Chain (LCEL)
     rag_chain = (
         {"context": retriever | format_docs, "input": RunnablePassthrough()} 
         | prompt
@@ -65,7 +58,6 @@ def create_rag_chain(disease_label=None):
     
     return rag_chain
 
-# ======= Blok testing =========
 if __name__ == "__main__":
     chain = create_rag_chain()
     
@@ -75,7 +67,7 @@ if __name__ == "__main__":
     
     try:
         jawaban = chain.invoke(pertanyaan)
-        print("=== JAWABAN RAG ===")
+        print("JAWABAN RAG")
         print(jawaban)
     except Exception as e:
         print(f"Terjadi kesalahan: {e}")

src/ingestion/loader.py

@@ -18,7 +18,6 @@ def load_data(source: str):
         loader = TextLoader(source, encoding="utf-8")
         documents = loader.load()
         
-    # TAMBAHAN BARU UNTUK CSV
     elif source.lower().endswith(".csv"):
         if not os.path.exists(source): raise FileNotFoundError(f"File tidak ditemukan: {source}")
         loader = CSVLoader(source, encoding="utf-8") 

src/retrieval/vector_store.py

@@ -5,12 +5,10 @@ CHROMA_DB_DIR = "chroma_data"
 COLLECTION_NAME = "chilicare_kb"
 
 def get_embedder():
-    # Wajib sama persis dengan yang ada di db_setup.py
     return HuggingFaceEmbeddings(model_name="Qwen/Qwen3-Embedding-0.6B")
 
 def get_vector_store():
     embedder = get_embedder()
-    # Memuat database Chroma yang sudah dibuat oleh db_setup.py
     return Chroma(
         persist_directory=CHROMA_DB_DIR, 
         embedding_function=embedder,