Update app.py

app.py

@@ -2,28 +2,14 @@ import streamlit as st
 from transformers import pipeline
 import nltk
 from nltk.tokenize import word_tokenize
-from nltk.corpus import stopwords
-
-# Download data NLTK
-nltk.download('punkt')
-nltk.download('stopwords')
 
 # Buat objek terjemahan
 translator = pipeline("translation", model="Helsinki-NLP/opus-mt-id-en")
 terjemah = pipeline("translation", model="Helsinki-NLP/opus-mt-en-id")
-pipe1 = pipeline("text-generation", model="TinyLlama/TinyLlama-1.1B-Chat-v1.0")
-#pipe2 = pipeline("text-generation", model="SumayyaAli/tiny-llama-1.1b-chat-medical")
-
-# Pra-pemrosesan teks menggunakan NLTK
-def pra_pemrosesan_teks(teks):
-    tokens = word_tokenize(teks)
-    tokens = [token for token in tokens if token.isalnum()]  # Menghapus kata bukan alfanumerik
-    tokens = [token for token in tokens if token.lower() not in stopwords.words('indonesian')]  # Menghapus stopword
-    teks_diproses = " ".join(tokens)
-    return teks_diproses
+pipe = pipeline("text-generation", model="TinyLlama/TinyLlama-1.1B-Chat-v1.0")
 
 # Antarmuka Streamlit
-st.title("Diagnosa Penyakit Berdasarkan Gejala")
+st.title("Diagnosa Berdasarkan Gejala Kehamilan")
 
 # Masukkan gejala, usia, dan jenis kelamin
 usia = st.number_input("Masukkan usia Anda:", min_value=0, max_value=120)
@@ -32,54 +18,37 @@ gejala_id = st.text_area("Masukkan gejala Anda:")
 
 if st.button("Diagnosa"):
     if gejala_id:
-        # Terjemahkan gejala dari Indonesia ke Inggris
-        terjemahan = translator(gejala_id, max_length=100)
-        gejala_en = terjemahan[0]["translation_text"]
+        # Terjemahkan gejala dari Bahasa Indonesia ke Bahasa Inggris
+        gejala_en = translator(gejala_id, max_length=100)[0]["translation_text"]
+        informasi_pasien = f"I am {usia} years old, {jenis_kelamin}. Current symptoms are: {gejala_en}"
 
-        informasi_pasien = f"I am {usia} years old, {jenis_kelamin}. And my current symptoms are {gejala_en}"
-
-        # Pesan untuk model
+        # Contoh kasus kehamilan untuk diagnosis
         pesan = [
             {"role": "system",
-     "content": """
-You are a doctor diagnosing patients based on symptoms.
+             "content": """
+You are a doctor diagnosing pregnancy-related conditions based on symptoms.
 
 Example 1:
-Patient symptoms: High fever, headache, nausea
-Diagnosis: Dengue fever
+Patient symptoms: Missed period, nausea, tender breasts
+Diagnosis: Possible pregnancy
 
 Example 2:
-Patient symptoms: Shortness of breath, chest pain, cough with phlegm
-Diagnosis: Pneumonia.
-
-
+Patient symptoms: Severe abdominal pain, bleeding
+Diagnosis: Possible miscarriage or ectopic pregnancy.
 """},
-            {"role": "user", "content": f"Based on your assessment, {informasi_pasien}, what illness could it be?"}
+            {"role": "user", "content": f"Based on your assessment, {informasi_pasien}, what is your diagnosis?"}
         ]
 
         # Fungsi untuk mendapatkan konten dari role 'assistant'
         def get_assistant_content(response):
-            generated_text = response [0]['generated_text']
-            return generated_text
-
-        # Dapatkan respon dari pipe1 (model TinyLlama)
-        response1 = pipe1(pesan, num_return_sequences=1, truncation=True)
-        asisten_konten1 = get_assistant_content(response1)
-
-        # Dapatkan respon dari pipe2 (model SumayyaAli)
-        #response2 = pipe2(pesan, num_return_sequences=1, truncation=True)
-        #asisten_konten2 = get_assistant_content(response2)
-
-        # Gabungkan hasil dari pipe1 dan pipe2 untuk pertanyaan akhir
-        #pertanyaan_akhir = [{'role': 'user', 'content': f"{asisten_konten1}. {asisten_konten2}. Based on these two sentences, what is your final conclusion of my current symptom? Please provide a brief answer with one diagnosis."}]
+            return response[0]['generated_text']
 
-        # Dapatkan hasil akhir diagnosis
-        #hasil_diagnosis = pipe1(pertanyaan_akhir)
-        #asisten_konten3 = get_assistant_content(hasil_diagnosis)
+        # Dapatkan respon dari pipe
+        response = pipe(pesan, num_return_sequences=1, truncation=True)
+        diagnosis = get_assistant_content(response)
 
-        # Terjemahkan hasil akhir ke bahasa Indonesia
-        terjemahan_hasil = terjemah(asisten_konten1, max_length=100)
-        diagnosa_terjemahan = terjemahan_hasil[0]["translation_text"]
+        # Terjemahkan hasil ke Bahasa Indonesia
+        diagnosa_terjemahan = terjemah(diagnosis, max_length=100)[0]["translation_text"]
 
         # Tampilkan hasil ke Streamlit
         st.subheader("Hasil Diagnosis:")