Alibaba Cloud Model Studio：Penelitian Mendalam (Qwen-Deep-Research)

import os
import dashscope

# Konfigurasikan API key
# Jika Anda belum mengonfigurasi variabel lingkungan, ganti baris berikut dengan Model Studio API key Anda: API_KEY = "sk-xxx"
API_KEY = os.getenv('DASHSCOPE_API_KEY')

def call_deep_research_model(messages, step_name):
    print(f"\n=== {step_name} ===")
    
    try:
        responses = dashscope.Generation.call(
            api_key=API_KEY,
            model="qwen-deep-research",
            messages=messages,
            # Model qwen-deep-research saat ini hanya mendukung streaming output
            stream=True
            # Untuk menggunakan output inkremental, tambahkan parameter incremental_output=True
        )
        
        return process_responses(responses, step_name)
        
    except Exception as e:
        print(f"Terjadi error saat memanggil API: {e}")
        return ""


# Menampilkan konten fase
def display_phase_content(phase, content, status):
    if content:
        print(f"\n[{phase}] {status}: {content}")
    else:
        print(f"\n[{phase}] {status}")

# Memproses respons
def process_responses(responses, step_name):
    current_phase = None
    phase_content = ""
    research_goal = ""
    web_sites = []
    references = []
    keepalive_shown = False  # Bendera untuk memeriksa apakah prompt KeepAlive telah ditampilkan

    for response in responses:
        # Periksa kode status respons
        if hasattr(response, 'status_code') and response.status_code != 200:
            print(f"Kode balasan HTTP: {response.status_code}")
            if hasattr(response, 'code'):
                print(f"Kode error: {response.code}")
            if hasattr(response, 'message'):
                print(f"Pesan error: {response.message}")
            print("Untuk informasi lebih lanjut, lihat https://www.alibabacloud.com/help/en/model-studio/error-code")
            continue

        if hasattr(response, 'output') and response.output:
            message = response.output.get('message', {})
            phase = message.get('phase')
            content = message.get('content', '')
            status = message.get('status')
            extra = message.get('extra', {})

            # Mendeteksi perubahan fase
            if phase != current_phase:
                if current_phase and phase_content:
                    # Menampilkan deskripsi penyelesaian berbeda berdasarkan nama fase dan nama langkah
                    if step_name == "Step 1: Model follow-up question for confirmation" and current_phase == "answer":
                        print(f"\n Fase pertanyaan lanjutan model selesai")
                    else:
                        print(f"\n Fase {current_phase} selesai")
                current_phase = phase
                phase_content = ""
                keepalive_shown = False  # Mengatur ulang bendera prompt KeepAlive

                # Menampilkan deskripsi berbeda berdasarkan nama fase dan nama langkah
                if step_name == "Step 1: Model follow-up question for confirmation" and phase == "answer":
                    print(f"\n Memasuki fase pertanyaan lanjutan model")
                else:
                    print(f"\n Memasuki fase {phase}")
                    
            # Memproses informasi referensi untuk fase Answer
            if phase == "answer":
                if extra.get('deep_research', {}).get('references'):
                    new_references = extra['deep_research']['references']
                    if new_references and new_references != references:  # Hindari tampilan duplikat
                        references = new_references
                        print(f"\n   Referensi ({len(references)}):")
                        for i, ref in enumerate(references, 1):
                            print(f"     {i}. {ref.get('title', 'Tidak ada judul')}")
                            if ref.get('url'):
                                print(f"        URL: {ref['url']}")
                            if ref.get('description'):
                                print(f"        Deskripsi: {ref['description'][:100]}...")
                            print()

            # Memproses informasi khusus untuk fase WebResearch
            if phase == "WebResearch":
                if extra.get('deep_research', {}).get('research'):
                    research_info = extra['deep_research']['research']

                    # Memproses status streamingQueries
                    if status == "streamingQueries":
                        if 'researchGoal' in research_info:
                            goal = research_info['researchGoal']
                            if goal:
                                research_goal += goal
                                print(f"\n   Tujuan penelitian: {goal}", end='', flush=True)

                    # Memproses status streamingWebResult
                    elif status == "streamingWebResult":
                        if 'webSites' in research_info:
                            sites = research_info['webSites']
                            if sites and sites != web_sites:  # Hindari tampilan duplikat
                                web_sites = sites
                                print(f"\n   Ditemukan {len(sites)} situs web relevan:")
                                for i, site in enumerate(sites, 1):
                                    print(f"     {i}. {site.get('title', 'Tidak ada judul')}")
                                    print(f"        Deskripsi: {site.get('description', 'Tidak ada deskripsi')[:100]}...")
                                    print(f"        URL: {site.get('url', 'Tidak ada tautan')}")
                                    if site.get('favicon'):
                                        print(f"        Ikon: {site['favicon']}")
                                    print()

                    # Memproses status WebResultFinished
                    elif status == "WebResultFinished":
                        print(f"\n   Pencarian web selesai. Ditemukan {len(web_sites)} sumber referensi.")
                        if research_goal:
                            print(f"   Tujuan penelitian: {research_goal}")

            # Mengumpulkan dan menampilkan konten
            if content:
                phase_content += content
                # Menampilkan konten secara real-time
                print(content, end='', flush=True)

            # Menampilkan perubahan status fase
            if status and status != "typing":
                print(f"\n   Status: {status}")

                # Menampilkan deskripsi status
                if status == "streamingQueries":
                    print("   → Menghasilkan tujuan penelitian dan kueri pencarian (fase WebResearch)")
                elif status == "streamingWebResult":
                    print("   → Melakukan pencarian, membaca halaman web, dan eksekusi kode (fase WebResearch)")
                elif status == "WebResultFinished":
                    print("   → Fase pencarian web selesai (fase WebResearch)")

            # Saat status adalah 'finished', tampilkan penggunaan token
            if status == "finished":
                if hasattr(response, 'usage') and response.usage:
                    usage = response.usage
                    print(f"\n    Statistik penggunaan token:")
                    print(f"      Token input: {usage.get('input_tokens', 0)}")
                    print(f"      Token output: {usage.get('output_tokens', 0)}")
                    print(f"      ID permintaan: {response.get('request_id', 'Tidak diketahui')}")

            if phase == "KeepAlive":
                # Tampilkan prompt hanya saat pertama kali memasuki fase KeepAlive
                if not keepalive_shown:
                    print("Langkah saat ini selesai. Mempersiapkan langkah berikutnya.")
                    keepalive_shown = True
                continue

    if current_phase and phase_content:
        if step_name == "Step 1: Model follow-up question for confirmation" and current_phase == "answer":
            print(f"\n Fase pertanyaan lanjutan model selesai")
        else:
            print(f"\n Fase {current_phase} selesai")

    return phase_content

def main():
    # Periksa API key
    if not API_KEY:
        print("Error: Variabel lingkungan DASHSCOPE_API_KEY belum diatur.")
        print("Atur variabel lingkungan atau modifikasi langsung variabel API_KEY dalam kode.")
        return
    
    print("Pengguna memulai percakapan: Teliti penerapan artificial intelligence dalam pendidikan")
    
    # Step 1: Model follow-up question for confirmation
    # Model menganalisis pertanyaan pengguna dan mengajukan pertanyaan lanjutan untuk mengklarifikasi arah penelitian.
    messages = [{'role': 'user', 'content': 'Teliti penerapan artificial intelligence dalam pendidikan'}]
    step1_content = call_deep_research_model(messages, "Step 1: Model follow-up question for confirmation")

    # Step 2: Deep research
    # Berdasarkan konten pertanyaan lanjutan pada Langkah 1, model menjalankan proses penelitian lengkap.
    messages = [
        {'role': 'user', 'content': 'Teliti penerapan artificial intelligence dalam pendidikan'},
        {'role': 'assistant', 'content': step1_content},  # Berisi pertanyaan lanjutan model
        {'role': 'user', 'content': 'Saya terutama tertarik pada pembelajaran personalisasi dan penilaian cerdas.'}
    ]
    
    call_deep_research_model(messages, "Step 2: Deep research")
    print("\n Penelitian selesai!")

if __name__ == "__main__":
    main()

echo "Step 1: Model follow-up question for confirmation"
curl --location 'https://dashscope.aliyuncs.com/api/v1/services/aigc/text-generation/generation' \
--header 'X-DashScope-SSE: enable' \
--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
--header 'Content-Type: application/json' \
--data '{
    "input": {
        "messages": [
            {
                "content": "Research the applications of artificial intelligence in education", 
                "role": "user"
            }
        ]
    },
    "model": "qwen-deep-research"
}'

echo -e "\n\n" 
echo "Step 2: Deep research"
curl --location 'https://dashscope.aliyuncs.com/api/v1/services/aigc/text-generation/generation' \
--header 'X-DashScope-SSE: enable' \
--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
--header 'Content-Type: application/json' \
--data '{
    "input": {
        "messages": [
            {
                "content": "Research the applications of artificial intelligence in education", 
                "role": "user"
            },
            {
                "content": "Which specific application scenarios of artificial intelligence in education would you like to focus on?", 
                "role": "assistant"
            },
            {
                "content": "I am mainly interested in personalized learning.", 
                "role": "user"
            }
        ]
    },
    "model": "qwen-deep-research"
}'