app.py

import gradio as gr
import re
import inspect

from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim


codes = """001 - Vehicle Registration (New)
002 - Vehicle Registration Renewal
003 - Vehicle Ownership Transfer
004 - Vehicle De-registration
005 - Lost Registration Certificate Replacement
006 - Address Change Update
007 - Vehicle Data Correction
008 - Ownership Name Correction
009 - Vehicle Tax Payment
010 - Late Payment Fee Processing
011 - Vehicle Type/Specification Update
012 - BBNKB (Transfer Fee of Vehicle Ownership)
013 - STNK Issuance (Vehicle Registration Certificate)
014 - STNK Renewal
015 - Motor Vehicle Roadworthiness Inspection
016 - Plate Number Renewal
017 - Lost Plate Replacement
018 - Vehicle Export Registration
019 - Vehicle Import Registration
020 - Fleet Vehicle Registration
021 - Bulk Vehicle Registration Update
022 - Vehicle Insurance Assistance
023 - Vehicle Accident Reporting
024 - Vehicle Usage Change Declaration (e.g., personal to commercial)
025 - Legal Document Verification
026 - Ownership Transfer for Inherited Vehicle
027 - STNK Temporary Suspension
028 - Proof of Ownership Document Update
029 - Vehicle Ownership History Check
030 - Vehicle Tax Recalculation Request
031 - Tax Exemption Application (for special cases)
032 - Deceased Owner’s Vehicle Ownership Transfer""".split("\n")

undetected = "099 - Other/Undetected"


# codes = """001 - Pendaftaran Kendaraan (Baru)
# 002 - Pembaruan Pendaftaran Kendaraan
# 003 - Alih Kepemilikan Kendaraan
# 004 - Pembatalan Pendaftaran Kendaraan
# 005 - Penggantian Sertifikat Pendaftaran Kendaraan yang Hilang
# 006 - Pembaruan Perubahan Alamat
# 007 - Koreksi Data Kendaraan
# 008 - Koreksi Nama Kepemilikan
# 009 - Pembayaran Pajak Kendaraan
# 010 - Proses Denda Keterlambatan Pembayaran
# 011 - Pembaruan Jenis/Spesifikasi Kendaraan
# 012 - Pembayaran Pajak Kendaraan Melalui E-Samsat
# 013 - Penerbitan STNK (Sertifikat Pendaftaran Kendaraan)
# 014 - Pembaruan STNK
# 015 - Pemeriksaan Kelayakan Jalan Kendaraan Bermotor
# 016 - Pembaruan Nomor Plat Kendaraan
# 017 - Penggantian Plat yang Hilang
# 018 - Pendaftaran Ekspor Kendaraan
# 019 - Pendaftaran Impor Kendaraan
# 020 - Pendaftaran Kendaraan Armada
# 021 - Pembaruan Pendaftaran Kendaraan Massal
# 022 - Bantuan Asuransi Kendaraan
# 023 - Pelaporan Kecelakaan Kendaraan
# 024 - Deklarasi Perubahan Penggunaan Kendaraan (misalnya, pribadi ke komersial)
# 025 - Verifikasi Dokumen Hukum
# 026 - Alih Kepemilikan Kendaraan Warisan
# 027 - Penangguhan Sementara STNK
# 028 - Pembaruan Dokumen Bukti Kepemilikan
# 029 - Pemeriksaan Riwayat Kepemilikan Kendaraan
# 030 - Permintaan Perhitungan Ulang Pajak Kendaraan
# 031 - Permohonan Pembebasan Pajak (untuk kasus khusus)
# 032 - Alih Kepemilikan Kendaraan Pemilik yang Meninggal""".split("\n")

codes = """001 - Pendaftaran Kendaraan
002 - Pembaruan Data Kendaraan
003 - Alih Kepemilikan (Balik Nama)
004 - Pelaporan Dokumen atau Plat yang Hilang
005 - Pembayaran dan Pengelolaan Pajak Kendaraan
006 - Pemeriksaan dan Verifikasi Kendaraan
007 - Pendaftaran Kendaraan Ekspor, Impor, atau Armada
008 - Pelaporan dan Bantuan Terkait Kendaraan
009 - Penangguhan atau Deklarasi Perubahan Penggunaan Kendaraan""".split("\n")


vehicle_tax_info = {
    "B 1234 BCA": {
        "no_rangka": "1237191234",
        "type": "SUV",
        "tanggal": "23 Desember 2024",
        "status": "Belum Bayar",
        "harga_jual": 500_000_000  # In Rupiah
    },
    "B 5678 XYZ": {
        "no_rangka": "9876543210",
        "type": "Sedan",
        "tanggal": "15 Januari 2025",
        "status": "Belum Bayar",
        "harga_jual": 375_800_000  # In Rupiah
    },
    "D 3456 DEF": {
        "no_rangka": "4561237890",
        "type": "MPV",
        "tanggal": "10 Februari 2025",
        "status": "Sudah Bayar",
        "harga_jual": 400_000_000  # In Rupiah
    }
}

# Table for detail calculations (perhitungan)
detail_perhitungan = {
    "001": {
        "name": "Pendaftaran Kendaraan",
        "formula": lambda harga_jual: harga_jual * 0.1,
        # Example formula: 10% of harga_jual
    },
    "002": {
        "name": "Pembaruan Data Kendaraan",
        "formula": lambda harga_jual: harga_jual * 0.05,
        # Example formula: 5% of harga_jual
    },
    "003": {
        "name": "Alih Kepemilikan Kendaraan",
        "formula": lambda harga_jual: harga_jual * 0.1,  
        # Example formula: 10% of harga_jual
    },
    "004": {
        "name": "Penggantian Dokumen atau Plat yang Hilang",
        "formula": lambda harga_jual: harga_jual * 0.03,
        # Example formula: 3% of harga_jual
    },
    "005": {
        "name": "Pembayaran dan Pengelolaan Pajak Kendaraan",
        "formula": lambda harga_jual: harga_jual * 0.12,
        # Example formula: 12% of harga_jual
    },
    "006": {
        "name": "Pemeriksaan dan Verifikasi Kendaraan",
        "formula": lambda harga_jual: 100000,
        # Example formula: 2% of harga_jual
    },
    "007": {
        "name": "Pendaftaran Kendaraan Ekspor, Impor, atau Armada",
        "formula": lambda harga_jual: harga_jual * 0.15,
        # Example formula: 15% of harga_jual
    },
    "008": {
        "name": "Pelaporan dan Bantuan Terkait Kendaraan",
        "formula": lambda harga_jual: harga_jual * 0.04,
        # Example formula: 4% of harga_jual
    },
    "009": {
        "name": "Penangguhan atau Deklarasi Perubahan Penggunaan Kendaraan",
        "formula": lambda harga_jual: harga_jual * 0.06,
        # Example formula: 6% of harga_jual
    }
}

undetected = "099 - Lainnya/Tidak Terdeteksi"

model_ids = [
    "BAAI/bge-m3",
    "sentence-transformers/paraphrase-multilingual-mpnet-base-v2",
    "intfloat/multilingual-e5-small",
    "sentence-transformers/distiluse-base-multilingual-cased-v2",
    "Alibaba-NLP/gte-multilingual-base",
    "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
]
# model_id = "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2"
# model_id = "Alibaba-NLP/gte-multilingual-base"
# model_id = "BAAI/bge-m3"
# model_id = "sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
# model_id = "intfloat/multilingual-e5-small"
# model_id = "sentence-transformers/distiluse-base-multilingual-cased-v2"
model_id = model_ids[-1]
model = SentenceTransformer(model_id, trust_remote_code=True)

codes_emb = model.encode([x[6:] for x in codes])


def censor_middle(number, num_to_hide=4):
  number_str = str(number)
  middle_index = len(number_str) // 2
  start_index = middle_index - num_to_hide // 2
  end_index = middle_index + num_to_hide // 2

  censored_part = "*" * num_to_hide
  censored_number = number_str[:start_index] + censored_part + number_str[end_index:]

  return censored_number


def get_calculation(request_code, plate_number):
    print(request_code, plate_number, "GET CALC")
    calc = detail_perhitungan.get(request_code)
    vehicle = vehicle_tax_info.get(plate_number)
    
    if vehicle != None and calc != None:
        harga_jual = vehicle.get("harga_jual")
        formula = calc.get("formula")
        result = formula(harga_jual)
        description = inspect.getsource(formula).split(":", 2)[-1].strip()
        result_detail = request_code + " - " + calc.get("name")
        return result, str(description), result_detail

    elif calc != None:
        formula = calc.get("formula")
        description = inspect.getsource(formula).split(":", 2)[-1].strip()
        result_detail = request_code + " - " + calc.get("name")
        return None, str(description), result_detail

    else:
        return None, None, None


def build_output_formula(descriptions, result_details):
    out = "----------------------------------------------------\n\n"
    out = "Daftar Kode Permohonan:\n"
    for i, (desc,detail) in enumerate(zip(descriptions, result_details)):
        # harga_jual = str(vehicle.get("harga_jual"))
        out += f"{i+1}. {detail}\nRumus: {desc}\n"
    return out


def build_output_vehicle(plate_number):
    vehicle = vehicle_tax_info.get(plate_number)
    out = "----------------------------------------------------\n\n"
    out = "Nomor Polisi: " + plate_number + "\n"
    vehicle["no_rangka"] = censor_middle(vehicle["no_rangka"])
    vehicle["harga_jual"] = "{:,}".format(vehicle["harga_jual"])
    out += "\n".join([k + " : " + str(v) for  k,v in vehicle.items()])
    return out

def build_output(result, description, result_detail, plate_number):
    return build_outputs([result], [description], [result_detail], plate_number)


def build_outputs(results, descriptions, result_details, plate_number):
    vehicle = vehicle_tax_info.get(plate_number)
    vehicle["harga_jual"] = "{:,}".format(vehicle["harga_jual"])

    out = "----------------------------------------------------\n\n"
    out = "Nomor Polisi: " + plate_number + "\n"
    out += "\n".join([k + " : " + str(v) if k != "no_rangka" else k + " : " + censor_middle(v) for  k,v in vehicle.items()])
    # out += "\n----------------------------------------------"
    
    # out += f"\nWajib Pajak dengan NoPol {plate_number} ingin melakukan proses berikut:\n"
    out += "\n\nDaftar Kode Permohonan:\n"

    for i, (res,desc,detail) in enumerate(zip(results, descriptions, result_details)):
        harga_jual = vehicle["harga_jual"]
        out += f"{i+1}. {detail}\nRumus: {desc}\nDetail perhitungan: {desc.replace('harga_jual', harga_jual)} = {res}\n"

    # out += "----------------------------\nEstimasi biaya: "
    out += "\n\n\nEstimasi Biaya: "

    if len(results) > 1:
        out += " + ".join([f"Rp{x}" for x in results])
        out += f" = {sum(results)}"
    else:
        out += str(results[0])

    out += "\n\n----------------------------"
    # out += "\n----------------------------\n\n--------------------------------"
    return out
    

def respond(
    message,
    history: list[tuple[str, str]],
    threshold,
    is_multiple
):
    global codes_emb
    global undetected

    undetected_code = undetected[:3]

    # if history and history[-1][-1][21:24] == undetected_code: 
    #     list_his = ""
    #     for his in history[::-1]:
    #         if his[-1][21:24] != undetected_code:
    #             break
    #         list_his = his[0] + "\n" + list_his

    #     message += "\n" + list_his
    
    # pattern = r'\b([A-Z]{1,2})\s?(\d{4})\s?([A-Z]{3})\b'
    # pattern = r'\b([A-Z]{1,2})\s?(\d{4})\s?([A-Z]{1,3})\b'
    pattern = r'\b([A-Za-z]{1,2})\s?(\d{4})\s?([A-Za-z]{1,3})\b'

    matches = re.findall(pattern, message)

    plates = [" ".join(x).upper() for i,x in enumerate(matches)]

    plate_numbers = ", ".join(plates)
    
    text_emb = model.encode(message)
    scores = cos_sim(codes_emb, text_emb)[:,0]

    if is_multiple:
        request_details = []
        request_numbers = []
        request_scores = []
        # request_undetected = False
        for i,score in enumerate(scores):
            if score > threshold:
                request_details.append(codes[i][6:])
                request_numbers.append(codes[i][:3])
                request_scores.append(str( round(score.tolist(), 3) ) )

        if not request_details:
            request_details.append(undetected[6:])
            request_numbers.append(undetected_code)
            

        request_numbers_copy = request_numbers
        request_details_copy = request_details

        request_numbers = "\n".join(request_numbers)
        request_details = "\n".join(request_details)
        request_scores = "\n".join(request_scores)

        # if len(request_numbers_copy) > 0:
        #     for code, detail in zip(request_numbers_copy, request_details_copy):
        #         kode_mohon = detail_perhitungan.get(code)
        #         formula = kode_mohon.get("formula")
        #         description = inspect.getsource(formula).split(":", 2)[-1].strip()

        # if request_undetected and len(plates) == 0:
            # + f"\n\nConfidence score:\n{request_scores}"
            # return "Request code number:\n" + request_numbers + "\n\nRequest detail:\n" + request_details + "\n\nPlate numbers: " + plate_numbers
        
        out = ""
        for plate in plates:
            results, descriptions, result_details = [], [], []
            for code in request_numbers_copy:
                result, description, result_detail = get_calculation(code, plate)

                if result != None:
                    results.append(result)
                if descriptions != None:
                    descriptions.append(description)
                    result_details.append(result_detail)

            if results:
                out += "\n\n" + build_outputs(results, descriptions, result_details, plate)
            elif vehicle_tax_info.get(plate):
                out += "\n\n" + build_output_vehicle(plate)

        if out == "":
            descriptions, result_details = [], []
            for code in request_numbers_copy:
                result, description, result_detail = get_calculation(code, "")
                if description != None:
                    descriptions.append(description)
                    result_details.append(result_detail)
            if descriptions:
                out += "\n\n" + build_output_formula(descriptions, result_details)
            else:
                return "Request code number: " + request_numbers + "\nRequest detail: " + request_details + "\nPlate numbers: " + plate_numbers

        return out

        # result, description, result_detailget_calculation(request_code, plate_number)



    s_max = scores.argmax()

    if scores[s_max] < threshold:
        # request_code = "033 - Other/Undetected"
        request_code = undetected
    else:
        request_code = codes[scores.argmax()]
        # "{:.2f}".format(a)

    # out = "Request code number: " + request_code[:3] + "\nRequest detail: " + request_code[6:] + f"\nConfidence score: {round(scores[s_max].tolist(),3)}" + "\nPlate numbers: " + plate_numbers

    out = ""
    for plate in plates:
        results, descriptions, result_details = [], [], []
        result, description, result_detail = get_calculation(request_code[:3], plate)

        if result != None:
            results.append(result)
            out += "\n\n" + build_outputs(results, descriptions, result_details, plate)
        elif vehicle_tax_info.get(plate):
            out += "\n\n" + build_output_vehicle(plate)
        # else:
        #     # + f"\nConfidence score: {round(scores[s_max].tolist(),3)}"
        #     out += "Request code number: " + request_code[:3] + "\nRequest detail: " + "\nPlate numbers: " + plate_numbers

    if out == "":
        if request_code[:3] == undetected_code:
            return "Request code number: " + request_code[:3] + "\nRequest detail: " + request_code[6:] + "\nPlate numbers: " + plate_numbers
        else:
            result, description, result_detail = get_calculation(request_code[:3], "")
            out = build_output_formula([description], [result_detail])

    return out
    
    # if vehicle_tax_info.get(request_detail)
    
    # for val in history:
    #     if val[0]:
    #         messages.append({"role": "user", "content": val[0]})
    #     if val[1]:
    #         messages.append({"role": "assistant", "content": val[1]})

    # messages.append({"role": "user", "content": message})

    # response = ""

    # for message in client.chat_completion(
    #     messages,
    #     max_tokens=max_tokens,
    #     stream=True,
    #     temperature=temperature,
    #     top_p=top_p,
    # ):
    #     token = message.choices[0].delta.content

    #     response += token
    #     yield response


"""
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
"""
# demo = gr.ChatInterface(
#     respond,
# )


def reload(chosen_model_id):
    global model
    global model_id
    global codes_emb

    if chosen_model_id != model_id:
        model = SentenceTransformer(chosen_model_id, trust_remote_code=True)
        model_id = chosen_model_id
        codes_emb = model.encode([x[6:] for x in codes])
        return f"Model {chosen_model_id} has been succesfully loaded!"
    return f"Model {chosen_model_id} is ready!"
    


with gr.Blocks() as demo:
    # Add header title and description
    gr.Markdown("# List of Request Numbers")
    gr.Markdown("<br>".join(codes) + "<br>" + undetected)
    gr.Markdown("# Valid License Plate Number Criteria:")
    gr.Markdown("(1-2 letters) (4 numbers) (1-3 letters)")
    gr.Markdown("# Example Valid Plate Numbers")
    gr.Markdown("<br>".join(vehicle_tax_info.keys()))


    gr.Markdown("# Choose & Load Model:")
    reload_model = gr.Interface(
        fn=reload, 
        inputs=[gr.Dropdown(choices=model_ids, value=model_id)], 
        outputs="text",
        # gr.HighlightedText(
        #     label="status",
        #     combine_adjacent=True,
        #     show_legend=True,
        #     color_map={"+": "red", "-": "green"}
        # ),
    )

    gr.Markdown("# Chatbot Interface:")
    chat_interface = gr.ChatInterface(
        respond, 
        additional_inputs=[
            gr.Number(0.5, label="confidence threshold", show_label=True, minimum=0., maximum=1.0, step=0.1),
            gr.Checkbox(label="multiple", info="Allow multiple request code numbers"),
            
        ],
        # type="messages",
        chatbot=gr.Chatbot(height=800),
        # textbox=gr.Textbox(placeholder="Ask me a yes or no question", container=False, scale=7),
        # title="SamSat Virtual Assistant",
        # description="Ask Yes Man any question",
        # theme="soft",
        # examples=["balik nama D 3456 DEF", "bayar pajak B 1234 BCA", "halo, selamat pagi!"],
        # cache_examples=True,
    )

if __name__ == "__main__":
    demo.launch()