app.py

import gradio as gr
import jieba
import jieba.analyse
import aiohttp
import asyncio
import ssl
from aiohttp import ClientSession

ssl_context = ssl.create_default_context()
ssl_context.check_hostname = False
ssl_context.verify_mode = ssl.CERT_NONE

def setup_jieba_dictionaries():
    dictionaries = [
        'flag/RVT_AddressCh.txt',
        'flag/RVT_AddressEn.txt',
        'flag/RVT_Area.txt',
        'flag/RVT_BuildingCh.txt',
        'flag/RVT_BuildingEn.txt'
    ]
    for file_path in dictionaries:
        jieba.load_userdict(file_path)

def process_text(text):
    setup_jieba_dictionaries()

    user_dict_terms = set()
    for file_path in [
        'flag/RVT_AddressCh.txt',
        'flag/RVT_AddressEn.txt',
        'flag/RVT_Area.txt',
        'flag/RVT_BuildingCh.txt',
        'flag/RVT_BuildingEn.txt'
    ]:
        try:
            with open(file_path, 'r', encoding='utf-8') as f:
                user_dict_terms.update(line.strip().split()[0] for line in f)
        except FileNotFoundError:
            print(f'File not found: {file_path}')
        except Exception as e:
            print(f'Error reading file {file_path}: {e}')

    lines = text.splitlines()
    results = []

    for line in lines:
        line = line.strip()
        keywords = jieba.analyse.textrank(line, topK=20, withWeight=False, allowPOS=('ns', 'n', 'vn', 'v'))
        keyword_text = ' '.join(keyword for keyword in keywords if keyword in user_dict_terms)
        results.append(keyword_text)

    return results

def reformat_text(text):
    lines = text.splitlines()
    return [line.strip() for line in lines if line.strip()]

def process_text_only(text, reformat):
    extracted_keywords = process_text(text)
    if reformat:
        extracted_keywords = reformat_text('\n'.join(extracted_keywords))
    # Join keywords with newline characters
    return '\n'.join(extracted_keywords)

async def lookup_address(query, language='zh-Hant'):
    url = 'https://www.als.gov.hk/lookup'
    headers = {
        'Accept': 'application/json',
        'Accept-Language': language
    }
    payload = {
        'q': query
    }

    async with ClientSession() as session:
        try:
            async with session.post(url, headers=headers, data=payload, ssl=ssl_context) as response:
                if response.status == 200:
                    return await response.json()
                else:
                    print(f'Error fetching data: Status Code {response.status}')
                    return {'error': f'Error fetching data: Status Code {response.status}'}
        except aiohttp.ClientError as e:
            print(f'Client Error: {e}')
            return {'error': f'Client Error: {e}'}
        except Exception as e:
            print(f'General Error: {e}')
            return {'error': f'General Error: {e}'}

async def get_address_lookup_results(keywords):
    results = []
    for keyword in keywords:
        keyword = keyword.strip()
        if not keyword:  # Skip empty keywords
            continue
        lookup_results = await lookup_address(keyword)
        if 'SuggestedAddress' in lookup_results and isinstance(lookup_results['SuggestedAddress'], list):
            first_match = lookup_results['SuggestedAddress'][0]  # Use the first match
            full_address = 'No matches found'
            geo_address = 'N/A'
            latitude = 'N/A'
            longitude = 'N/A'
            matched_building = 'No Building Name'

            if first_match:
                premises_address = first_match['Address']['PremisesAddress']
                raw_address = premises_address.get('ChiPremisesAddress', {})
                matched_building = raw_address.get('BuildingName', 'No Building Name')
                full_address = matched_building
                geo_address = premises_address.get('GeoAddress', 'N/A')
                geo_info = premises_address.get('GeospatialInformation', {})
                latitude = geo_info.get('Latitude', 'N/A')
                longitude = geo_info.get('Longitude', 'N/A')

            results.append({
                'Keyword': keyword,
                'Full Address': full_address,
                'Geo Address': geo_address,
                'Latitude': latitude,
                'Longitude': longitude
            })
        else:
            results.append({
                'Keyword': keyword,
                'Full Address': 'No matches found',
                'Geo Address': 'N/A',
                'Latitude': 'N/A',
                'Longitude': 'N/A'
            })

    return results

async def gradio_function(text, reformat, perform_lookup):
    extracted_keywords = process_text_only(text, reformat)
    keywords_list = extracted_keywords.splitlines()

    address_results = []
    if perform_lookup:
        address_results = await get_address_lookup_results(keywords_list)

    return extracted_keywords, address_results

def gradio_interface(text, reformat, perform_lookup):
    return asyncio.run(gradio_function(text, reformat, perform_lookup))

interface = gr.Interface(
    fn=gradio_interface,
    inputs=[
        gr.Textbox(lines=20, placeholder="Paste text here, each line will be processed separately..."),
        gr.Checkbox(label="Reformat text (remove empty lines)"),
        gr.Checkbox(label="Perform Address Lookup")
    ],
    outputs=[
        gr.Textbox(label="Extracted Address Keywords"),
        gr.JSON(label="Address Lookup Results")
    ],
    title="Address Extraction and Lookup with Natural Language Processing",
    description="Extract address keywords using NLP and optionally perform address lookup using ALS."
)

interface.launch()