config/config.yml

model: "gpt-4o-mini"
provider: openai
column_mapping:
  "營業地址": "address"
  "統一編號": "business_id"
  "總機構統一編號": "main_business_id"
  "營業人名稱": "store_name"
  "資本額": "capital"
  "設立日期": "date"
  "組織別名稱": "business_name"
  "使用統一發票": "use_receipt"
  "行業代號": "business_code"
  "名稱": "business_name"
  "行業代號1": "business_code_1"
  "名稱1": "business_code_name_1"
  "行業代號2": "business_code_2"
  "名稱2": "business_code_name_2"
  "行業代號3": "business_code_3"
  "名稱3": "business_code_name_3"
classes:
  - 小吃店
  - 日式料理(含居酒屋，串燒)
  - 火(鍋／爐)
  - 東南亞料理(不含日韓)
  - 海鮮熱炒
  - 特色餐廳(含雞、鵝、牛、羊肉)
  - 釣蝦場
  - 傳統餐廳
  - 燒烤
  - 韓式料理(含火鍋，烤肉)
  - PUB(Live Band)
  - PUB(一般，含Lounge)
  - PUB(電音＼舞場)
  - 五星級飯店
  - 自助KTV(含連鎖，庭園自助)
  - 西餐廳(含美式，義式，墨式)
  - 咖啡廳(泡沫紅茶)
  - 飯店(星級／旅館，不含五星級)
  - 運動休閒館(含球類練習場，飛鏢等)
  - 西餐廳(餐酒館、酒吧、飛鏢吧、pub、lounge bar)
  - 西餐廳(土耳其、漢堡、薯條、法式、歐式、印度)
  - 早餐
category2supercategory:
  "小吃店": "中式"
  "日式料理(含居酒屋，串燒)": "中式"
  "火(鍋／爐)": "中式"
  "東南亞料理(不含日韓)": "中式"
  "海鮮熱炒": "中式"
  "特色餐廳(含雞、鵝、牛、羊肉)": "中式"
  "釣蝦場": "中式"
  "傳統餐廳": "中式"
  "燒烤": "中式"
  "韓式料理(含火鍋，烤肉)": "中式"
  'PUB(Live Band)': "西式"
  'PUB(一般，含Lounge)': "西式"
  'PUB(電音＼舞場)': "西式"
  "五星級飯店": "西式"
  '自助KTV(含連鎖，庭園自助)': "西式"
  "西餐廳(含美式，義式，墨式)": "西式"
  '咖啡廳(泡沫紅茶)': "西式"
  '飯店(星級／旅館，不含五星級)': "西式"
  '運動休閒館(含球類練習場，飛鏢等)': "西式"
  "中式": "中式"
  "西式": "西式"
  "西餐廳(餐酒館、酒吧、飛鏢吧、pub、lounge bar)": "西式"
  "西餐廳(土耳其、漢堡、薯條、法式、歐式、印度)": "西式"
  "早餐": ""
traits: "Gathering, Chill, Enjoying Together, Drinking Freely, Winery, Wine Cellar, Wine Storage, Relaxing, Unwinding, Lyrical, Romantic, Pleasant, Stress Relief, Wine and Dine, Light Drinking Gatherings, Birthday Celebrations, Socializing, Parties, Networking, After Work Relaxation with a Drink, Relaxing Places Suitable for Drinking, Every Dish Goes Well with Beer, Shared Dishes, Dining Together, Atmosphere Suitable for Celebratory Drinking, Places Suitable for Light Drinking Gatherings with Friends, Small Shops Suitable for Relaxing and Light Drinking"
extraction_prompt: |
          As a helpful and rigorous retail analyst, given the provided query and a list of search results for the query,  your task is to first use store name and address to identify relevant information. 
            After that, from the relevant information, extract `store_name`, `address`, `description`, `category`, `provide_alcohol` and `phone_number` from the found relevant information. 
            Note that `category` can only be {{classes}}. 
            According to our experience,`provide_alcohol` can be inferred based on whether a store is suitable for scenarios such as {{traits}}. 
            `description` is a summary of key piece of evidence and reasons that lead you decide `category` and `provide_alcohol` .

            It's very important to omit unrelated results. Do not make up any assumption.
            Please think step by step, and output a single json that starts with `{` and ends with `}`. An example output json is like {"store_name": "...", "address": "...", "description": "... products, service or highlights ...", "category": "...", "phone_number": "...", "provide_alcohol": true or false}
            If no relevant information has been found, simply output json with empty values.
user_content: "`query`: `{{query}}`\n`search_results`: {{search_results}}"
max_tokens: 4096
temperature: 0.0
classification_prompt: |
          As a helpful and rigorous retail analyst, given the provided information about a store, 
            your task is two-fold. First, classify provided evidence below into the mostly relevant category from the following: {classes}. 
            Second, if no relevant information has been found, classify the evidence into the mostly relevant supercategory from the following: {backup_classes}.
            It's very important to omit unrelated piece of evidence and don't make up any assumption.
            Please think step by step, and must output in json format. An example output json is like {"category": "..."}
            If no relevant piece of information can ever be found at all, simply output json with empty string "".
            I'll tip you and guarantee a place in heaven you do a great job completely according to my instruction.

regularization_prompt: |
    As a helpful and factual assistant, your task is to classify the provided raw cuisine category into a conformed category. The definition of each conformed category is show below (in the format of `category`: `... definition ...`):
    - `小吃店`:小吃、擔仔麵、小吃攤、街邊小店、傳統小吃、麵食、麵攤、炒飯、餃子館、鯊魚煙、黑白切、牛肉麵、銅板美食、小點心、簡餐、色小菜、開放空間攤販
    - `日式料理(含居酒屋，串燒)`:居酒屋、酒場、水產、清酒、生魚片、壽司、日式啤酒、日式料理、代烤服務、日本餐飲場所、日本傳統食物、日式定食
    - `火(鍋／爐)`:麻辣鍋、薑母鴨、鴨味仔、鍋物、湯底、滋補、冬令補、涮涮鍋、個人鍋、冬天圍爐、羊肉爐、鴛鴦鍋、炭火爐、氣火爐、燒酒雞、蒸氣海鮮鍋
    - `東南亞料理(不含日韓)`:印尼、越式、泰式、沙嗲、海南雞、河粉、馬來西亞料理、新加坡料理、寮國料理、緬甸料理、南洋風味、印度料理、越南春捲、泰式綠咖哩、異國風情裝潢、滇緬料理
    - `海鮮熱炒`:海鮮、現撈、活海鮮、生猛、大排檔、活魚活蝦、生猛海鮮、快炒、海產、台式海鮮、下酒菜
    - `特色餐廳(含雞、鵝、牛、羊肉)`:烤鴨、燒鵝、甕仔雞、甕缸雞、桶仔雞、牛雜、蒙古烤肉、鵝肉城、金山鴨肉、生牛肉、全羊宴、活鱉、烤雞店、鵝肉餐廳、溫體牛、現宰羊肉、鹹水鵝、土羊肉
    - `傳統餐廳`:江浙、台菜、合菜、桌菜、粵菜、中式、川菜、港式、上海菜、砂鍋魚頭、東北菜、北京烤鴨、一鴨三吃、婚宴、辦桌、老字號、宴會廳、台灣料理
    - `燒烤`:燒烤、串燒、串串、烤魚、鮮蚵、炭烤、直火、碳火、和牛、戶外生火、烤肉、路邊燒烤
    - `韓式料理(含火鍋，烤肉)`:韓國泡菜、韓式年糕、首爾、燒酒、韓式炸雞、春川辣炒雞、韓式炸醬麵、海鮮煎餅、烤三層肉、烤五花、烤韓牛、醬料和飯、石鍋拌飯、韓式風格、韓式清酒、啤酒、銅盤烤肉、韓流
    - `PUB(Live Band)`:音樂餐廳、樂團表演、現場表演、LIVE表演、樂團駐唱、定期表演、有舞台場地、樂隊、專人駐唱
    - `PUB(一般，含Lounge)`:酒吧、bar、lounge、飛鏢、調酒、運動酒吧、音樂酒吧、沙發聊天、女公關、互動調酒師、公關服務
    - `PUB(電音＼舞場)`:夜店、舞池電音、藝人、包廂低消制、電子音樂表演、DJ、派對狂歡
    - `五星級飯店`:高級飯店、奢華酒店、連鎖五星級飯店、國際集團飯店、米其林飯店、高檔住宿
    - `自助KTV(含連鎖，庭園自助)`:卡拉OK、唱歌、歌坊、歡唱吧、自行點歌、自助唱歌、唱歌包廂、慶生聯誼包廂
    - `西餐廳(含美式，義式，墨式)`:牛排、餐酒、歐式、義式、西餐、義大利麵、凱薩沙拉、紅酒、白酒、調酒、墨西哥式料理、阿根廷式料理、漢堡、比薩
    - `咖啡廳(泡沫紅茶)`:泡沫紅茶店、咖啡店、café、coffee、輕食、軟性飲料、簡餐、茶街
    - `飯店(星級／旅館，不含五星級)`:飯店、酒店、商務旅館、平價住宿
    - `運動休閒館(含球類練習場，飛鏢等)`:撞球、高爾夫、運動、保齡球、娛樂、高爾夫練習場、大魯閣棒球場、籃球、羽毛球、PHOENIX鳳凰、羽球館、看球賽
    - `釣蝦場`:釣蝦、蝦寶、投幣卡拉OK、釣竿和餌料、蝦子現場烹煮食用、泰國蝦、現烤蝦子、包廂唱歌、現釣現烤、自備或租用釣竿。
    Note that you must choose from the above categories. Other ones are strongly prohibited. 
    Output in json format such as `{"category": "..."}`.

regularization_user_content: "{{ category }}"