Merge pull request #4 from PrompTartLab/dev

feature: launch the app

.github/auto_assign.yml

@@ -0,0 +1,26 @@
+# Set to true to add reviewers to pull requests
+addReviewers: true
+
+# Set to true to add assignees to pull requests
+addAssignees: author
+
+# A list of reviewers to be added to pull requests (GitHub user name)
+reviewers: 
+  - DimensionSTP
+  - thispath98
+  - aeolian83
+  - jiyoon9704
+  - milhaud1201
+
+# A number of reviewers added to the pull request
+# Set 0 to add all the reviewers (default: 0)
+numberOfReviewers: 0
+
+# A number of assignees to add to the pull request
+# Set to 0 to add all of the assignees.
+# Uses numberOfReviewers if unset.
+# numberOfAssignees: 1
+
+# A list of keywords to be skipped the process that add reviewers if pull requests include it
+# skipKeywords:
+#   - wip

.github/labeler.yml

@@ -0,0 +1,54 @@
+js:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.js"
+
+ts:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.ts"
+
+py:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.py"
+
+java:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.java"
+
+c:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.c"
+
+c++:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.cpp"
+
+c#:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.cs"
+
+swift:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.swift"
+
+kotlin:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.kt"
+
+go:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.go"
+
+elixir:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "**/*.exs"

.github/pull_request_template.md

@@ -0,0 +1,16 @@
+## PR Checklist
+
+- **Tag**: Select the appropriate tag for this PR:
+  - [ ] `feature`
+  - [ ] `fix`
+  - [ ] `refactor`
+  - [ ] `test`
+
+- **Module**: Select the relevant module(s) affected by this PR:
+  - [ ] `sql`
+  - [ ] `rag`
+  - [ ] `base`
+  - [ ] `deploy`
+
+## Description
+- 

.github/workflows/auto-assign-label.yaml

@@ -0,0 +1,24 @@
+name: Auto Assign and Label
+
+on:
+  pull_request:
+    types: [opened, reopened]
+
+jobs:
+  assign-reviewers-assignees:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: kentaro-m/[email protected]
+
+  label-lang:
+    runs-on: ubuntu-latest
+    continue-on-error: true
+
+    permissions:
+      contents: read
+      pull-requests: write
+
+    steps:
+      - uses: actions/labeler@v5
+        with:
+          repo-token: ${{ github.token }}

.gitignore

@@ -0,0 +1,3 @@
+# folder
+.streamlit/
+*.pyc

configs/default_translate_params.py

@@ -0,0 +1,6 @@
+class DefaultTranslateParams:
+    def __init__(self):
+        self.temperature = 0.0
+        self.template = """
+You are a professional translator. Given the query "{query}", provide the best result of Korean to English.
+"""

configs/default_web_params.py

@@ -0,0 +1,9 @@
+class DefaultWebParams:
+    def __init__(self):
+        self.temperature = 0.0
+        self.template = """
+You are a search assistant. Given the query "{query}", provide the best search results and a summary.
+"""
+        self.serpapi_params = {
+            "engine": "google",
+        }

configs/knowledge.py

@@ -0,0 +1,28 @@
+busan_general_knowledge ='''
+광안리는 수영구에 있다.
+부산역은 동구에 있다.
+해운대는 해운대구에 있다.
+부산항은 중구와 동구에 걸쳐 있다.
+감천항은 사하구에 있다.
+광안대교는 수영구와 해운대구를 연결한다.
+송도해수욕장은 서구에 있다.
+다대포해수욕장은 사하구에 있다.
+국제시장과 자갈치시장은 중구에 있다.
+영도구는 중구와 다리로 연결되어 있다.
+동백섬은 해운대구에 있다.
+송정해수욕장은 해운대구에 있다.
+부산 영화의 전당은 해운대구 센텀시티에 위치해 있다.
+충렬사는 동래구에 있다.
+오륜대는 연제구에 있다.
+태종대는 영도구에 있다.
+서구는 사하구, 사상구, 동구, 중구와 붙어있다.
+해운대구는 수영구, 동래구, 기장군과 붙어있다.
+남구는 수영구, 연제구, 사상구, 서구와 붙어있다.
+기장군은 해운대구와 동래구에 접해 있다.
+동래구는 연제구, 금정구, 해운대구와 붙어있다.
+금정구는 동래구, 연제구, 사상구와 접해 있다.
+연제구는 남구, 동래구, 금정구, 해운대구와 붙어있다.
+동구는 중구, 서구, 사하구, 남구와 접해 있다.
+사상구는 북구, 금정구, 서구와 접해 있다.
+북구는 사상구, 금정구와 접해 있다.
+'''

configs/prompts.py

@@ -0,0 +1,62 @@
+from langchain_core.prompts import PromptTemplate
+
+# Template for SQL generation (initial attempt)
+sql_generation_template = PromptTemplate(
+    input_variables=["question", "datasource", "schema", "external_knowledge"],
+    template="""
+    You are an expert in generating SQL queries. Based on the user's question, external knowledge and the specified data source, generate an SQL query. 
+
+    Data source: {datasource}
+    Table schema: {schema}
+    External knowledge:{external_knowledge}
+
+    Ensure the query matches the schema of the data source and retrieves the most relevant information. Do not filter with extra columns that are not explicitly requested in the query.
+    Use the 'LIKE' operator instead of equal(=) to filter text-type columns such as MENU_NAME or reviews and utilize the 'IN' operator for multiple possible options.
+    Note that all columns other than those ending in English are Korean characters, and make sure that categorical columns are selected only in Possible values.
+    Let's think step by step; first divide the user query into several conditions, tailor each condition to a possible value in the schema, and then convert each condition into a SQL statement.
+    After the thoughts, provide the final SQL query in the <SQL> </SQL> tag.
+
+    For your information, I'll provide examples of query-answer pairs.
+
+    <QUERY> 부산역 근처 돼지국밥 맛집 알려주세요.
+    <ANSWER> "부산역 근처" indicates proximity to Busan Station. 부산역 is located in 동구. This can be addressed using the DISTRICT_NAME column.
+            "돼지국밥" specifies the type of food, which can be matched with the MENU_NAME column.
+            맛집" implies that the restaurant should have a high rating, likely using the RATING column.
+            <SQL> SELECT RESTAURANT_NAME_KOREAN, ADDRESS_KOREAN FROM restaurants WHERE DISTRICT_NAME = '동구' AND MENU_NAME LIKE '%돼지국밥%' ORDER BY RATING DESC LIMIT 5; </SQL>
+
+    <QUERY> 광안리 근처에서 아이들이랑 가기 좋은 식당 추천해주세요.
+    <ANSWER> "광안리 근처" indicates proximity to 광안리. 광안리 is in 수영구. This can be addressed using the DISTRICT_NAME column.
+            Childeren might needs menu for childeren, which can be matched with MENU_FOR_CHILDREN_YN column. This column has Boolean type.
+            <SQL> SELECT RESTAURANT_NAME_KOREAN, MENU_NAME FROM restaurants WHERE DISTRICT_NAME = '수영구' AND MENU_FOR_CHILDREN_YN = True ORDER BY RATING DESC LIMIT 5; </SQL>
+
+    <QUERY> 외국인 친구랑 갈만한 관광지 알려주세요.
+    <ANSWER> Foreign friends may like historical experience, which can be matched with the Major_Category column with '역사관광지'.
+            <SQL> SELECT Tourist_Spot_Name_Korean FROM foreign_tourist_spots WHERE Major_Category = '역사관광지' ORDER BY Number_of_Visit DESC LIMIT 5; </SQL>
+
+    <QUERY> 서구 근처에서 할 만한 액티비티 알려주세요.
+    <ANSWER> 서구 근처" indicates proximity to 서구. 서구 is connected to 사하구, 사상구, 동구, and 중구. This can be addressed using the District column.
+            "액티비티" specifies the type of tour, which can be matched with the Major_Category column with '육상 레포츠' '수상 레포츠' and '레포츠소개'.
+            <SQL> SELECT Tourist_Spot_Name_Korean FROM local_tourist_spots WHERE District IN ('사하구', '사상구', '동구', '중구') AND Major_Category IN ('육상 레포츠', '수상 레포츠', '레포츠소개') ORDER BY Number_of_Visit DESC LIMIT 5; </SQL>
+
+    <QUERY> {question}
+    <ANSWER>
+    """
+)
+
+# Template for SQL generation (retry attempt)
+sql_retry_template = PromptTemplate(
+    input_variables=["question", "datasource", "schema", "external_knowledge", "previous_answer"],
+    template="""
+    You are an expert in generating SQL queries. Based on the user's question, external knowledge, and the specified data source, generate an SQL query.
+
+    In a prior attempt, you generated a SQL query that returned no results. Relax constraints to retrieve relevant data.
+
+    Data source: {datasource}
+    Table schema: {schema}
+    External knowledge:{external_knowledge}
+    Prior SQL: {previous_answer}
+    User query: {question}
+
+    Provide the final SQL query in the <SQL> </SQL> tag.
+    """
+)

configs/schemas.py

@@ -0,0 +1,45 @@
+schemas = {
+    "local_tourist_spots": '''
+        - Description: This table contains information about tourist spots for local visitors.
+        - Columns:
+            - Tourist_Spot_Name_Korean (TEXT): The name of the tourist spot in Korean.
+            - District (TEXT): The district (구) where the tourist spot is located.
+            - Major_Category (CATEGORICAL): The category of the tourist spot. Possible values: ['역사관광지', '건축/조형물', '휴양관광지', '쇼핑', '체험관광지', '자연관광지', '육상 레포츠', '음식점', '문화시설', '산업관광지', '숙박시설', '수상 레포츠', '레포츠소개'].
+            - Number_of_Visit (INTEGER): The number of visitors to the tourist spot.
+    ''',
+    "foreign_tourist_spots": '''
+        - Description: This table contains information about tourist spots for foreign visitors.
+        - Columns:
+            - Tourist_Spot_Name_English (TEXT): The name of the tourist spot in English.
+            - Tourist_Spot_Name_Korean (TEXT): The name of the tourist spot in Korean.
+            - District (TEXT): The district (구) where the tourist spot is located.
+            - Major_Category (CATEGORICAL): The category of the tourist spot. Possible values: ['역사관광지', '휴양관광지', '체험관광지', '쇼핑', '자연관광지', '건축/조형물', '육상 레포츠', '산업관광지', '문화시설', '수상 레포츠', '음식점', '공연/행사', '숙박시설', '레포츠소개'].
+            - Number_of_Visit (INTEGER): The number of visitors to the tourist spot.
+    ''',
+    "restaurants":'''
+        - Description: This table contains information about popular restaurants in Busan.
+        - Columns:
+            - MENU_TYPE (TEXT): The type of menu offered at the restaurant (e.g., 한식, 중식, 분식).
+            - RESTAURANT_NAME_KOREAN (TEXT): The Korean name of the restaurant.
+            - ADDRESS_KOREAN (TEXT): The Korean address of the restaurant.
+            - MENU_NAME (TEXT): The list of menu items offered by the restaurant.
+            - DISTRICT_NAME (TEXT): The district (구) where the restaurant is located.
+            - LATITUDE (REAL): The latitude of the restaurant.
+            - LONGITUDE (REAL): The longitude of the restaurant.
+            - RATING (REAL): The rating of the restaurant (out of 5).
+            - RATING_COUNT (INTEGER): The number of reviews that contributed to the restaurant's rating.
+            - PRICE_LEVEL (INTEGER): The price level of the restaurant (1 = low, 2 = medium, 3 = high).
+            - TAKEOUT_YN (BOOLEAN): Whether the restaurant offers takeout (True/False).
+            - RESERVABLE (BOOLEAN): Whether the restaurant accepts reservations (True/False).
+            - BREAKFAST_YN (BOOLEAN): Whether breakfast is served (True/False).
+            - LUNCH_YN (BOOLEAN): Whether lunch is served (True/False).
+            - DINNER_YN (BOOLEAN): Whether dinner is served (True/False).
+            - BEER_YN (BOOLEAN): Whether the restaurant serves beer (True/False).
+            - OUTDOOR_SEAT_YN (BOOLEAN): Whether the restaurant has outdoor seating (True/False).
+            - MENU_FOR_CHILDREN_YN (BOOLEAN): Whether the restaurant offers a children's menu (True/False).
+            - RESTROOM_YN (BOOLEAN): Whether the restaurant has restrooms (True/False).
+            - PARKING_LOT_YN (BOOLEAN): Whether the restaurant provides parking (True/False).
+            - PAYMENT_OPTIONS (TEXT): Payment options available (e.g., credit cards).
+            - REVIEW (TEXT): Customer reviews for the restaurant.
+'''
+}