SQL Extraction with LLM

- SQL Extraction Example (GPT4o)

Output Example 1

• Query: Please recommend an industrial public works program that can be supported until 2018.

{
   "sql_where_clause": "end_date >= '2018-01-01'",
   "explanation": "Filtering for programs that were available for support until at least the year 2018."
}

Output Example 2

• Query: Youth Support Programs Available in 2022.

{
   "sql_where_clause": "target_audience = 'youth' AND start_date <= '2022-12-31' AND end_date >= '2022-01-01'",
   "explanation": "The filter condition searches for programs targeted at 'youth' that were available at any time during the year 2022."
}

- SQL Extraction Example (GPT-OSS 20b)

Output Example 1

• Query: ...

...

Docker Environment

- Docker Build

docker build -t sql_extraction_test .

- Docker Run

docker run -it --gpus all --name sql_extraction_test_env --shm-size=64G -p {port}:{port} -e GRANT_SUDO=yes --user root -v {root_folder}:/workspace/sql_extraction_test -w /workspace/sql_extraction_test sql_extraction_test bash

- Docker Exec

docker exec -it sql_extraction_test_env bash

SQL Extraction Test

- SQL Extraction (GPT4o)

• When using the GPT API, enter the API KEY (token) in the '.env' file created after executing the initialization command

python sql_extraction.py --config_path config/llm_config.yaml --user_query "{user_query}" --llm gpt_api

- SQL Extraction (GPT-OSS 20b)

python sql_extraction.py --config_path config/llm_config.yaml --user_query "{user_query}" --llm gpt_oss_20b

Author

- LinkedIn: https://www.linkedin.com/in/taeyong-kong-016bb2154

- Blog URL: https://blog.naver.com/qbxlvnf11

- Email: qbxlvnf11@google.com, qbxlvnf11@naver.com