FlyingFeather / DEA-SQL

[ACL Findings 2024] Decomposition for Enhancing Attention: Improving LLM-based Text-to-SQL through Workflow Paradigm
Apache License 2.0
22 stars 1 forks source link

BIRD Support #2

Open JaviS-Rei opened 1 month ago

JaviS-Rei commented 1 month ago

Hello, Is there a version available that supports the BIRD dataset? I noticed that there is support for BIRD in the code, but I am not sure if the fewshot folder needs to be regenerated on the BIRD training set (processor.py) or just use it without regeneration?

Nisha03-wd commented 1 month ago

Decomposition for Enhancing Attention: Improving LLM-based Text-to-SQL through Workflow Paradigm can u plz give how to run the i did all but i m not able to access database. Will you plz help me

FlyingFeather commented 1 month ago

Thank you very much for your attention to our work. The fewshot file for 'bird' is not using the current results in the repository. Recently, we have supplemented experiments on the bird dataset, but the relevant code has not been updated yet. It is expected to be updated within the next two weeks.

Best regards

FlyingFeather commented 1 month ago

Decomposition for Enhancing Attention: Improving LLM-based Text-to-SQL through Workflow Paradigm can u plz give how to run the i did all but i m not able to access database. Will you plz help me

Could you please be more specific? Are you unable to connect to the database with SQLite? You might want to check if the file exists and if the dependent environment is fully installed.

Nisha03-wd commented 1 month ago

Downloaded the data set from the Spider website and i made the data folder and uplode the dataset but i get the error that

PS C:\Users\91862\OneDrive\Desktop\Text To Sql\DEA-SQL> python evaluation/test-suite-sql-eval/evaluation.py --gold "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/evaluation/gold_files/spider_dev_gold.sql" --pred "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/outputs/spider/dea-sql.txt" --db "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/data/spider/database/example.db" --print_file_name "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/outputs/spider/spider-dea-sql.txt" --table "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/data/spider/tables.json" --etype exec 0 Traceback (most recent call last): File "C:\Users\91862\OneDrive\Desktop\Text To Sql\DEA-SQL\evaluation\test-suite-sql-eval\evaluation.py", line 1013, in

evaluate(args.gold, args.pred, args.db, args.etype, kmaps, args.plug_value, File "C:\Users\91862\OneDrive\Desktop\Text To Sql\DEA-SQL\evaluation\test-suite-sql-eval\evaluation.py", line 593, in evaluate schema = Schema(get_schema(db)) File "C:\Users\91862\OneDrive\Desktop\Text To Sql\DEA-SQL\evaluation\test-suite-sql-eval\process_sql.py", line 88, in get_schema conn = sqlite3.connect(db) sqlite3.OperationalError: unable to open database file Can You please tell me how to run this like the line of code beacause i use this "python evaluation/test-suite-sql-eval/evaluation.py --gold "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/evaluation/gold_files/spider_dev_gold.sql" --pred "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/outputs/spider/dea-sql.txt" --db "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/data/spider/database/example.db" --print_file_name "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/outputs/spider/spider-dea-sql.txt" --table "C:/Users/91862/OneDrive/Desktop/Text To Sql/DEA-SQL/data/spider/tables.json" --etype exec " Which i get from chtgpt On Sun, 14 Jul 2024 at 14:12, Feather ***@***.***> wrote: > Decomposition for Enhancing Attention: Improving LLM-based Text-to-SQL > through Workflow Paradigm can u plz give how to run the i did all but i m > not able to access database. Will you plz help me > > Could you please be more specific? Are you unable to connect to the > database with SQLite? You might want to check if the file exists and if the > dependent environment is fully installed. > > — > Reply to this email directly, view it on GitHub > , > or unsubscribe > > . > You are receiving this because you commented.Message ID: > ***@***.***> >
JaviS-Rei commented 1 month ago

Thank you very much for your attention to our work. The fewshot file for 'bird' is not using the current results in the repository. Recently, we have supplemented experiments on the bird dataset, but the relevant code has not been updated yet. It is expected to be updated within the next two weeks. Best regards

Thanks for reply. Waiting for bird version and after that this issue will be close.

downeykking commented 1 month ago

Decomposition for Enhancing Attention: Improving LLM-based Text-to-SQL through Workflow Paradigm can u plz give how to run the i did all but i m not able to access database. Will you plz help me

Hello, thank you for attention to our work! We have updated a more detailed environment setup in README.md. You can re-install the environment by following the instructions in setup steps. Then run python evaluation/test-suite-sql-eval/evaluation.py --gold "evaluation/gold_files/spider_dev_gold.sql" --pred "outputs/spider/dea-sql.txt" --db ./data/spider/database --print_file_name "outputs/spider/spider-dea-sql.txt" --table './data/spider/tables.json' --etype exec and check whether you can get result like this:

image

Hope this could help you.

JaviS-Rei commented 3 weeks ago

It has been two weeks since our last communication, and there is any progress on opening BIRD related data and code? They are extremely beneficial to our ongoing research, and potentially help our work to contribute to the SE and DB community. Thank you for your time and consideration.

FlyingFeather commented 2 weeks ago

It has been two weeks since our last communication, and there is any progress on opening BIRD related data and code? They are extremely beneficial to our ongoing research, and potentially help our work to contribute to the SE and DB community. Thank you for your time and consideration.

Thank you very much for your interest in our work. Because we are currently intensively conducting a new research work related to the bird data set, it is expected that the update of this code will be postponed. Once updated, we will contact you about this issue and hope to be able to help you by then.

In addition, the fewshot file you mentioned earlier is not used in the bird data set. As explained in the article, it is not a core module and the effect is not significant in GPT-3.5.