The Code Generation & Summarization Problems
摘要
To gain a broader perspective on the Text-to-SQL and SQL-to-Text tasks discussed in previous chapters, it is useful to examine how similar tasks are addressed in the wider field of generating natural language to and from code. This perspective helps highlight both the shared challenges and the unique characteristics that have motivated the development of systems specifically tailored to SQL. This chapter aims to provide a concise but comprehensive overview of the Code Generation and the Code Summarization problems. It presents the tasks, provides an overview of the most widely used datasets and evaluation metrics and describes representative systems in the field.