FlashExtract: A Framework for Data Extraction by Examples

Abstract not retrievable programmatically at ingestion; this summary is written from the title and the forwarder's note, not the paper's text. FlashExtract (Le and Gulwani, PLDI 2014) is a programming-by-example framework: the user highlights a few fields in a text file, web page or spreadsheet and the system synthesises a reusable extraction program that pulls the same fields from the rest of the document. Forwarded as the general statement of the locator idea, an extractor written once from examples and run without further interaction, and as the ancestor of the LLM-generated regex and XPath papers filed beside it.