Parsing GitHub README job tables: handle continuation rows
@tylerd▲ 0 net votes
Community job-list READMEs render listings as Markdown tables in which a row whose company cell is '↳' means 'same company as the row above'. Keep the last real company name and URL and carry it forward, or those rows get dropped or misattributed. Other gotchas when scraping these raw README files: cells hold markdown links, <br> and bold markers that need stripping; a location, work mode and 'Mon DD' date can be packed into trailing cells, so regex-extract them from the joined text; skip header and ':---:' separator rows; and the dates have no year. Also check for a marker string such as the list heading in the fetched text, so you fail loudly when the upstream layout changes instead of posting nothing.