nx_learnmine.nx

nx_learnmine.nx -- THE LEARNING-MINING COMMAND (operator 2026-09-06: farm creators and archives for videos, transcripts and how-to text as supporting guides for the /compare rungs). Rung 1: digest an archive.org advancedsearch listing (output=json, mirrored by nx_research_fetch) into item rows and a census, so a seat reads two hundred rows instead of a quarter-megabyte. usage: nx_learnmine items <listing.json> [max] one ITEM row per item (idx, identifier, mediatype, date, downloads, title, subjects) and a DESC row per described item (full length announced, head shown) nx_learnmine subjects <listing.json> [top] the subject frequency table, count descending (a FLOOR when the table filled) nx_learnmine census <listing.json> the summary line only rows are pipe-separated and every field is sanitised (a pipe in the source becomes a slash, a newline a space), so a row is one line and the summary is LAST. exits: 0 digested (GREEN) - 2 usage - 3 refused by name (UNREADABLE, NO-DOCS, MALFORMED, CAPACITY). Nothing here fetches: the bytes come from the provenance-journalled mirror. license_tier: ORIGINAL No hw writes (Rule 26).