what does lulu screened do?
Reads a finished screening round back and turns it into the document, with every question it measured printed beside its verdict and the names the same draws reached for underneath.
how do you run it?
In the parser’s own words, read a finished screening round back, and what it found. Every flag this build accepts, on one line:
lulu screened --draft PATH [--ledger LEDGER] [--tex PATH] [--pdf PATH]
lulu screened --draft PATH ± on the terminallulu screened --draft PATH --pdf screened.pdf ± and as a typeset document
why read a round back at all?
lulu draft prints its screening round once, while it is spending, and then that round is a scrollback and two files of JSON. That is enough to run the next command and not enough to hand to anybody.
Nothing is spent and nothing is asked. Every figure comes out of the draft ledger and the round it names, so the same two files give the same document forever, including to whoever is checking whether it was true.
what happens to a question that named nobody?
It is printed in full rather than summarised away. On the first paid screening round this library ran, eleven of twenty questions came back with none of the declared rivals named in any draw, which is a larger statement about a market than the one score that round produced, and it had nowhere to be read until this command existed.
That round also showed why a barren verdict is a statement about a list. Its rivals were read off the arm that answers from its own weights and the round was collected on the arm that searches, and the two arms reach for different companies. So both surfaces print the names the answers used that the declared list never mentioned, off the same draws, at no extra cost.