nx_dr_chain_cli.nx
buildroot/runtime/nx_dr_chain_cli.nx
about
nx_dr_chain_cli.nx -- ONE CALL runs the whole sovereign deep-research loop (DR-CHAIN).
judge (tokenize+containment of every source vs the shared groundtruth insight set)
-> conduct (verify-yield -> Elo tournament -> opportunity-cost budget dispatch)
-> score (insight-recall of what was actually selected, vs what was achievable)
nx_dr_chain <threshold_permil> <budget> <n_insights> <insight_1..n> <cost_1> <doc_1> ...
e.g. nx_dr_chain 600 3 3 "alpha one" "beta two" "gamma three" \
2 "alpha one beta two" 3 "gamma three extra words here" 1 "alpha one"
-> {"recall_all_permil":1000,"recall_selected_permil":666,"budget_cost_permil":334,...}
The recall_all vs recall_selected delta is the MEASURED COST of the opportunity-cost
constraint -- the number no single-agent research tool reports.
license_tier: ORIGINAL No hw writes (Rule 26). expect_exit: 0
dependencies 6 imports · 0 importers
imports: nx_dr_chain.nxnx_itoa_lib.nxnx_dr_run.nxnx_dr_conductor.nxnx_dr_drbench.nxnx_syscalls.nx
imported by: nobody (leaf or entry point)
call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown
structs
| none |
consts
| 18 | const K_MAGIC_8192: i64 = 8192 |
| 19 | const K_MAGIC_1024: i64 = 1024 |
| 20 | const K_MAGIC_65536: i64 = 65536 |
functions
| 22 | func ch_puts(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(1, s, n); return 0 } |
| 23 | func ch_q() -> i64 { let b: *u8 = sys_mmap(1); b[0] = 34 as u8; sys_write(1, b, 1); return 0 } |
| 24 | func ch_key(name: *u8) -> i64 { ch_q(); ch_puts(name); ch_q(); ch_puts(":" as *u8); return 0 } |
| 29 | func ch_num(v: i64) -> i64 { nxi_out(v); return 0 } |
| 30 | func ch_atoi(a: *u8) -> i64 called by 1: main |
| 40 | func main(argc: i64, argv: *i64) -> i64 |