code wiki / _hdl_build / nx_external_search_gate.nx
nx_external_search_gate.nx source
↩ module page · 96 lines · 7091 B
1// nx_external_search_gate.nx -- MEASURED gate for EXTERNAL / offsite search (the charter's scope=offsite, now
2// real). Rebuilds the external index from the sovereign-fetched-web corpus (knowledge/index/web_src.tsv, built
3// by nx_external_corpus from knowledge/fetched/*.raw) via the reusable ingestor, then forks the REAL
4// /tmp/nx_onsite_search.sov.elf and checks each query lands the RIGHT external page at rank #1 -- proving the
5// SAME engine serves offsite as onsite. MEASURED from the live client, never asserted. GREEN iff 4/4. Durable
6// knowledge/status/external_search_gate.log. license_tier: ORIGINAL
7import "nx_syscalls.nx"
8
9func x_puts(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} sys_write(1,s,n); return 0 }
10func x_num(v: i64) -> i64 { let bb: *u8=sys_mmap(28); var m: i64=v; if m<0{m=0-m;sys_write(1,"-" as *u8,1)}; let t: *u8=sys_mmap(28); var k: i64=0; if m==0{t[0]=(48 as u8);k=1}; while m>0{t[k]=((48+(m%10)) as u8);m=m/10;k=k+1}; var i: i64=0; while i<k{bb[i]=t[k-1-i];i=i+1}; sys_write(1,bb,k); return 0 }
11func x_w(fd: i64, s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} sys_write(fd,s,n); return 0 }
12func x_wn(fd: i64, v: i64) -> i64 { let bb: *u8=sys_mmap(28); var m: i64=v; if m<0{m=0-m}; let t: *u8=sys_mmap(28); var k: i64=0; if m==0{t[0]=(48 as u8);k=1}; while m>0{t[k]=((48+(m%10)) as u8);m=m/10;k=k+1}; var i: i64=0; while i<k{bb[i]=t[k-1-i];i=i+1}; sys_write(fd,bb,k); return 0 }
13func x_strlen(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} return n }
14func x_write_file(path: *u8, data: *u8) -> i64 { let fd: i64=sys_openat_wr(path, 0x1a4); if fd<0 { return 0-1 } sys_write(fd, data, x_strlen(data)); sys_close(fd); return 0 }
15func x_read_small(path: *u8, buf: *u8, cap: i64) -> i64 {
16 let fd: i64=sys_openat_rd(path); if fd<0 { return 0 }
17 var total: i64=0; var nrd: i64=sys_read(fd, buf, cap)
18 while nrd>0 { total=total+nrd; if total>=cap { nrd=0 } else { nrd=sys_read(fd, ((buf as i64)+total) as *u8, cap-total) } }
19 sys_close(fd); return total
20}
21func x_index(a1: *u8, a2: *u8, a3: *u8, a4: *u8) -> i64 {
22 let pid: i64=sys_fork()
23 if pid==0 {
24 let o: i64=sys_openat_wr("/tmp/xg_idx.txt" as *u8, 0x1a4); if o>=0 { sys_dup3(o,1,0) }
25 let argv: *i64=sys_mmap(8*8) as *i64
26 argv[0]="/tmp/nx_onsite_index.sov.elf" as *u8 as i64; argv[1]=a1 as i64; argv[2]=a2 as i64; argv[3]=a3 as i64; argv[4]=a4 as i64; argv[5]=0
27 let envp: *i64=sys_mmap(16) as *i64; envp[0]="PATH=/usr/bin:/bin" as *u8 as i64; envp[1]=0
28 sys_execve("/tmp/nx_onsite_index.sov.elf" as *u8, argv, envp); sys_exit(127)
29 }
30 let st: *i64=sys_mmap(16) as *i64; sys_wait4(pid, st, 0); return (st[0]>>8)&0xff
31}
32func x_client(sites: *u8, terms: *i64, nt: i64, out_path: *u8) -> i64 {
33 let pid: i64=sys_fork()
34 if pid==0 {
35 let o: i64=sys_openat_wr(out_path, 0x1a4); if o>=0 { sys_dup3(o,1,0) }
36 let argv: *i64=sys_mmap(8*16) as *i64
37 argv[0]="/tmp/nx_onsite_search.sov.elf" as *u8 as i64; argv[1]=sites as i64; argv[2]="web" as *u8 as i64
38 var i: i64=0; while i<nt { argv[3+i]=terms[i]; i=i+1 } argv[3+nt]=0
39 let envp: *i64=sys_mmap(16) as *i64; envp[0]="PATH=/usr/bin:/bin" as *u8 as i64; envp[1]=0
40 sys_execve("/tmp/nx_onsite_search.sov.elf" as *u8, argv, envp); sys_exit(127)
41 }
42 let st: *i64=sys_mmap(16) as *i64; sys_wait4(pid, st, 0); return (st[0]>>8)&0xff
43}
44func x_find(hay: *u8, n: i64, needle: *u8) -> i64 {
45 let nl: i64=x_strlen(needle); if nl==0 { return 0-1 }
46 var i: i64=0
47 while i+nl<=n { var j: i64=0; var hit: i64=1; while j<nl { if hay[i+j]!=needle[j] { hit=0; j=nl } else { j=j+1 } } if hit==1 { return i } i=i+1 }
48 return 0-1
49}
50func x_top_is(buf: *u8, n: i64, expected: *u8) -> i64 {
51 let at: i64=x_find(buf, n, expected); if at<0 { return 0 }
52 var ls: i64=at
53 while ls>0 { if buf[ls-1]==(10 as u8) { ls=0-ls } else { ls=ls-1 } }
54 if ls<0 { ls=0-ls }
55 let h: i64=x_find(((buf as i64)+ls) as *u8, at-ls+1, "#" as *u8); if h<0 { return 0 }
56 var p: i64=ls+h+1; var rank: i64=0
57 while p<n { let c: i64=buf[p] as i64; if c>=48 { if c<=57 { rank=rank*10+(c-48); p=p+1 } else { p=n } } else { p=n } }
58 if rank==1 { return 1 }
59 return 0
60}
61func x_row(sites: *u8, label: *u8, terms: *i64, nt: i64, expect: *u8, obuf: *u8) -> i64 {
62 x_client(sites, terms, nt, "/tmp/xg_out.txt" as *u8)
63 let n: i64=x_read_small("/tmp/xg_out.txt" as *u8, obuf, 65536)
64 let ok: i64=x_top_is(obuf, n, expect)
65 x_puts(" q='" as *u8); x_puts(label); x_puts("' -> " as *u8); x_puts(expect)
66 if ok==1 { x_puts(" #1 PASS\n" as *u8) } else { x_puts(" NOT#1 FAIL\n" as *u8) }
67 return ok
68}
69
70func main() -> i64 {
71 x_puts("=== EXTERNAL SEARCH GATE (offsite: sovereign-fetched web pages, reusable client; MEASURED) ===\n" as *u8)
72 let rc: i64=x_index("knowledge/index/web_src.tsv" as *u8, "https://en.wikipedia.org/wiki/" as *u8,
73 "knowledge/index/web.idx" as *u8, "knowledge/index/web.manifest" as *u8)
74 if rc!=0 { x_puts("GATE-FAIL: external index rc=" as *u8); x_num(rc); x_puts(" (run nx_external_corpus + rebuild /tmp deps?)\n" as *u8); sys_exit(1); return 1 }
75 x_write_file("/tmp/web_sites.tsv" as *u8, "web\tknowledge/index/web.idx\tknowledge/index/web.manifest\t\n" as *u8)
76 let sites: *u8="/tmp/web_sites.tsv" as *u8
77 let obuf: *u8=sys_mmap(65536)
78
79 let q1: *i64=sys_mmap(8*6) as *i64; q1[0]="okapi" as *u8 as i64; q1[1]="probabilistic" as *u8 as i64; q1[2]="ranking" as *u8 as i64
80 let q2: *i64=sys_mmap(8*6) as *i64; q2[0]="retrieval" as *u8 as i64; q2[1]="augmented" as *u8 as i64; q2[2]="generation" as *u8 as i64
81 let q3: *i64=sys_mmap(8*6) as *i64; q3[0]="faceted" as *u8 as i64; q3[1]="navigation" as *u8 as i64; q3[2]="filters" as *u8 as i64
82 let q4: *i64=sys_mmap(8*6) as *i64; q4[0]="elasticsearch" as *u8 as i64; q4[1]="distributed" as *u8 as i64; q4[2]="java" as *u8 as i64
83
84 var pass: i64=0
85 let rows: i64=4
86 pass=pass + x_row(sites, "okapi probabilistic ranking" as *u8, q1, 3, "/wiki/Okapi_BM25" as *u8, obuf)
87 pass=pass + x_row(sites, "retrieval augmented generation" as *u8, q2, 3, "/wiki/Retrieval-augmented_generation" as *u8, obuf)
88 pass=pass + x_row(sites, "faceted navigation filters" as *u8, q3, 3, "/wiki/Faceted_search" as *u8, obuf)
89 pass=pass + x_row(sites, "elasticsearch distributed java" as *u8, q4, 3, "/wiki/Elasticsearch" as *u8, obuf)
90
91 x_puts("----\nEXTERNAL-SEARCH rows=" as *u8); x_num(rows); x_puts(" pass=" as *u8); x_num(pass); x_puts("\n" as *u8)
92 let lg: i64=sys_openat_append("knowledge/status/external_search_gate.log" as *u8, 0x1a4)
93 if lg>=0 { x_w(lg, "EXTERNAL-SEARCH rows=" as *u8); x_wn(lg, rows); x_w(lg, " pass=" as *u8); x_wn(lg, pass); if pass==rows { x_w(lg, " verdict=GREEN\n" as *u8) } else { x_w(lg, " verdict=RED\n" as *u8) } sys_close(lg) }
94 if pass==rows { x_puts("EXTERNAL-SEARCH GREEN (offsite web pages retrieved + ranked by the reusable engine)\n" as *u8); sys_exit(0); return 0 }
95 x_puts("EXTERNAL-SEARCH RED\n" as *u8); sys_exit(1); return 1
96}