nx_research_extract_sovereign_test.nx source

1// nx_research_extract_sovereign_test.nx -- the SOVEREIGN half of the 2// B3 research loop, now fed by the TEAM's OWN fetch (no Playwright). 3// 4// Reads the sovereign-fetched raw HTTP response 5// knowledge/fetched/wiki_web_dev_sovereign.raw (from nx_research_fetch_wiki) 6// strips the HTTP headers (\r\n\r\n boundary), renders HTML->text with the 7// team's nx_html_to_text, saves the rendered text to 8// knowledge/fetched/wiki_web_dev_sovereign.txt 9// (apples-to-apples with the Playwright benchmark's wiki_web_dev.txt), 10// then drives the team's deterministic ce_extract over it. 11// 12// Division (unchanged): FETCH = sovereign (was the Playwright benchmark); 13// RENDER + EXTRACT = sovereign Nishi; only the qualitative prose->feature 14// semantic residual is the LLM/Nishi-AI seat. 15// 16// Lives in _hdl_build/ so the import search reaches BOTH nx_claim_extract 17// (here) and nx_html_to_text (../). 18// 19// expect_exit: 0 20// license_tier: ORIGINAL 21 22import "nx_syscalls.nx" 23import "nx_claim_extract.nx" 24import "nx_html_to_text.nx" 25 26func rr_puts(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} sys_write(1,s,n); return 0 } 27func rr_putn(v: i64) -> i64 { 28 let bb: *u8=sys_mmap(28); var m: i64=v 29 if m<0{m=0-m;sys_write(1,"-" as *u8,1)} 30 let t: *u8=sys_mmap(28); var k: i64=0 31 if m==0{t[0]=48 as u8;k=1} 32 while m>0{t[k]=(48+(m%10)) as u8;m=m/10;k=k+1} 33 var i: i64=0; while i<k{bb[i]=t[k-1-i];i=i+1} sys_write(1,bb,k); return 0 34} 35 36func main() -> i64 { 37 let raw: *u8 = sys_mmap(2097152) 38 let rn: i64 = ce_read_file("knowledge/fetched/wiki_web_dev_sovereign.raw" as *u8, raw, 2097151) 39 if rn <= 0 { rr_puts("SOV-EXTRACT: raw fetch file not found/empty\n" as *u8); sys_exit(1); return 1 } 40 41 // ---- Find the HTTP header/body boundary (\r\n\r\n) ---- 42 var bo: i64 = 0 43 var p: i64 = 0 44 while p + 3 < rn { 45 if raw[p]==(13 as u8) { if raw[p+1]==(10 as u8) { if raw[p+2]==(13 as u8) { if raw[p+3]==(10 as u8) { 46 bo = p + 4 47 p = rn 48 } } } } 49 p = p + 1 50 } 51 if bo == 0 { rr_puts("SOV-EXTRACT: no header/body boundary found\n" as *u8); sys_exit(2); return 2 } 52 let body_len: i64 = rn - bo 53 54 // ---- Sovereign HTML -> text render ---- 55 let txt: *u8 = sys_mmap(2097152) 56 let tn: i64 = nx_html_to_text((raw as i64 + bo) as *u8, body_len, txt, 2097151) 57 if tn <= 0 { rr_puts("SOV-EXTRACT: html_to_text produced no text\n" as *u8); sys_exit(3); return 3 } 58 59 // ---- Persist rendered text (apples-to-apples vs Playwright file) ---- 60 let ofd: i64 = sys_openat_wr("knowledge/fetched/wiki_web_dev_sovereign.txt\x00" as *u8, 0x1A4) 61 if ofd > 0 { sys_write(ofd, txt, tn); sys_close(ofd) } 62 63 // ---- Team's deterministic extract over the SOVEREIGN text ---- 64 let out: *i64 = sys_mmap(64) as *i64 65 let facts: i64 = ce_extract(txt, tn, out) 66 67 rr_puts("SOVEREIGN RESEARCH EXTRACT (no Playwright; team fetch+render+extract):\n raw_bytes=" as *u8); rr_putn(rn) 68 rr_puts("\n header_bytes=" as *u8); rr_putn(bo) 69 rr_puts("\n html_body_bytes=" as *u8); rr_putn(body_len) 70 rr_puts("\n rendered_text_bytes=" as *u8); rr_putn(tn) 71 rr_puts("\n sentences=" as *u8); rr_putn(out[0]) 72 rr_puts("\n fact_claims(number+unit)=" as *u8); rr_putn(out[1]) 73 rr_puts("\n first_claim_value=" as *u8); rr_putn(out[2]) 74 rr_puts("\n (FETCH now sovereign; Playwright = benchmark only)\n" as *u8) 75 sys_exit(0) 76 return 0 77}