code wiki / _hdl_build / nx_vizsla_capture_gate.nx
`nx_vizsla_capture_gate.nx` source

↩ module page · 183 lines · 8260 B
1// nx_vizsla_capture_gate.nx -- VIZSLA capture gate: the opt-in page->person extractor pulls a contact from a
2// loaded page's structured data (JSON-LD Person > og:title > <title>), strips the site suffix, and emits an
3// action-ready suggestion -- NEVER auto-adding. 100% sovereign: writes synthetic page fixtures, runs the
4// extractor, greps the suggestion. (The realm choice + confirm happen downstream; this is just the extractor.)
5//
6// Rows:
7//   1 jsonld-person    JSON-LD {"@type":"Person","name":"Ada Lovelace"} -> id=ada-lovelace name=Ada-Lovelace source=jsonld
8//   2 og-title         <meta og:title "Grace Hopper - Rear Admiral | LinkedIn"> -> id=grace-hopper source=og (suffix stripped)
9//   3 title-fallback   <title>Alan Turing | LinkedIn</title> -> id=alan-turing source=title (suffix stripped)
10//   4 priority-jsonld  JSON-LD + og both present -> JSON-LD wins (source=jsonld, NOT the og "Wrong Name")
11//   5 form-output      emits VIZSLA-CAPTURE-FORM id=ada-lovelace&name=Ada-Lovelace (action-ready body)
12//   6 neg-no-person    a page with no Person/og/title -> exit 1
13// Evidence: VIZSLA-CAPTURE-GATE -> stdout + knowledge/status/vizsla_gate.log; exit 0 iff 6/6.
14// license_tier: ORIGINAL
15import "nx_syscalls.nx"
16
17func og_slen(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } return n }
18func og_p(s: *u8) -> i64 { sys_write(1, s, og_slen(s)); return 0 }
19
20func og_cat(dst: *u8, off: i64, s: *u8) -> i64 {
  var i: i64 = 0
  while s[i] != (0 as u8) { dst[off + i] = s[i]; i = i + 1 }
  return off + i
24}
25
26func og_catn(dst: *u8, off: i64, v: i64) -> i64 {
  var o: i64 = off
  var m: i64 = v
  if m < 0 { dst[o] = 45 as u8; o = o + 1; 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 { dst[o + i] = t[k - 1 - i]; i = i + 1 }
  return o + k
37}
38
39func og_write(path: *u8, content: *u8) -> i64 {
  let fd: i64 = sys_openat_wr(path, 0x1a4)
  if fd < 0 { return 0 - 1 }
  sys_write(fd, content, og_slen(content))
  sys_close(fd)
  return 0
45}
46
47func og_readall(path: *u8, szout: *i64) -> *u8 {
  let fd: i64 = sys_openat_rd(path)
  if fd < 0 { szout[0] = 0 - 1; return 0 as *u8 }
  let sz: i64 = sys_lseek(fd, 0, 2)
  sys_lseek(fd, 0, 0)
  let buf: *u8 = sys_mmap(sz + 64)
  var got: i64 = 0
  var n: i64 = 1
  while n > 0 { n = sys_read(fd, (buf as i64 + got) as *u8, 65536); if n > 0 { got = got + n } }
  sys_close(fd)
  szout[0] = got
  return buf
59}
60
61func og_runv(elf: *u8, args: *i64, outpath: *u8) -> i64 {
  let pid: i64 = sys_fork()
  if pid == 0 {
      if (outpath as i64) != 0 {
          let ofd: i64 = sys_openat_wr(outpath, 0x1a4)
          if ofd >= 0 { sys_dup3(ofd, 1, 0); sys_dup3(ofd, 2, 0) }
      }
      let argv: *i64 = sys_mmap(128) as *i64
      argv[0] = elf as i64
      var i: i64 = 0
      var go: i64 = 1
      while go == 1 { if args[i] == 0 { go = 0 } else { argv[i + 1] = args[i]; i = i + 1 } }
      argv[i + 1] = 0
      let envp: *i64 = sys_mmap(16) as *i64
      envp[0] = 0
      sys_execve(elf, argv, envp)
      sys_exit(127)
  }
  let st: *i64 = sys_mmap(16) as *i64
  sys_wait4(pid, st, 0)
  let sig: i64 = st[0] & 0x7f
  if sig != 0 { return 128 + sig }
  return (st[0] >> 8) & 0xff
84}
85
86func og_has(path: *u8, needle: *u8) -> i64 {
  let szp: *i64 = sys_mmap(16) as *i64
  let b: *u8 = og_readall(path, szp)
  let sz: i64 = szp[0]
  let n: i64 = og_slen(needle)
  if sz < n { return 0 }
  var i: i64 = 0
  while i + n <= sz {
      var ok: i64 = 1
      var j: i64 = 0
      while j < n { if b[i + j] != needle[j] { ok = 0; j = n } else { j = j + 1 } }
      if ok == 1 { return 1 }
      i = i + 1
  }
  return 0
101}
102
103func og_row(name: *u8, pass: i64) -> i64 {
  og_p("ROW " as *u8); og_p(name)
  if pass == 1 { og_p(" PASS\n" as *u8) } else { og_p(" FAIL\n" as *u8) }
  return pass
107}
108
109func og_args(a1: *u8, a2: *u8, a3: *u8, a4: *u8, a5: *u8) -> *i64 {
  let a: *i64 = sys_mmap(64) as *i64
  a[0] = a1 as i64; a[1] = a2 as i64; a[2] = a3 as i64; a[3] = a4 as i64; a[4] = a5 as i64; a[5] = 0
  return a
113}
114
115func main(argc: i64, argv: *i64) -> i64 {
  og_p("=== VIZSLA CAPTURE GATE: opt-in page->person extractor (JSON-LD/og/title) ===\n" as *u8)
  var ob: *u8 = "/tmp/nx_vizsla_capture.sov.elf" as *u8
  if argc > 1 { ob = argv[1] as *u8 }
  let pr: i64 = sys_openat_rd(ob)
  if pr >= 0 { sys_close(pr) }
  else {
      og_p(" instrument missing -> rebuilding via durable runner\n" as *u8)
      og_runv("_offc/nx_sov_build_run.elf" as *u8, og_args("nx_vizsla_capture" as *u8, 0 as *u8, 0 as *u8, 0 as *u8, 0 as *u8), "/tmp/vzcap_rebuild.out" as *u8)
  }
125
  og_write("/tmp/cap_jsonld.html" as *u8, "<html><head><script type=\"application/ld+json\">{\"@context\":\"https://schema.org\",\"@type\":\"Person\",\"name\":\"Ada Lovelace\",\"jobTitle\":\"Mathematician\"}</script></head><body>x</body></html>" as *u8)
  og_write("/tmp/cap_og.html" as *u8, "<html><head><meta property=\"og:title\" content=\"Grace Hopper - Rear Admiral | LinkedIn\"></head><body></body></html>" as *u8)
  og_write("/tmp/cap_title.html" as *u8, "<html><head><title>Alan Turing | LinkedIn</title></head><body></body></html>" as *u8)
  og_write("/tmp/cap_pri.html" as *u8, "<html><head><script type=\"application/ld+json\">{\"@type\":\"Person\",\"name\":\"Ada Lovelace\"}</script><meta property=\"og:title\" content=\"Wrong Name\"></head></html>" as *u8)
  og_write("/tmp/cap_none.html" as *u8, "<html><head></head><body>just some text, nobody here</body></html>" as *u8)
131
  var pass: i64 = 0
  var r: i64 = 0
134
  og_runv(ob, og_args("extract" as *u8, "/tmp/cap_jsonld.html" as *u8, 0 as *u8, 0 as *u8, 0 as *u8), "/tmp/cap_r1.txt" as *u8)
  r = og_has("/tmp/cap_r1.txt" as *u8, "VIZSLA-CAPTURE id=ada-lovelace name=Ada-Lovelace source=jsonld" as *u8)
  pass = pass + og_row("jsonld-person" as *u8, r)
138
  og_runv(ob, og_args("extract" as *u8, "/tmp/cap_og.html" as *u8, 0 as *u8, 0 as *u8, 0 as *u8), "/tmp/cap_r2.txt" as *u8)
  r = og_has("/tmp/cap_r2.txt" as *u8, "id=grace-hopper name=Grace-Hopper source=og" as *u8)
  pass = pass + og_row("og-title" as *u8, r)
142
  og_runv(ob, og_args("extract" as *u8, "/tmp/cap_title.html" as *u8, 0 as *u8, 0 as *u8, 0 as *u8), "/tmp/cap_r3.txt" as *u8)
  r = og_has("/tmp/cap_r3.txt" as *u8, "id=alan-turing name=Alan-Turing source=title" as *u8)
  pass = pass + og_row("title-fallback" as *u8, r)
146
  og_runv(ob, og_args("extract" as *u8, "/tmp/cap_pri.html" as *u8, 0 as *u8, 0 as *u8, 0 as *u8), "/tmp/cap_r4.txt" as *u8)
  r = 0
  if og_has("/tmp/cap_r4.txt" as *u8, "source=jsonld" as *u8) == 1 { if og_has("/tmp/cap_r4.txt" as *u8, "Wrong" as *u8) == 0 { r = 1 } }
  pass = pass + og_row("priority-jsonld-wins" as *u8, r)
151
  r = og_has("/tmp/cap_r1.txt" as *u8, "VIZSLA-CAPTURE-FORM id=ada-lovelace&name=Ada-Lovelace" as *u8)
  pass = pass + og_row("form-output" as *u8, r)
154
  let rc6: i64 = og_runv(ob, og_args("extract" as *u8, "/tmp/cap_none.html" as *u8, 0 as *u8, 0 as *u8, 0 as *u8), "/tmp/cap_r6.txt" as *u8)
  r = 0
  if rc6 == 1 { r = 1 }
  pass = pass + og_row("neg-no-person" as *u8, r)
159
  let permil: i64 = (pass * 1000) / 6
  let logfd: i64 = sys_openat_append("knowledge/status/vizsla_gate.log" as *u8, 0x1a4)
  var fdi: i64 = 0
  while fdi < 2 {
      var fd: i64 = 1
      if fdi == 1 { fd = logfd }
      if fd > 0 {
          let line: *u8 = sys_mmap(256)
          var o: i64 = 0
          o = og_cat(line, o, "VIZSLA-CAPTURE-GATE epoch=" as *u8)
          o = og_catn(line, o, sys_now_realtime_sec())
          o = og_cat(line, o, " rows=6 pass=" as *u8)
          o = og_catn(line, o, pass)
          o = og_cat(line, o, " permil=" as *u8)
          o = og_catn(line, o, permil)
          if pass == 6 { o = og_cat(line, o, " verdict=GREEN\n" as *u8) } else { o = og_cat(line, o, " verdict=RED\n" as *u8) }
          sys_write(fd, line, o)
      }
      fdi = fdi + 1
  }
  if logfd > 0 { sys_close(logfd) }
  if pass == 6 { return 0 }
  return 1
183}
nx_vizsla_capture_gate.nx source

`nx_vizsla_capture_gate.nx` source
