3e88f9fbc9
The sentence helper is an optional download (Settings → Sentence helper): Gemma 4 E2B in Google's web build, run by Google's LiteRT-LM web runtime on the graphics chip, in a worker. ✨ beside the message turns the words chosen into three sentences, fitting what the partner said if it was typed in, or suggests four replies when no words are chosen yet. The suggestions take the word predictions' place, at their height, so no cell moves (SPEC §3); the person chooses one and speaks it. Nothing is spoken by itself, and nothing typed leaves the device. - ai_helper.rs: the catalogue (aac-ai-index/1), prompts, whether a device can run it (a WebGPU hardware adapter, never the processor), download, continue, remove and re-check, and suggestions that a later request supersedes. The model's parts (64 MiB, SHA-256 checked) and the one runtime build this browser runs live in the model store; the worker streams the parts from there into GPU memory. - third_party/litert-lm: @litert-lm/core 0.17.1 vendored and pinned (npm integrity, each file's size and SHA-256), with our worker bundled by esbuild 0.28.2; the runtime's CDN default is emptied. - tools/ai: build_ai.py (the pinned model → parts and catalogue), check_packs.py, languages.json (English, Spanish, German and Hindi ready; Swahili not yet), and eval/: the 20 exchanges, a harness that runs the app's own prompts (a test keeps them word for word), and the judged answers. - Measured: Chrome on an M5 Max, 0.4-0.7 s an answer. A real iPad Air (M3) in Safari 27: 101 s to download and check 2 GB, about 80 s for the first start, 5.9 s for later starts, 0.8-1.4 s an answer. Co-Authored-By: Claude Opus 5.5 <[email protected]>
80 lines
2.9 KiB
Python
80 lines
2.9 KiB
Python
#!/usr/bin/env python3
|
|
"""Check a sentence-helper directory before a site build ships it.
|
|
|
|
python3 tools/ai/check_packs.py DIR
|
|
|
|
DIR must hold what tools/ai/build_ai.py built for the committed catalogue
|
|
(data/ai/index.json): its index.json points at index.<sha12>.json, whose
|
|
bytes are the committed catalogue's, and every part and licence file that
|
|
catalogue names is there with its exact size and SHA-256. Otherwise this exits
|
|
non-zero and says what is wrong, so a build never ships a model that
|
|
disagrees with the catalogue, or that is incomplete.
|
|
|
|
Standard library only; Python 3.9 or later (macOS's own python3).
|
|
"""
|
|
|
|
import hashlib
|
|
import json
|
|
import sys
|
|
from pathlib import Path
|
|
|
|
HERE = Path(__file__).resolve().parent
|
|
COMMITTED = HERE.parents[1] / "data" / "ai" / "index.json"
|
|
SHOW = 10
|
|
|
|
|
|
def sha256(path):
|
|
h = hashlib.sha256()
|
|
with open(path, "rb") as f:
|
|
for chunk in iter(lambda: f.read(1 << 22), b""):
|
|
h.update(chunk)
|
|
return h.hexdigest()
|
|
|
|
|
|
def main(argv):
|
|
if len(argv) != 2:
|
|
print(__doc__, file=sys.stderr)
|
|
return 2
|
|
root = Path(argv[1])
|
|
problems = []
|
|
try:
|
|
pointer = json.loads((root / "index.json").read_text(encoding="utf-8"))
|
|
name = pointer["current"]
|
|
cat_bytes = (root / name).read_bytes()
|
|
except (OSError, ValueError, KeyError) as e:
|
|
print(f"✗ {root}: no readable index.json and catalogue ({e})", file=sys.stderr)
|
|
return 1
|
|
if cat_bytes != COMMITTED.read_bytes():
|
|
problems.append(f"{name} is not the committed catalogue ({COMMITTED})")
|
|
if name != f"index.{hashlib.sha256(cat_bytes).hexdigest()[:12]}.json":
|
|
problems.append(f"{name} is not named by its contents")
|
|
cat = json.loads(cat_bytes)
|
|
refs = []
|
|
for model in cat.get("models", {}).values():
|
|
refs += [(p["url"], p["bytes"], p["sha256"]) for p in model.get("parts", [])]
|
|
lf = model.get("licence_file")
|
|
if lf:
|
|
data = (root / lf).read_bytes() if (root / lf).is_file() else b""
|
|
refs.append((lf, len(data) if data else -1, hashlib.sha256(data).hexdigest() if data else ""))
|
|
if not (root / "LICENSES.md").is_file():
|
|
problems.append("LICENSES.md is missing")
|
|
for rel, size, digest in refs:
|
|
f = root / rel
|
|
if not f.is_file():
|
|
problems.append(f"missing {rel}")
|
|
elif f.stat().st_size != size:
|
|
problems.append(f"{rel} is {f.stat().st_size} bytes, the catalogue says {size}")
|
|
elif sha256(f) != digest:
|
|
problems.append(f"{rel} does not match its SHA-256")
|
|
if problems:
|
|
for p in problems[:SHOW]:
|
|
print(f"✗ {p}", file=sys.stderr)
|
|
if len(problems) > SHOW:
|
|
print(f" … and {len(problems) - SHOW} more", file=sys.stderr)
|
|
return 1
|
|
print(f"✓ {root}: {len(refs)} files match {name}")
|
|
return 0
|
|
|
|
|
|
if __name__ == "__main__":
|
|
sys.exit(main(sys.argv))
|