Files
qingjian/tools/release/data-fetch.sh
T
yenharvey 4a5d60952b build: 安装包与数据包不再带含章·知微
三个桌面平台都是有含章·通变就不加载知微,而通变一直随包,知微从来用不上,白占 53MB。
用户自己放进数据目录 models/hanzhang-zhiwei/ 的仍然认。data-fetch 兼容还带着知微的 data-v7。
2026-10-21 15:06:25 +08:00

90 lines
3.3 KiB
Bash
Executable File
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
#!/usr/bin/env bash
# 按 tools/release/data.lock 下载并校验单包产品数据;旧锁仍可读取分离的模型附件。
#
# tools/release/data-fetch.sh # 下载 + 校验 + 解开
# tools/release/data-fetch.sh --verify # 只校验 target/release-data/ 里已下载的文件
set -euo pipefail
ROOT="$(cd "$(dirname "$0")/../.." && pwd)"
LOCK="$ROOT/tools/release/data.lock"
OUT="$ROOT/target/release-data"
REPO="qingjian-team/qingjian"
ASSETS=(qingjian-data.tar.gz)
cd "$ROOT"
[[ -f "$LOCK" ]] || { echo "缺少 $LOCK" >&2; exit 1; }
lock_value() { sed -nE "s/^$1 *= *//p" "$LOCK" | head -1 | tr -d '[:space:]'; }
sha256() { if command -v sha256sum >/dev/null 2>&1; then sha256sum "$1"; else shasum -a 256 "$1"; fi | cut -d' ' -f1; }
TAG="$(lock_value tag)"
[[ -n "$TAG" ]] || { echo "$LOCK 里没有 tag" >&2; exit 1; }
LEGACY_MODEL="$(lock_value model.qjm)"
if [[ -n "$LEGACY_MODEL" ]]; then
ASSETS+=(model.qjm)
[[ -n "$(lock_value model-p2c.qjm)" ]] && ASSETS+=(model-p2c.qjm)
fi
mkdir -p "$OUT"
if [[ "${1:-}" != "--verify" ]]; then
for f in "${ASSETS[@]}"; do
rm -f "$OUT/$f"
if command -v gh >/dev/null 2>&1 && gh auth status >/dev/null 2>&1; then
gh release download "$TAG" --repo "$REPO" --pattern "$f" --dir "$OUT"
else
curl -fL --retry 3 -o "$OUT/$f" "https://github.com/$REPO/releases/download/$TAG/$f"
fi
done
fi
for f in "${ASSETS[@]}"; do
expected="$(lock_value "$f")"
actual="$(sha256 "$OUT/$f")"
[[ -n "$expected" && "$actual" == "$expected" ]] || { echo "$f 与 data.lock 不符($TAG):期望 $expected,实际 $actual" >&2; exit 1; }
echo "$f $actual"
done
[[ "${1:-}" == "--verify" ]] && exit 0
if [[ -n "$LEGACY_MODEL" ]]; then
mkdir -p data/generated data/models/hanzhang-zhiwei
tar -xzf "$OUT/qingjian-data.tar.gz" -C data/generated
cp "$OUT/model.qjm" data/models/hanzhang-zhiwei/model.qjm
MODEL_SOURCE=data/models/hanzhang-zhiwei/model.qjm
P2C_MODEL_SOURCE=""
if [[ -n "$(lock_value model-p2c.qjm)" ]]; then
mkdir -p data/models/hanzhang-tongbian
cp "$OUT/model-p2c.qjm" data/models/hanzhang-tongbian/model.qjm
P2C_MODEL_SOURCE=data/models/hanzhang-tongbian/model.qjm
fi
else
tar -xzf "$OUT/qingjian-data.tar.gz" -C "$ROOT"
# data-v7 及以前的包里还带着含章·知微,之后的不带
MODEL_SOURCE=""
[[ -f data/models/hanzhang-zhiwei/hanzhang-zhiwei-small.qjm ]] && MODEL_SOURCE=data/models/hanzhang-zhiwei/hanzhang-zhiwei-small.qjm
P2C_MODEL_SOURCE=data/models/hanzhang-tongbian/hanzhang-tongbian-small.qjm
[[ -f "$P2C_MODEL_SOURCE" && -f data/generated/dict.qj ]] || {
echo "产品数据包缺少词库或含章·通变($TAG)" >&2; exit 1;
}
fi
# 解出来的 mtime 比 checkout 出来的 TSV 旧,bundle.sh 会以为要重打
find data/generated -type f -exec touch {} +
[[ -n "$MODEL_SOURCE" ]] && touch "$MODEL_SOURCE"
[[ -n "$P2C_MODEL_SOURCE" ]] && touch "$P2C_MODEL_SOURCE"
echo "产品数据 $TAG 已就位"
if [[ -n "${GITHUB_ENV:-}" ]]; then
{
echo "DATA_TAG=$TAG"
echo "DATA_SHA256=$(lock_value qingjian-data.tar.gz)"
if [[ -n "$MODEL_SOURCE" ]]; then
echo "MODEL_SHA256=$(sha256 "$MODEL_SOURCE")"
else
echo "MODEL_SHA256="
fi
if [[ -n "$P2C_MODEL_SOURCE" ]]; then
echo "P2C_MODEL_SHA256=$(sha256 "$P2C_MODEL_SOURCE")"
else
echo "P2C_MODEL_SHA256="
fi
} >> "$GITHUB_ENV"
fi