rust-pyfunc 0.76.9__tar.gz → 0.76.14__tar.gz

{rust_pyfunc-0.76.9 → rust_pyfunc-0.76.14}/.github/workflows/CI.yml

@@ -40,8 +40,9 @@ jobs:
         uses: PyO3/maturin-action@v1
         with:
           target: ${{ matrix.platform.target }}
-          args: --release --out dist --find-interpreter
+          args: --release --out dist --find-interpreter --no-default-features
           sccache: 'true'
+          before-script-windows: choco install cmake
           manylinux: auto
           before-script-linux: ${{ matrix.platform.target == 'aarch64' && 'pip install cmake' || '' }}
       - name: Upload wheels
@@ -179,6 +180,8 @@ jobs:
       - uses: actions/setup-python@v5
         with:
           python-version: '3.9'
+      - name: Install dependencies
+        run: pip install pandas numpy
       - name: Install rust_pyfunc
         run: pip install rust_pyfunc --no-index --find-links dist
       - name: Test import

{rust_pyfunc-0.76.9 → rust_pyfunc-0.76.14}/.gitignore

@@ -197,6 +197,9 @@ Thumbs.db
 .windsurf/
 .crush/
 AGENT.md
+CLAUDE.md
+AGENTS.md
+CRUSH.md
 
 # Logs
 *.log
@@ -234,3 +237,5 @@ docs/theme_feature_expansion_explanation.md
 .vtcode/skills/INDEX.md
 .vtcode/terminals/INDEX.md
 .mindfs/sessions/session-list.db
+.maturin_zig/*
+dist_versions/*

{rust_pyfunc-0.76.9 → rust_pyfunc-0.76.14}/Cargo.lock

@@ -2181,7 +2181,7 @@ checksum = "082f11ffa03bbef6c2c6ea6bea1acafaade2fd9050ae0234ab44a2153742b058"
 
 [[package]]
 name = "rust_pyfunc"
-version = "0.76.9"
+version = "0.76.14"
 dependencies = [
  "arrow",
  "base64 0.21.7",

{rust_pyfunc-0.76.9 → rust_pyfunc-0.76.14}/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "rust_pyfunc"
-version = "0.76.9"
+version = "0.76.14"
 edition = "2021"
 description = "A collection of high-performance Python functions implemented in Rust"
 readme = "README.md"
@@ -46,8 +46,12 @@ log = "0.4"
 once_cell = "1.19"
 csv = "1.3"
 ndarray-npy = "0.8"
-hdf5-metno = { version = "0.12", features = ["static", "zlib"] }
+hdf5-metno = { version = "0.12", features = ["static", "zlib"], optional = true }
 zstd = "0.13"
+
+[features]
+default = ["hdf5"]
+hdf5 = ["hdf5-metno"]
 # Unix-specific dependencies for fork mode
 [target.'cfg(unix)'.dependencies]
 nix = { version = "0.27", features = ["process", "signal", "fs"] }

{rust_pyfunc-0.76.9 → rust_pyfunc-0.76.14}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: rust_pyfunc
-Version: 0.76.9
+Version: 0.76.14
 Classifier: Programming Language :: Rust
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Programming Language :: Python :: Implementation :: PyPy

rust_pyfunc-0.76.14/README.md

@@ -0,0 +1,33 @@
+# rust_pyfunc
+
+用 Rust 写的高性能 Python 库。专治各种计算密集场景：金融数据分析、时间序列处理、统计计算——Python 搞不定的性能问题，这里管。
+
+## 安装
+
+```shell
+pip install rust_pyfunc
+```
+
+## 使用
+
+```python
+import rust_pyfunc as rp
+```
+
+## 开发
+
+```bash
+pip install maturin
+maturin develop --release
+```
+
+## 项目结构
+
+```
+src/          # Rust 源码
+python/       # Python 包 & 类型提示
+```
+
+## License
+
+MIT

{rust_pyfunc-0.76.9 → rust_pyfunc-0.76.14}/alter.sh

@@ -22,8 +22,9 @@ conda activate chenzongwei311
 # 3. 切换到你的项目目录 (如果当前目录不是)
 # cd /path/to/your/maturin/project
 
-# 4. 运行 maturin develop
-maturin develop --release
+# 4. 运行 maturin develop (使用 mold 链接器 + sccache 编译缓存加速)
+export PATH="/home/chenzongwei/.local/bin:$PATH"
+/home/chenzongwei/.local/bin/mold -run maturin develop --release
 
 # 如果需要，你也可以在脚本结束时停用环境
 # conda deactivate

rust_pyfunc-0.76.14/freeze_version.sh

@@ -0,0 +1,210 @@
+#!/bin/bash
+# freeze_version.sh — 将当前项目打包为版本锁定的独立 wheel
+#
+# 用法:
+#   ./freeze_version.sh                                    # 从 Cargo.toml 自动读取版本号
+#   ./freeze_version.sh 0_76_8                             # 手动指定版本标签
+#   ./freeze_version.sh --manylinux2014                    # 构建 glibc 2.17+ 兼容 wheel
+#   ./freeze_version.sh 0_76_8 --manylinux_2_31            # 构建 glibc 2.31+ 兼容 wheel
+#
+# 输出:
+#   dist_versions/rust_pyfunc_<version>-<ver>-cp*-cp*-linux_x86_64.whl
+#   加上 --manylinux* 参数后:
+#   dist_versions/rust_pyfunc_<version>-<ver>-cp*-cp*-manylinux_*_x86_64.whl
+#
+# 安装后使用:
+#   import rust_pyfunc_0_76_8 as rp
+
+set -euo pipefail
+
+PROJECT_DIR="$(cd "$(dirname "$0")" && pwd)"
+cd "$PROJECT_DIR"
+
+# ── 0. 确保 zig 可用（用于 --manylinux* 交叉链接）─────────────
+ZIG_BIN="$(python3 -c 'import ziglang, os; print(os.path.join(os.path.dirname(ziglang.__file__), "zig"))' 2>/dev/null || true)"
+ZIG_DIR="${PROJECT_DIR}/.maturin_zig"
+if [ -n "$ZIG_BIN" ] && [ -x "$ZIG_BIN" ]; then
+    mkdir -p "${ZIG_DIR}/bin"
+    ln -sf "$ZIG_BIN" "${ZIG_DIR}/bin/zig"
+    # 创建 cargo 包装器：把 cargo → cargo-zigbuild
+    # 注意：
+    #   - CARGO_BUILD_TARGET 在包装器内部设置，这样 maturin 读不到带后缀的 triple
+    #   - CARGO 指向真正的 cargo，避免 cargo-zigbuild 子调用又回到包装器形成死循环
+    CARGO_ZIGBUILD="/home/chenzongwei/.conda/envs/chenzongwei311/bin/cargo-zigbuild"
+    REAL_CARGO="$(command -v cargo)"
+    cat > "${ZIG_DIR}/bin/cargo" << CARGO_WRAPPER
+#!/bin/bash
+export CARGO_BUILD_TARGET="x86_64-unknown-linux-gnu.2.31"
+export CARGO="${REAL_CARGO}"
+exec ${CARGO_ZIGBUILD} "\$@"
+CARGO_WRAPPER
+    chmod +x "${ZIG_DIR}/bin/cargo"
+fi
+
+# ── 1. 确定版本标签与兼容性参数 ────────────────────────────────
+VERSION_TAG=""
+COMPAT_TAG=""
+
+for arg in "$@"; do
+    case "$arg" in
+        --manylinux*)
+            COMPAT_TAG="${arg#--}"
+            ;;
+        *)
+            VERSION_TAG="$arg"
+            ;;
+    esac
+done
+
+if [ -z "$VERSION_TAG" ]; then
+    VERSION_TAG=$(grep '^version = ' Cargo.toml | head -1 | sed 's/.*"\(.*\)".*/\1/' | tr '.' '_')
+fi
+
+PACKAGE_NAME="rust_pyfunc_${VERSION_TAG}"
+TMPDIR="/tmp/build_${PACKAGE_NAME}_$$"
+OUTPUT_DIR="${PROJECT_DIR}/dist_versions"
+PYTHON_SRC="${PROJECT_DIR}/python/rust_pyfunc"
+
+echo "========================================"
+echo " 打包版本: ${PACKAGE_NAME}"
+echo " 兼容目标: ${COMPAT_TAG:-系统原生 (glibc $(ldd --version 2>&1 | head -1 | grep -oP '\d+\.\d+$'))}"
+echo " 临时目录: ${TMPDIR}"
+echo " 输出目录: ${OUTPUT_DIR}"
+echo "========================================"
+
+# ── 2. 创建临时构建目录 ────────────────────────────────────────
+rm -rf "$TMPDIR"
+mkdir -p "$TMPDIR"
+
+# ── 3. 复制 Python 源码，并重命名包目录 ────────────────────────
+cp -r "$PYTHON_SRC" "${TMPDIR}/${PACKAGE_NAME}"
+
+# ── 4. 修补 import 语句（绝对自导入 → 相对导入） ────────────────
+#    这些文件用了 from rust_pyfunc.xxx import yyy（绝对导入），
+#    包名改成 rust_pyfunc_0_76_8 后必须改为相对导入 from .xxx import yyy
+
+cd "${TMPDIR}/${PACKAGE_NAME}"
+
+echo "[patch] __init__.py ..."
+# __init__.py 第 3 行: from rust_pyfunc.rust_pyfunc import *  →  from .rust_pyfunc import *
+sed -i 's/^from rust_pyfunc\.rust_pyfunc import \*/from .rust_pyfunc import */' __init__.py
+# __init__.py 第 5 行: from rust_pyfunc import *  →  删除（冗余）
+sed -i '/^from rust_pyfunc import \*$/d' __init__.py
+# __init__.py 第 6 行: from rust_pyfunc.rolling_future import ...  →  from .rolling_future import ...
+sed -i 's/^from rust_pyfunc\.rolling_future import/from .rolling_future import/' __init__.py
+# __init__.py 第 7 行: from rust_pyfunc.rolling_past import ...  →  from .rolling_past import ...
+sed -i 's/^from rust_pyfunc\.rolling_past import/from .rolling_past import/' __init__.py
+
+echo "[patch] rolling_future.py ..."
+sed -i 's/^from rust_pyfunc\.rust_pyfunc import/from .rust_pyfunc import/' rolling_future.py
+
+echo "[patch] rolling_past.py ..."
+sed -i 's/^from rust_pyfunc\.rust_pyfunc import/from .rust_pyfunc import/' rolling_past.py
+
+echo "[patch] treevisual.py ..."
+sed -i 's/^from rust_pyfunc\.rust_pyfunc import/from .rust_pyfunc import/' treevisual.py
+
+echo "[patch] pandas_corrwith.py ..."
+sed -i 's/^from rust_pyfunc\.rust_pyfunc import/from .rust_pyfunc import/' pandas_corrwith.py
+
+# 确认修补结果
+echo ""
+echo "修补后的导入语句："
+grep -n '^from \.rust_pyfunc import' __init__.py rolling_future.py rolling_past.py treevisual.py pandas_corrwith.py 2>/dev/null || true
+echo ""
+
+# ── 5. 复制 Rust 源码 + 辅助文件 ────────────────────────────────
+cp -r "${PROJECT_DIR}/src" "$TMPDIR/src"
+cp "${PROJECT_DIR}/Cargo.toml" "$TMPDIR/Cargo.toml"
+[ -f "${PROJECT_DIR}/Cargo.lock" ] && cp "${PROJECT_DIR}/Cargo.lock" "$TMPDIR/Cargo.lock" || true
+[ -f "${PROJECT_DIR}/rust-toolchain.toml" ] && cp "${PROJECT_DIR}/rust-toolchain.toml" "$TMPDIR/" || true
+cp "${PROJECT_DIR}/README.md" "$TMPDIR/README.md" 2>/dev/null || touch "$TMPDIR/README.md"
+
+# ── 6. 创建临时的 pyproject.toml ──────────────────────────────
+cat > "$TMPDIR/pyproject.toml" << PYPROJECT
+[build-system]
+requires = ["maturin>=1.3,<2.0"]
+build-backend = "maturin"
+
+[project]
+name = "${PACKAGE_NAME}"
+requires-python = ">=3.8"
+dynamic = ["version"]
+
+[tool.maturin]
+features = ["pyo3/extension-module"]
+python-source = "."
+module-name = "${PACKAGE_NAME}.rust_pyfunc"
+PYPROJECT
+
+echo "pyproject.toml:"
+cat "$TMPDIR/pyproject.toml"
+echo ""
+
+# ── 7. 构建 wheel ─────────────────────────────────────────────
+
+build_for_python() {
+    local python_bin="$1"
+    local compat="$2"
+    local label="$3"
+
+    cd "$TMPDIR"
+
+    local opts="--release --interpreter $python_bin"
+    if [ -n "$compat" ]; then
+        opts="$opts --compatibility $compat --no-default-features"
+    fi
+
+    echo "━━━ 构建 [$label] ━━━"
+    echo "解释器: $($python_bin --version 2>&1)"
+    echo "选项: maturin build $opts"
+    echo ""
+
+    # 需要 zig 做交叉链接
+    if [ -n "$compat" ]; then
+        # 用 cargo-zigbuild 替代 cargo —— 自动用 zig 编译和链接，
+        # 产生目标 glibc 版本的二进制
+        export PATH="${ZIG_DIR}/bin:$PATH"  # 内有 cargo 包装器（含 CARGO/CARGO_BUILD_TARGET 设置）
+        echo "  [zig] 通过 cargo-zigbuild 构建 (glibc 2.31)"
+    fi
+    CARGO_TARGET_DIR="${PROJECT_DIR}/target" maturin build $opts 2>&1
+
+    echo ""
+    echo "[$label] 构建完成"
+    echo ""
+}
+
+# 第一次构建：当前 Python，遵循 --manylinux* 参数
+CURRENT_PYTHON=$(command -v python3)
+build_for_python "$CURRENT_PYTHON" "$COMPAT_TAG" "当前环境"
+
+# 第二次构建：Python 3.9 + manylinux_2_31
+PY39="/opt/anaconda3/bin/python3.9"
+if [ -x "$PY39" ]; then
+    build_for_python "$PY39" "manylinux_2_31" "Python 3.9 + manylinux_2_31"
+else
+    echo "[跳过] python3.9 未找到: ${PY39}"
+fi
+
+# ── 8. 收集产物 ────────────────────────────────────────────────
+mkdir -p "$OUTPUT_DIR"
+cp "${PROJECT_DIR}"/target/wheels/"${PACKAGE_NAME}"*.whl "$OUTPUT_DIR/"
+
+# ── 9. 清理 ────────────────────────────────────────────────────
+rm -rf "$TMPDIR"
+
+# ── 10. 输出结果 ───────────────────────────────────────────────
+echo ""
+echo "========================================"
+echo " ✅ 打包完成"
+echo "========================================"
+ls -lh "${OUTPUT_DIR}/${PACKAGE_NAME}"*.whl 2>/dev/null || echo "（wheel 文件未找到）"
+
+echo ""
+echo "安装命令："
+echo "  pip install ${OUTPUT_DIR}/${PACKAGE_NAME}-*.whl"
+echo ""
+echo "使用方式："
+echo "  import ${PACKAGE_NAME} as rp"
+echo "  rp.fast_rank(...)"
+echo ""

{rust_pyfunc-0.76.9 → rust_pyfunc-0.76.14}/python/rust_pyfunc/__init__.py

@@ -50,6 +50,7 @@ _LAZY_EXPORTS = {
     "PriceTreeViz": (".treevisual", "PriceTreeViz"),
     "haha": (".treevisual", "haha"),
     "treevisual": (".treevisual", None),
+    "read_minute_data": (".minute_data_reader", "read_minute_data"),
 }
 
 _LAZY_MODULE_EXPORTS = {

{rust_pyfunc-0.76.9 → rust_pyfunc-0.76.14}/python/rust_pyfunc/__init__.pyi

@@ -34,6 +34,7 @@ from .theme_cluster_factors import *
 from .theme_feature_expansion import *
 from .theme_feature_expansion_data import *
 from .trading_data_utils import *
+from .minute_data_reader import *
 
 
 
@@ -257,6 +258,7 @@ __all__ = [
     "read_trade",
     "read_market",
     "read_market_pair",
+    "read_minute_data",
     "get_features_factors_single",
     "get_features_factors",
     "get_features_names",

rust_pyfunc-0.76.14/python/rust_pyfunc/factor_task.py

@@ -0,0 +1,191 @@
+#!/usr/bin/env python3
+"""factor-task — 因子计算任务 CLI 工具
+
+通过 HTTP 与 factor_taskd 守护进程通信，管理因子计算任务。
+
+用法:
+  factor-task submit <script.py> [--n-jobs N] [--start-date YYYYMMDD] [--end-date YYYYMMDD]
+  factor-task list
+  factor-task show <id>
+  factor-task cancel <id>
+  factor-task adjust <id> <n_jobs>
+  factor-task log <id>
+"""
+from __future__ import annotations
+
+import argparse
+import json
+import os
+import sys
+import urllib.error
+import urllib.request
+
+DAEMON_URL = os.environ.get("FACTOR_TASK_URL", "http://127.0.0.1:9099")
+TIMEOUT = 10
+
+
+def _request(method: str, path: str, body: dict | None = None) -> dict:
+    url = f"{DAEMON_URL}{path}"
+    data = json.dumps(body).encode() if body else None
+    req = urllib.request.Request(url, data=data, method=method)
+    req.add_header("Content-Type", "application/json")
+    try:
+        with urllib.request.urlopen(req, timeout=TIMEOUT) as resp:
+            return json.loads(resp.read().decode())
+    except urllib.error.HTTPError as e:
+        err = e.read().decode()
+        print(f"❌ HTTP {e.code}: {err}", file=sys.stderr)
+        sys.exit(1)
+    except urllib.error.URLError as e:
+        print(f"❌ 连接失败 (确保 factor_taskd 正在运行): {e.reason}", file=sys.stderr)
+        sys.exit(1)
+
+
+def cmd_submit(args):
+    body = {
+        "script_path": os.path.abspath(args.script),
+        "n_jobs": args.n_jobs or 50,
+    }
+    if args.start_date:
+        body["start_date"] = args.start_date
+    if args.end_date:
+        body["end_date"] = args.end_date
+
+    resp = _request("POST", "/api/tasks", body)
+    print(f"✅ 任务已提交")
+    print(f"   任务 ID: {resp['id']}")
+    print(f"   名称:    {resp['name']}")
+    print(f"   状态:    {resp['status']}")
+    print(f"   查看:    factor-task show {resp['id']}")
+
+
+def cmd_list(args):
+    tasks = _request("GET", "/api/tasks")
+    if not tasks:
+        print("📭 暂无任务")
+        return
+    print(f"{'ID':>4}  {'名称':<20} {'状态':<12} {'n_jobs':<6} {'PID':<7} {'进度':<10}")
+    print("-" * 70)
+    for t in tasks:
+        progress = ""
+        if t.get("progress") and t["progress"].get("latest"):
+            latest = t["progress"]["latest"]
+            if "collected" in latest and "total_tasks" in latest:
+                pct = (
+                    latest["collected"] / latest["total_tasks"] * 100
+                    if latest["total_tasks"]
+                    else 0
+                )
+                progress = f"{latest['collected']}/{latest['total_tasks']}({pct:.0f}%)"
+        pid = t.get("pid") or ""
+        n_jobs = t.get("n_jobs") or ""
+        status = t.get("status", "unknown")
+        print(
+            f"{t['id']:>4}  {t.get('name',''):<20} {status:<12} {str(n_jobs):<6} {str(pid):<7} {progress:<10}"
+        )
+
+
+def cmd_show(args):
+    task = _request("GET", f"/api/tasks/{args.id}")
+    print(f"任务 ID:      {task['id']}")
+    print(f"名称:         {task.get('name', '-')}")
+    print(f"脚本路径:     {task.get('script_path', '-')}")
+    print(f"状态:         {task.get('status', '-')}")
+    print(f"PID:          {task.get('pid', '-')}")
+    print(f"n_jobs:       {task.get('n_jobs', '-')}")
+    print(f"起止日期:     {task.get('start_date', '-')} ~ {task.get('end_date', '-')}")
+    print(f"创建时间:     {task.get('created_at', '-')}")
+    print(f"开始时间:     {task.get('started_at', '-')}")
+    print(f"完成时间:     {task.get('finished_at', '-')}")
+    print(f"日志路径:     {task.get('log_path', '-')}")
+    print(f"运行中:       {task.get('active', False)}")
+    if task.get("error"):
+        print(f"错误:         {task['error']}")
+
+    progress = task.get("progress")
+    if progress and progress.get("records"):
+        records = progress["records"]
+        latest = progress.get("latest", {})
+        if "collected" in latest and "total_tasks" in latest:
+            pct = (
+                latest["collected"] / latest["total_tasks"] * 100
+                if latest["total_tasks"]
+                else 0
+            )
+            print(
+                f"进度:         {latest['collected']}/{latest['total_tasks']} ({pct:.1f}%)"
+            )
+        if len(records) > 1:
+            last = records[-1]
+            print(
+                f"最后备份:     {last.get('batch', '-')}/{last.get('total_batches', '-')}"
+            )
+
+
+def cmd_cancel(args):
+    resp = _request("POST", f"/api/tasks/{args.id}/cancel")
+    print(f"✅ 任务 {args.id} 已取消")
+
+
+def cmd_adjust(args):
+    body = {"n_jobs": args.n_jobs}
+    resp = _request("POST", f"/api/tasks/{args.id}/adjust-njobs", body)
+    print(f"✅ 任务 {args.id} n_jobs 已调整为 {args.n_jobs}")
+
+
+def cmd_log(args):
+    resp = _request("GET", f"/api/tasks/{args.id}/log")
+    log = resp.get("log", "")
+    if log:
+        print(log)
+    else:
+        print("📭 日志为空")
+
+
+def main():
+    parser = argparse.ArgumentParser(description="因子计算任务 CLI 工具")
+    parser.add_argument(
+        "--url", default=DAEMON_URL, help=f"daemon 地址 (默认: {DAEMON_URL})"
+    )
+    sub = parser.add_subparsers(dest="command")
+
+    p_submit = sub.add_parser("submit", help="提交新任务")
+    p_submit.add_argument("script", help="Python 脚本路径")
+    p_submit.add_argument("--n-jobs", type=int, default=None, help="并行数")
+    p_submit.add_argument("--start-date", type=int, help="起始日期 (YYYYMMDD)")
+    p_submit.add_argument("--end-date", type=int, help="结束日期 (YYYYMMDD)")
+    p_submit.set_defaults(func=cmd_submit)
+
+    p_list = sub.add_parser("list", help="列出所有任务")
+    p_list.set_defaults(func=cmd_list)
+
+    p_show = sub.add_parser("show", help="查看任务详情")
+    p_show.add_argument("id", type=int, help="任务 ID")
+    p_show.set_defaults(func=cmd_show)
+
+    p_cancel = sub.add_parser("cancel", help="取消任务")
+    p_cancel.add_argument("id", type=int, help="任务 ID")
+    p_cancel.set_defaults(func=cmd_cancel)
+
+    p_adjust = sub.add_parser("adjust", help="调节并行数")
+    p_adjust.add_argument("id", type=int, help="任务 ID")
+    p_adjust.add_argument("n_jobs", type=int, help="新的并行数")
+    p_adjust.set_defaults(func=cmd_adjust)
+
+    p_log = sub.add_parser("log", help="查看任务日志")
+    p_log.add_argument("id", type=int, help="任务 ID")
+    p_log.set_defaults(func=cmd_log)
+
+    args = parser.parse_args()
+    if not args.command:
+        parser.print_help()
+        sys.exit(1)
+
+    import factor_task  # noqa: F811
+    factor_task.DAEMON_URL = args.url
+
+    args.func(args)
+
+
+if __name__ == "__main__":
+    main()