sino-bc-cli 1.8.24 → 1.8.26

package/dist/skills/custom/biz-line-strategy-generate/scripts/cli_json.py

@@ -142,39 +142,10 @@ def parse_json_text(raw: str):
         return None
 
 
-def load_json_file(path: Path, drop_page_rows: bool = False):
-    """从文件解析 JSON。优先 json.load，避免先整文件读成 str 再 loads。
-
-    drop_page_rows：分页信封在 object_hook 里记下行数并丢掉行 list，调用方拿不到整页行。
-    解析失败只读文件头，不把后面埋着的 JSON 当成功回包。
-    """
+def load_json_file(path: Path):
     if not path.exists():
         return None
-
-    def hook(obj):
-        if isinstance(obj.get("data"), list) and obj.get("total") is not None:
-            obj["_page_got"] = len(obj["data"])
-            obj["data"] = []
-        if isinstance(obj.get("result"), list) and obj.get("total") is not None:
-            obj["_page_got"] = obj.get("_page_got") or len(obj["result"])
-            obj["result"] = []
-        return obj
-
-    try:
-        with path.open(encoding="utf-8") as handle:
-            if drop_page_rows:
-                return json.load(handle, object_hook=hook)
-            return json.load(handle)
-    except json.JSONDecodeError:
-        return parse_json_text(read_text_head(path))
-
-
-def read_text_head(path: Path, limit: int = 2000) -> str:
-    """失败摘要只读文件头部，不把整页 stdout 再读进内存。"""
-    if not path.exists():
-        return ""
-    with path.open(encoding="utf-8", errors="replace") as handle:
-        return handle.read(limit)
+    return parse_json_text(path.read_text(encoding="utf-8"))
 
 
 def unwrap_page(envelope: dict) -> tuple[list, int]:
@@ -200,23 +171,8 @@ def unwrap_page(envelope: dict) -> tuple[list, int]:
     return [], 0
 
 
-def take_page_stats(envelope: dict) -> tuple[int, int]:
-    """取出本页行数与 total，并丢掉行 list，让整页数据可被回收。"""
-    rows, total = unwrap_page(envelope)
-    got = len(rows)
-    data = envelope.get("data")
-    if isinstance(data, dict):
-        if isinstance(data.get("data"), list):
-            data["data"] = []
-        if isinstance(data.get("result"), list):
-            data["result"] = []
-    elif isinstance(data, list):
-        envelope["data"] = []
-    return got, total
-
-
-def exec_cli_to_file(cmd: list[str], out_path: Path) -> int:
-    """两个输出流直写本次文件，只返回退出码，不把 stdout 读回内存。"""
+def exec_cli_to_file(cmd: list[str], out_path: Path) -> tuple[int, str]:
+    """两个输出流直写本次文件；取消时等待子进程退出后才允许清理。"""
     out_path.parent.mkdir(parents=True, exist_ok=True)
     cancel = getattr(_local, "cancel", None)
     if cancel and cancel.is_set():
@@ -239,7 +195,8 @@ def exec_cli_to_file(cmd: list[str], out_path: Path) -> int:
                 except subprocess.TimeoutExpired:
                     proc.kill()
                     proc.wait()
-    return proc.returncode
+    raw = out_path.read_text(encoding="utf-8")
+    return proc.returncode, raw
 
 
 def list_query_body(*, name: str = "", biz_id: int = 0) -> dict:
@@ -258,7 +215,7 @@ def list_query_body(*, name: str = "", biz_id: int = 0) -> dict:
     return body
 
 
-def run_cli(args: list[str], body: dict | None, out_path: Path, *, drop_rows: bool = False) -> dict:
+def run_cli(args: list[str], body: dict | None, out_path: Path) -> dict:
     out_path.parent.mkdir(parents=True, exist_ok=True)
     cmd = [*resolve_cli(), *args, "--format", "json"]
     if body is not None:
@@ -267,12 +224,11 @@ def run_cli(args: list[str], body: dict | None, out_path: Path, *, drop_rows: bo
             body_path = out_path.with_suffix(".body.json")
         body_path.write_text(json.dumps(body, ensure_ascii=False), encoding="utf-8")
         cmd.extend(["--data", json.dumps(body, ensure_ascii=False)])
-    rc = exec_cli_to_file(cmd, out_path)
-    parsed = load_json_file(out_path, drop_page_rows=drop_rows)
-    stderr_path = out_path.with_suffix(".stderr.txt")
+    rc, raw = exec_cli_to_file(cmd, out_path)
+    parsed = parse_json_text(raw.strip())
+    stderr = out_path.with_suffix(".stderr.txt").read_text(encoding="utf-8", errors="replace")
     if not isinstance(parsed, dict) or rc != 0 or parsed.get("ok") is not True:
-        stderr = read_text_head(stderr_path)
-        error = parse_json_text(stderr) or stderr or parsed or read_text_head(out_path) or "无有效 CLI 回包"
+        error = parse_json_text(stderr) or stderr or parsed or raw or "无有效 CLI 回包"
         parsed = {"ok": False, "error": error, "rc": rc}
     return parsed
 
@@ -293,46 +249,24 @@ def optional_num(row, key):
         return None
 
 
-METRIC_FORMULAS = {
-    "CPI": ("spend_usd", "installs", 1),
-    "CPC": ("spend_usd", "clicks", 1),
-    "CPM": ("spend_usd", "impressions", 1000),
-    "ROAS": ("purchase_value", "spend_usd", 1),
-    "CTR": ("clicks", "impressions", 100),
-}
-
-
-def new_metric_acc(metric="CPI"):
-    a, b, scale = METRIC_FORMULAS[metric]
-    return {"a": a, "b": b, "scale": scale, "num_a": 0.0, "num_b": 0.0, "ok": True, "seen": False}
-
-
-def add_metric_acc(acc, row) -> None:
-    if not acc["ok"]:
-        return
-    acc["seen"] = True
-    left, right = optional_num(row, acc["a"]), optional_num(row, acc["b"])
-    if left is None or right is None:
-        acc["ok"] = False
-        return
-    acc["num_a"] += left
-    acc["num_b"] += right
-
-
-def finish_metric_acc(acc):
-    if not acc["ok"] or not acc["seen"]:
-        return None
-    return acc["num_a"] / acc["num_b"] * acc["scale"] if acc["num_b"] > 0 else None
-
-
 def metric_value(rows, metric="CPI"):
     """同一粒度先合计基础指标再计算，任一缺失不当作零补齐。只扫一遍，可接迭代器。"""
-    acc = new_metric_acc(metric)
+    formulas = {"CPI": ("spend_usd", "installs", 1), "CPC": ("spend_usd", "clicks", 1),
+                "CPM": ("spend_usd", "impressions", 1000), "ROAS": ("purchase_value", "spend_usd", 1),
+                "CTR": ("clicks", "impressions", 100)}
+    a, b, scale = formulas[metric]
+    seen = False
+    num_a = num_b = 0.0
     for row in rows or ():
-        add_metric_acc(acc, row)
-        if not acc["ok"]:
+        seen = True
+        left, right = optional_num(row, a), optional_num(row, b)
+        if left is None or right is None:
             return None
-    return finish_metric_acc(acc)
+        num_a += left
+        num_b += right
+    if not seen:
+        return None
+    return num_a / num_b * scale if num_b > 0 else None
 
 
 def core_rule(text):

package/dist/skills/custom/biz-line-strategy-generate/scripts/fetch_data.py

@@ -15,7 +15,7 @@ SCRIPTS = Path(__file__).resolve().parent
 if str(SCRIPTS) not in sys.path:
     sys.path.insert(0, str(SCRIPTS))
 
-from cli_json import list_query_body, run_cli, unwrap_page, take_page_stats, require_ok, report_progress
+from cli_json import list_query_body, run_cli, unwrap_page, require_ok, report_progress
 from datetime import date
 
 # 维度/成效 size=1000 时，TikTok 出价按天明细可到 4～5 万行，旧上限 30 页会截断
@@ -57,19 +57,16 @@ def fetch_paged(shortcut: str, base_body: dict, stem: str, bl_dir: Path) -> int:
     while True:
         body = {**base_body, "page": page, "size": 1000}
         out = bl_dir / f"{stem}.p{page}.out.json"
-        env = run_cli(["biz-line", shortcut], body, out, drop_rows=True)
+        env = run_cli(["biz-line", shortcut], body, out)
         require_ok(env, stem)
         data = env.get("data")
         if not isinstance(data, dict) or not isinstance(data.get("data"), list) or not isinstance(data.get("total"), int) or data["total"] < 0:
             raise RuntimeError(f"{stem}: 分页信封缺少有效 data/total")
-        got = data.get("_page_got")
-        if isinstance(got, int):
-            current_total = data["total"]
-        else:
-            got, current_total = take_page_stats(env)
+        rows, current_total = unwrap_page(env)
         if page > 1 and current_total != total:
             raise RuntimeError(f"{stem}: 分页期间总数变化，请重新拉取")
         total = current_total
+        got = len(rows)
         accumulated += got
         counts.append(got)
         report_progress(f"{stem} page={page} got={got} acc={accumulated} total={total}")
@@ -77,7 +74,7 @@ def fetch_paged(shortcut: str, base_body: dict, stem: str, bl_dir: Path) -> int:
             raise RuntimeError(f"{stem}: 累计行数超过 total")
         if accumulated == total:
             break
-        if got == 0 or page >= PAGE_CAP:
+        if not rows or page >= PAGE_CAP:
             raise RuntimeError(f"{stem}: 分页不完整 {accumulated}/{total}")
         page += 1
     (bl_dir / f"{stem}.pages.json").write_text(