tinet-data-foundry-cli 0.1.5__tar.gz → 0.1.8__tar.gz

{tinet_data_foundry_cli-0.1.5 → tinet_data_foundry_cli-0.1.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: tinet-data-foundry-cli
-Version: 0.1.5
+Version: 0.1.8
 Summary: Tinet Data Foundry 取数 CLI（docs / catalog / query / schema）
 Author: Tinet
 License: MIT
@@ -55,7 +55,16 @@ tdf query \
 | `ends_with` | 左模糊，匹配后缀 | `{"op":"ends_with","field":"name","value":"公司"}` |
 | `contains` | 左右模糊 | `{"op":"contains","field":"name","value":"科技"}` |
 
-分区表可以不传分区字段。服务端按上海时区补默认闭区间：按日为当天，按月为当月，按年为当年，并在响应 `data.partitionWindow` 里返回实际 `field/from/to`。显式 `between` 时，按日不超过 30 天，按月不超过 12 个月，按年不超过 3 年，否则 422 `PARTITION_RANGE_INVALID`。
+仅按日分区表必须带分区条件。未传分区字段时，服务端按上海时区补当天闭区间，并在响应 `data.partitionWindow` 里返回实际 `field/from/to`。显式 `between` 不超过 30 天，否则 422 `PARTITION_RANGE_INVALID`。按月、按年分区字段不是必填条件。
+
+ODS 可以查询和过滤 `_source_ds`，值是来源数据源编码。编码清单：
+
+```bash
+tdf datasource-codes
+tdf datasource-codes --code nboss
+```
+
+`_df_pt` 是通用查询字段，表示分区查询时间，写在 `fields` 或 `filter` 里，只有分区表有这一列。建议范围：按日查当天，`between` 不超过 30 天；按月的值是每月 1 日，建议查当月 1 日，跨月不超过 12 个月；按年查当年。列上的 `suggestedRange` 给出当前建议的 `from`/`to`。
 
 若 pip 提示 scripts 目录不在 PATH，把该目录加入 PATH，或用 `python3 -m tinet_data_foundry_cli`。
 
@@ -87,12 +96,25 @@ tdf catalog --insecure
 | `tdf docs rev` | `GET /docs/credentials/{ak}?format=rev` |
 | `tdf docs pull` | 按 `copy.catalog.policy` 只拉取变化的 `part=public` 或 `part=policy` |
 | `tdf catalog` | `GET /data-api/v1/catalog` |
+| `tdf datasource-codes` | `GET /data-api/v1/datasource-codes` |
 | `tdf query` | `POST /data-api/v1/datasets/{id}/query` |
 | `tdf schema` | `GET /data-api/v1/datasets/{id}/schema` |
 
 文档站点默认 `https://data-foundry.tinetcloud.com`，可用 `DF_DOCS_BASE` 覆盖。`tdf docs` 只需要 `DF_AK`。缓存位于 `~/.cache/tinet-data-foundry/docs/<AK 哈希>/`，可用 `DF_DOCS_CACHE` 改根目录。stdout 非终端时 `tdf catalog` 改为 JSON。
 
-Agent / Skill 传查询体用文件，避免 shell 转义：
+### 业务 Skill 锁定数据集
+
+`tdf docs pull` 把公共目录与授权策略缓存到本机。业务 Skill 自行检索和阅读字段，不要把 `public.json` 或 `policy.json` 读进上下文：
+
+```bash
+tdf docs pull
+tdf docs search --query "<检索词>"
+tdf docs describe --dataset "DATASET_ID"
+```
+
+`search` 默认只返回有表权限的数据集，表名或中文名精确匹配排在前面，最多 20 条。`describe` 只返回该表已授权字段、分区粒度和脱敏规则。需要与网关当前结构核对时再用 `tdf schema --dataset "DATASET_ID"`。不要用不带 `--keyword` 的 `tdf catalog` 拉取整份目录。
+
+DataFoundry 查询 Skill 只执行已经写好的查询。查询体用文件传入，避免 shell 转义：
 
 ```bash
 tdf query --dataset "DATASET_ID" --body-file query.json

{tinet_data_foundry_cli-0.1.5 → tinet_data_foundry_cli-0.1.8}/README.md

@@ -38,7 +38,16 @@ tdf query \
 | `ends_with` | 左模糊，匹配后缀 | `{"op":"ends_with","field":"name","value":"公司"}` |
 | `contains` | 左右模糊 | `{"op":"contains","field":"name","value":"科技"}` |
 
-分区表可以不传分区字段。服务端按上海时区补默认闭区间：按日为当天，按月为当月，按年为当年，并在响应 `data.partitionWindow` 里返回实际 `field/from/to`。显式 `between` 时，按日不超过 30 天，按月不超过 12 个月，按年不超过 3 年，否则 422 `PARTITION_RANGE_INVALID`。
+仅按日分区表必须带分区条件。未传分区字段时，服务端按上海时区补当天闭区间，并在响应 `data.partitionWindow` 里返回实际 `field/from/to`。显式 `between` 不超过 30 天，否则 422 `PARTITION_RANGE_INVALID`。按月、按年分区字段不是必填条件。
+
+ODS 可以查询和过滤 `_source_ds`，值是来源数据源编码。编码清单：
+
+```bash
+tdf datasource-codes
+tdf datasource-codes --code nboss
+```
+
+`_df_pt` 是通用查询字段，表示分区查询时间，写在 `fields` 或 `filter` 里，只有分区表有这一列。建议范围：按日查当天，`between` 不超过 30 天；按月的值是每月 1 日，建议查当月 1 日，跨月不超过 12 个月；按年查当年。列上的 `suggestedRange` 给出当前建议的 `from`/`to`。
 
 若 pip 提示 scripts 目录不在 PATH，把该目录加入 PATH，或用 `python3 -m tinet_data_foundry_cli`。
 
@@ -70,12 +79,25 @@ tdf catalog --insecure
 | `tdf docs rev` | `GET /docs/credentials/{ak}?format=rev` |
 | `tdf docs pull` | 按 `copy.catalog.policy` 只拉取变化的 `part=public` 或 `part=policy` |
 | `tdf catalog` | `GET /data-api/v1/catalog` |
+| `tdf datasource-codes` | `GET /data-api/v1/datasource-codes` |
 | `tdf query` | `POST /data-api/v1/datasets/{id}/query` |
 | `tdf schema` | `GET /data-api/v1/datasets/{id}/schema` |
 
 文档站点默认 `https://data-foundry.tinetcloud.com`，可用 `DF_DOCS_BASE` 覆盖。`tdf docs` 只需要 `DF_AK`。缓存位于 `~/.cache/tinet-data-foundry/docs/<AK 哈希>/`，可用 `DF_DOCS_CACHE` 改根目录。stdout 非终端时 `tdf catalog` 改为 JSON。
 
-Agent / Skill 传查询体用文件，避免 shell 转义：
+### 业务 Skill 锁定数据集
+
+`tdf docs pull` 把公共目录与授权策略缓存到本机。业务 Skill 自行检索和阅读字段，不要把 `public.json` 或 `policy.json` 读进上下文：
+
+```bash
+tdf docs pull
+tdf docs search --query "<检索词>"
+tdf docs describe --dataset "DATASET_ID"
+```
+
+`search` 默认只返回有表权限的数据集，表名或中文名精确匹配排在前面，最多 20 条。`describe` 只返回该表已授权字段、分区粒度和脱敏规则。需要与网关当前结构核对时再用 `tdf schema --dataset "DATASET_ID"`。不要用不带 `--keyword` 的 `tdf catalog` 拉取整份目录。
+
+DataFoundry 查询 Skill 只执行已经写好的查询。查询体用文件传入，避免 shell 转义：
 
 ```bash
 tdf query --dataset "DATASET_ID" --body-file query.json

{tinet_data_foundry_cli-0.1.5 → tinet_data_foundry_cli-0.1.8}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "tinet-data-foundry-cli"
-version = "0.1.5"
+version = "0.1.8"
 description = "Tinet Data Foundry 取数 CLI（docs / catalog / query / schema）"
 readme = "README.md"
 requires-python = ">=3.10"

tinet_data_foundry_cli-0.1.8/src/tinet_data_foundry_cli/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.1.8"

{tinet_data_foundry_cli-0.1.5 → tinet_data_foundry_cli-0.1.8}/src/tinet_data_foundry_cli/cli.py

@@ -82,6 +82,56 @@ def _ok_data(parsed: Any) -> Any:
     return parsed.get("data") if isinstance(parsed, dict) else parsed
 
 
+def _print_datasource_codes(rows: List[Mapping[str, Any]]) -> None:
+    if not rows:
+        print("当前凭证没有可查询的数据源编码。")
+        return
+    cols = [
+        ("code", "编码", 24),
+        ("name", "名称", 28),
+        ("id", "id", 36),
+    ]
+    header = "  ".join(_cell(title, w) for _, title, w in cols)
+    print(header)
+    print("-" * len(header))
+    for row in rows:
+        print("  ".join(_cell(row.get(key), w) for key, _, w in cols))
+
+
+def cmd_datasource_codes(args: argparse.Namespace) -> int:
+    base, ak, sk = credentials(args.endpoint)
+    query = {
+        "code": args.code or "",
+        "keyword": args.keyword or "",
+    }
+    try:
+        _, parsed, _ = request(
+            "GET",
+            base,
+            "/data-api/v1/datasource-codes",
+            ak,
+            sk,
+            query=query,
+            timeout=args.timeout,
+            insecure=args.insecure,
+            dry_run=args.dry_run,
+        )
+    except ApiError as ex:
+        _exit_api(ex)
+    except TransportError as ex:
+        _exit_transport(ex)
+    fmt = args.format
+    if fmt == "table" and not sys.stdout.isatty():
+        fmt = "json"
+    if args.dry_run or fmt == "json":
+        _print_json(parsed)
+        return 0
+    data = _ok_data(parsed) or {}
+    rows = data.get("list") if isinstance(data, dict) else []
+    _print_datasource_codes(list(rows or []))
+    return 0
+
+
 def cmd_catalog(args: argparse.Namespace) -> int:
     base, ak, sk = credentials(args.endpoint)
     query = {
@@ -249,6 +299,13 @@ def build_parser() -> argparse.ArgumentParser:
     c.add_argument("--format", choices=["table", "json"], default="table")
     c.set_defaults(func=cmd_catalog)
 
+    codes = sub.add_parser("datasource-codes", help="列出当前凭证可查询数据集关联的数据源编码")
+    _add_insecure(codes, overlay=True)
+    codes.add_argument("--code", help="按编码精确匹配，忽略大小写")
+    codes.add_argument("--keyword", help="按编码或名称包含匹配")
+    codes.add_argument("--format", choices=["table", "json"], default="table")
+    codes.set_defaults(func=cmd_datasource_codes)
+
     q = sub.add_parser("query", help="查询 Dataset")
     _add_insecure(q, overlay=True)
     q.add_argument("--dataset", required=True)
@@ -259,8 +316,10 @@ def build_parser() -> argparse.ArgumentParser:
         "--filter",
         help=(
             "filter AST JSON。字符串字段可 starts_with（右模糊）、ends_with（左模糊）、contains（左右模糊）。"
-            "分区表省略分区字段时，按日补当天、按月补当月、按年补当年；"
-            "显式 between 按日不超过 30 天、按月不超过 12 个月、按年不超过 3 年"
+            "仅按日分区必填：省略分区字段时按上海时区补当天，显式 between 不超过 30 天。"
+            "按月、按年分区字段不是必填条件。"
+            "_df_pt 是通用查询字段，表示分区查询时间：按日建议当天（不超过 30 天），"
+            "按月建议当月 1 日（不超过 12 个月），按年建议当年"
         ),
     )
     q.add_argument("--sort", help="sort JSON 数组")
@@ -277,9 +336,13 @@ def build_parser() -> argparse.ArgumentParser:
     )
     q.set_defaults(func=cmd_query)
 
-    d = sub.add_parser("docs", help="核对数据集开发文档修订号并缓存 public/policy")
+    d = sub.add_parser("docs", help="核对数据集开发文档修订号，并在本地缓存中检索数据集")
     _add_insecure(d, overlay=True)
-    d.add_argument("action", choices=["rev", "pull"])
+    d.add_argument("action", choices=["rev", "pull", "search", "describe"])
+    d.add_argument("--query", help="search：按名称、描述、表名或业务域检索本地缓存。表名精确匹配优先")
+    d.add_argument("--dataset", help="describe：输出指定数据集的已授权字段")
+    d.add_argument("--all", action="store_true", help="search：包含无表权限的数据集")
+    d.add_argument("--limit", type=int, default=20, help="search：最多返回条数，默认 20")
     d.add_argument("--docs-base", help="覆盖文档站点。默认 DF_DOCS_BASE 或 https://data-foundry.tinetcloud.com")
     d.set_defaults(func=cmd_docs)
 

{tinet_data_foundry_cli-0.1.5 → tinet_data_foundry_cli-0.1.8}/src/tinet_data_foundry_cli/docs.py

@@ -47,6 +47,15 @@ def run_docs(args: Any) -> int:
     ak = access_key()
     base = docs_base(getattr(args, "docs_base", None))
     folder = cache_dir(ak)
+    if args.action == "search":
+        return _search(
+            folder,
+            str(getattr(args, "query", "") or ""),
+            bool(getattr(args, "all", False)),
+            int(getattr(args, "limit", None) or 20),
+        )
+    if args.action == "describe":
+        return _describe(folder, str(getattr(args, "dataset", "") or ""))
     url = docs_url(base, ak, "rev")
     if args.dry_run:
         _print({
@@ -64,6 +73,114 @@ def run_docs(args: Any) -> int:
     return _pull(folder, base, ak, args.timeout, args.insecure)
 
 
+def _cached_docs(folder: Path) -> tuple[Dict[str, Any], Dict[str, Any]]:
+    public = _read_json(folder / "public.json")
+    policy = _read_json(folder / "policy.json")
+    if not public or not policy:
+        raise SystemExit("本地文档缓存不存在，请先运行 tdf docs pull")
+    return public, policy
+
+
+def _search(folder: Path, query: str, include_denied: bool, limit: int = 20) -> int:
+    public, policy = _cached_docs(folder)
+    needle = query.strip().casefold()
+    if not needle:
+        raise SystemExit("docs search 需要 --query")
+    if limit < 1:
+        raise SystemExit("docs search --limit 必须大于 0")
+    grants = _index(policy)
+    ranked: List[tuple] = []
+    for dataset in public.get("datasets") or []:
+        if not isinstance(dataset, dict):
+            continue
+        grant = grants.get(str(dataset.get("id") or ""), {})
+        if not include_denied and not grant.get("tableGranted"):
+            continue
+        rank = _match_rank(dataset, needle)
+        if rank is None:
+            continue
+        ranked.append((rank, str(dataset.get("name") or ""), {
+            "id": dataset.get("id"),
+            "name": dataset.get("name"),
+            "chineseName": dataset.get("chineseName"),
+            "description": dataset.get("description"),
+            "businessDomainName": dataset.get("businessDomainName"),
+            "layer": dataset.get("layer"),
+            "schemaVersion": dataset.get("schemaVersion"),
+            "tableGranted": bool(grant.get("tableGranted")),
+        }))
+    ranked.sort(key=lambda row: (row[0], row[1]))
+    shown = ranked[:limit]
+    _print({
+        "query": query,
+        "total": len(ranked),
+        "truncated": len(ranked) > limit,
+        "datasets": [row[2] for row in shown],
+    })
+    return 0
+
+
+def _match_rank(dataset: Dict[str, Any], needle: str) -> Optional[int]:
+    identity = [str(dataset.get(key) or "").casefold() for key in ("id", "name", "chineseName")]
+    if needle in identity:
+        return 0
+    if any(value.startswith(needle) for value in identity if value):
+        return 1
+    if any(needle in value for value in identity if value):
+        return 2
+    extra = " ".join(str(dataset.get(key) or "") for key in (
+        "tableComment", "description", "businessDomainName",
+    )).casefold()
+    if needle in extra:
+        return 3
+    return None
+
+
+def _describe(folder: Path, dataset_id: str) -> int:
+    public, policy = _cached_docs(folder)
+    if not dataset_id.strip():
+        raise SystemExit("docs describe 需要 --dataset")
+    datasets = _index(public)
+    dataset = datasets.get(dataset_id)
+    if not dataset:
+        raise SystemExit(f"缓存中没有数据集 {dataset_id}；先执行 tdf docs search --query <关键词>")
+    grant = _index(policy).get(dataset_id, {})
+    rules = {
+        str(col.get("name")): col
+        for col in grant.get("columns") or []
+        if isinstance(col, dict) and col.get("granted") and col.get("name")
+    }
+    columns = []
+    for column in dataset.get("columns") or []:
+        if not isinstance(column, dict) or str(column.get("name")) not in rules:
+            continue
+        name = str(column["name"])
+        columns.append({
+            "name": name,
+            "type": column.get("type"),
+            "comment": column.get("comment"),
+            "pk": column.get("pk"),
+            "partition": column.get("partition"),
+            "partitionGrain": column.get("partitionGrain"),
+            "partitionRequired": column.get("partitionRequired"),
+            "suggestedRange": column.get("suggestedRange"),
+            "outputRule": rules[name].get("outputRule"),
+        })
+    _print({
+        "id": dataset.get("id"),
+        "name": dataset.get("name"),
+        "chineseName": dataset.get("chineseName"),
+        "description": dataset.get("description"),
+        "businessDomainName": dataset.get("businessDomainName"),
+        "schemaVersion": dataset.get("schemaVersion"),
+        "primaryKey": dataset.get("primaryKey"),
+        "partitionFields": dataset.get("partitionFields"),
+        "tableGranted": bool(grant.get("tableGranted")),
+        "columns": columns,
+    })
+    return 0
+
+
 def _pull(folder: Path, base: str, ak: str, timeout: float, insecure: bool) -> int:
     local = _read_json(folder / "rev.json")
     local_etag = str(local.get("etag") or "")

{tinet_data_foundry_cli-0.1.5 → tinet_data_foundry_cli-0.1.8}/src/tinet_data_foundry_cli.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: tinet-data-foundry-cli
-Version: 0.1.5
+Version: 0.1.8
 Summary: Tinet Data Foundry 取数 CLI（docs / catalog / query / schema）
 Author: Tinet
 License: MIT
@@ -55,7 +55,16 @@ tdf query \
 | `ends_with` | 左模糊，匹配后缀 | `{"op":"ends_with","field":"name","value":"公司"}` |
 | `contains` | 左右模糊 | `{"op":"contains","field":"name","value":"科技"}` |
 
-分区表可以不传分区字段。服务端按上海时区补默认闭区间：按日为当天，按月为当月，按年为当年，并在响应 `data.partitionWindow` 里返回实际 `field/from/to`。显式 `between` 时，按日不超过 30 天，按月不超过 12 个月，按年不超过 3 年，否则 422 `PARTITION_RANGE_INVALID`。
+仅按日分区表必须带分区条件。未传分区字段时，服务端按上海时区补当天闭区间，并在响应 `data.partitionWindow` 里返回实际 `field/from/to`。显式 `between` 不超过 30 天，否则 422 `PARTITION_RANGE_INVALID`。按月、按年分区字段不是必填条件。
+
+ODS 可以查询和过滤 `_source_ds`，值是来源数据源编码。编码清单：
+
+```bash
+tdf datasource-codes
+tdf datasource-codes --code nboss
+```
+
+`_df_pt` 是通用查询字段，表示分区查询时间，写在 `fields` 或 `filter` 里，只有分区表有这一列。建议范围：按日查当天，`between` 不超过 30 天；按月的值是每月 1 日，建议查当月 1 日，跨月不超过 12 个月；按年查当年。列上的 `suggestedRange` 给出当前建议的 `from`/`to`。
 
 若 pip 提示 scripts 目录不在 PATH，把该目录加入 PATH，或用 `python3 -m tinet_data_foundry_cli`。
 
@@ -87,12 +96,25 @@ tdf catalog --insecure
 | `tdf docs rev` | `GET /docs/credentials/{ak}?format=rev` |
 | `tdf docs pull` | 按 `copy.catalog.policy` 只拉取变化的 `part=public` 或 `part=policy` |
 | `tdf catalog` | `GET /data-api/v1/catalog` |
+| `tdf datasource-codes` | `GET /data-api/v1/datasource-codes` |
 | `tdf query` | `POST /data-api/v1/datasets/{id}/query` |
 | `tdf schema` | `GET /data-api/v1/datasets/{id}/schema` |
 
 文档站点默认 `https://data-foundry.tinetcloud.com`，可用 `DF_DOCS_BASE` 覆盖。`tdf docs` 只需要 `DF_AK`。缓存位于 `~/.cache/tinet-data-foundry/docs/<AK 哈希>/`，可用 `DF_DOCS_CACHE` 改根目录。stdout 非终端时 `tdf catalog` 改为 JSON。
 
-Agent / Skill 传查询体用文件，避免 shell 转义：
+### 业务 Skill 锁定数据集
+
+`tdf docs pull` 把公共目录与授权策略缓存到本机。业务 Skill 自行检索和阅读字段，不要把 `public.json` 或 `policy.json` 读进上下文：
+
+```bash
+tdf docs pull
+tdf docs search --query "<检索词>"
+tdf docs describe --dataset "DATASET_ID"
+```
+
+`search` 默认只返回有表权限的数据集，表名或中文名精确匹配排在前面，最多 20 条。`describe` 只返回该表已授权字段、分区粒度和脱敏规则。需要与网关当前结构核对时再用 `tdf schema --dataset "DATASET_ID"`。不要用不带 `--keyword` 的 `tdf catalog` 拉取整份目录。
+
+DataFoundry 查询 Skill 只执行已经写好的查询。查询体用文件传入，避免 shell 转义：
 
 ```bash
 tdf query --dataset "DATASET_ID" --body-file query.json

{tinet_data_foundry_cli-0.1.5 → tinet_data_foundry_cli-0.1.8}/tests/test_cli.py

@@ -17,13 +17,21 @@ class CliParserTest(unittest.TestCase):
         self.assertIn("starts_with", help_text)
         self.assertIn("ends_with", help_text)
         self.assertIn("contains", help_text)
+        self.assertIn("仅按日分区必填", help_text)
         self.assertIn("30", help_text)
         self.assertIn("当天", help_text)
+        self.assertIn("按月、按年分区字段不是必填条件", help_text)
+        self.assertIn("_df_pt", help_text)
+        self.assertIn("分区查询时间", help_text)
+        self.assertNotIn("按月补当月", help_text)
 
     def test_subcommands(self):
         parser = build_parser()
         catalog = parser.parse_args(["catalog", "--format", "json"])
         self.assertEqual(catalog.cmd, "catalog")
+        codes = parser.parse_args(["datasource-codes", "--code", "nboss"])
+        self.assertEqual(codes.cmd, "datasource-codes")
+        self.assertEqual(codes.code, "nboss")
         query = parser.parse_args(["query", "--dataset", "ds_1", "--page-no", "1", "--page-size", "10"])
         self.assertEqual(query.cmd, "query")
         catalog_k = parser.parse_args(["catalog", "--insecure"])

tinet_data_foundry_cli-0.1.8/tests/test_docs.py

@@ -0,0 +1,101 @@
+import json
+import tempfile
+import unittest
+from io import StringIO
+from pathlib import Path
+from unittest.mock import patch
+
+from tinet_data_foundry_cli.docs import _describe, _search, summarize
+
+
+class DocsSummaryTest(unittest.TestCase):
+    def test_search_and_describe_only_print_authorized_dataset_content(self):
+        with tempfile.TemporaryDirectory() as tmp, patch("sys.stdout", new_callable=StringIO) as out:
+            folder = Path(tmp)
+            (folder / "public.json").write_text(json.dumps({"datasets": [{
+                "id": "ds_bill", "name": "monthly_bill", "chineseName": "月账单",
+                "description": "京东云消费", "columns": [
+                    {"name": "amount", "type": "decimal", "comment": "金额"},
+                    {"name": "secret", "type": "text", "comment": "无权限"},
+                    {"name": "_df_pt", "type": "date", "comment": "分区查询时间",
+                     "suggestedRange": {"from": "2026-10-01", "to": "2026-10-01", "grain": "month"}},
+                ],
+            }]}), encoding="utf-8")
+            (folder / "policy.json").write_text(json.dumps({"datasets": [{
+                "id": "ds_bill", "tableGranted": True,
+                "columns": [
+                    {"name": "amount", "granted": True, "outputRule": "PLAIN"},
+                    {"name": "_df_pt", "granted": True, "outputRule": "PLAIN"},
+                ],
+            }]}), encoding="utf-8")
+            self.assertEqual(_search(folder, "京东云", False), 0)
+            search = json.loads(out.getvalue())
+            self.assertEqual(search["datasets"][0]["id"], "ds_bill")
+            self.assertEqual(search["total"], 1)
+            self.assertFalse(search["truncated"])
+            out.seek(0)
+            out.truncate(0)
+            self.assertEqual(_describe(folder, "ds_bill"), 0)
+            described = json.loads(out.getvalue())
+            self.assertEqual([col["name"] for col in described["columns"]], ["amount", "_df_pt"])
+            self.assertEqual(described["columns"][1]["suggestedRange"]["from"], "2026-10-01")
+
+    def test_search_prefers_exact_name_and_limits_rows(self):
+        with tempfile.TemporaryDirectory() as tmp, patch("sys.stdout", new_callable=StringIO) as out:
+            folder = Path(tmp)
+            datasets = [
+                {"id": "ds_desc", "name": "other", "description": "bill detail"},
+                {"id": "ds_exact", "name": "bill", "chineseName": "账单"},
+                {"id": "ds_more", "name": "bill_extra", "description": "bill"},
+            ]
+            grants = [{"id": row["id"], "tableGranted": True, "columns": []} for row in datasets]
+            (folder / "public.json").write_text(json.dumps({"datasets": datasets}), encoding="utf-8")
+            (folder / "policy.json").write_text(json.dumps({"datasets": grants}), encoding="utf-8")
+            self.assertEqual(_search(folder, "bill", False, 2), 0)
+            search = json.loads(out.getvalue())
+            self.assertEqual([row["id"] for row in search["datasets"]], ["ds_exact", "ds_more"])
+            self.assertEqual(search["total"], 3)
+            self.assertTrue(search["truncated"])
+
+    def test_summarize_catalog_and_policy(self):
+        old_public = {"datasets": [{"id": "a", "schemaVersion": "1"}, {"id": "b", "schemaVersion": "1"}]}
+        new_public = {"datasets": [{"id": "b", "schemaVersion": "2"}, {"id": "c", "schemaVersion": "1"}]}
+        old_policy = {"datasets": [{"id": "a", "tableGranted": True}, {"id": "b", "tableGranted": False}]}
+        new_policy = {"datasets": [{"id": "b", "tableGranted": True}, {"id": "c", "tableGranted": False}]}
+        lines = summarize(old_public, new_public, old_policy, new_policy)
+        self.assertIn("新增数据集 c", lines)
+        self.assertIn("下线数据集 a", lines)
+        self.assertIn("schemaVersion 变化 b", lines)
+        self.assertIn("表权限变化 b 有", lines)
+
+    def test_pull_fetches_only_policy_when_catalog_unchanged(self):
+        from tinet_data_foundry_cli import docs as docs_mod
+
+        rev = {"copy": "c", "catalog": "k", "policy": "p2", "etag": "c.k.p2"}
+        calls = []
+
+        def fake_get(url, etag, timeout, insecure):
+            calls.append(url)
+            if "format=rev" in url:
+                return 200, rev
+            if "part=policy" in url:
+                return 200, {"credentialType": "PERSONAL", "datasets": [{"id": "a", "tableGranted": True, "columns": []}]}
+            raise AssertionError(url)
+
+        with tempfile.TemporaryDirectory() as tmp:
+            folder = Path(tmp)
+            (folder / "rev.json").write_text(json.dumps({
+                "copy": "c", "catalog": "k", "policy": "p1", "etag": "c.k.p1",
+            }), encoding="utf-8")
+            (folder / "public.json").write_text(json.dumps({"datasets": [{"id": "a"}]}), encoding="utf-8")
+            with patch.object(docs_mod, "_get_json", fake_get):
+                code = docs_mod._pull(folder, "https://docs.example", "DFAK_x", 5, False)
+            self.assertEqual(code, 0)
+            self.assertTrue(any("part=policy" in url for url in calls))
+            self.assertFalse(any("part=public" in url for url in calls))
+            saved = json.loads((folder / "rev.json").read_text(encoding="utf-8"))
+            self.assertEqual(saved["policy"], "p2")
+
+
+if __name__ == "__main__":
+    unittest.main()

tinet_data_foundry_cli-0.1.5/src/tinet_data_foundry_cli/__init__.py

@@ -1 +0,0 @@
-__version__ = "0.1.5"

tinet_data_foundry_cli-0.1.5/tests/test_docs.py

@@ -1,52 +0,0 @@
-import json
-import tempfile
-import unittest
-from pathlib import Path
-from unittest.mock import patch
-
-from tinet_data_foundry_cli.docs import summarize
-
-
-class DocsSummaryTest(unittest.TestCase):
-    def test_summarize_catalog_and_policy(self):
-        old_public = {"datasets": [{"id": "a", "schemaVersion": "1"}, {"id": "b", "schemaVersion": "1"}]}
-        new_public = {"datasets": [{"id": "b", "schemaVersion": "2"}, {"id": "c", "schemaVersion": "1"}]}
-        old_policy = {"datasets": [{"id": "a", "tableGranted": True}, {"id": "b", "tableGranted": False}]}
-        new_policy = {"datasets": [{"id": "b", "tableGranted": True}, {"id": "c", "tableGranted": False}]}
-        lines = summarize(old_public, new_public, old_policy, new_policy)
-        self.assertIn("新增数据集 c", lines)
-        self.assertIn("下线数据集 a", lines)
-        self.assertIn("schemaVersion 变化 b", lines)
-        self.assertIn("表权限变化 b 有", lines)
-
-    def test_pull_fetches_only_policy_when_catalog_unchanged(self):
-        from tinet_data_foundry_cli import docs as docs_mod
-
-        rev = {"copy": "c", "catalog": "k", "policy": "p2", "etag": "c.k.p2"}
-        calls = []
-
-        def fake_get(url, etag, timeout, insecure):
-            calls.append(url)
-            if "format=rev" in url:
-                return 200, rev
-            if "part=policy" in url:
-                return 200, {"credentialType": "PERSONAL", "datasets": [{"id": "a", "tableGranted": True, "columns": []}]}
-            raise AssertionError(url)
-
-        with tempfile.TemporaryDirectory() as tmp:
-            folder = Path(tmp)
-            (folder / "rev.json").write_text(json.dumps({
-                "copy": "c", "catalog": "k", "policy": "p1", "etag": "c.k.p1",
-            }), encoding="utf-8")
-            (folder / "public.json").write_text(json.dumps({"datasets": [{"id": "a"}]}), encoding="utf-8")
-            with patch.object(docs_mod, "_get_json", fake_get):
-                code = docs_mod._pull(folder, "https://docs.example", "DFAK_x", 5, False)
-            self.assertEqual(code, 0)
-            self.assertTrue(any("part=policy" in url for url in calls))
-            self.assertFalse(any("part=public" in url for url in calls))
-            saved = json.loads((folder / "rev.json").read_text(encoding="utf-8"))
-            self.assertEqual(saved["policy"], "p2")
-
-
-if __name__ == "__main__":
-    unittest.main()