fix: 对齐规格解析 Mock 契约 (#257)

2026-08-17 18:14:50 +08:00
parent e5ac327d35
commit 1009e54531
2 changed files with 299 additions and 26 deletions
@@ -2,6 +2,8 @@

 import hashlib
 import json
+import re
+import unicodedata
 from copy import deepcopy
 from dataclasses import dataclass
 from datetime import datetime, timezone
@@ -195,8 +197,12 @@ class MockAdminGateway(AdminGateway):
    ) -> SpecResolutionReceipt:
        with self._lock:
            self._raise_forced_error()
-            task = self._validate_submission_target(task_id, idempotency_key)
-            self._validate_spec_observation(task, idempotency_key, observation)
+            self._validate_spec_observation(
+                task_id, idempotency_key, observation
+            )
+            task = self._validate_spec_resolution_target(
+                task_id, observation
+            )
            fingerprint = self._fingerprint(
                "spec_resolution", task_id, observation
            )
@@ -233,42 +239,59 @@ class MockAdminGateway(AdminGateway):

    @staticmethod
    def _validate_spec_observation(
-        task: AdminTask,
+        task_id: str,
        idempotency_key: str,
        observation: Mapping[str, Any],
    ) -> None:
-        if task.task_type.value != "purchase":
+        try:
+            encoded = json.dumps(
+                observation,
+                ensure_ascii=False,
+                separators=(",", ":"),
+            ).encode("utf-8")
+        except (TypeError, ValueError) as exc:
            raise AdminGatewayError(
-                "TASK_NOT_PURCHASE", "当前任务不是采购任务", False
+                "INVALID_BODY", "规格解析请求不是有效 JSON", False
+            ) from exc
+        if len(encoded) > 64 * 1024:
+            raise AdminGatewayError(
+                "INVALID_BODY", "规格解析请求超过 64 KiB", False
            )
        if observation.get("schema_version") != 1:
            raise AdminGatewayError(
                "INVALID_SPEC_RESOLUTION_SCHEMA", "规格解析版本无效", False
            )
-        if observation.get("task_version") != task.version:
-            raise AdminGatewayError(
-                "TASK_VERSION_CONFLICT", "规格解析任务版本不一致", False
-            )
-        payload = task.payload
-        if observation.get("pdd_goods_id") != payload.get("goods_id"):
-            raise AdminGatewayError(
-                "PDD_GOODS_MISMATCH", "规格解析商品编号不一致", False
-            )
-        if dict(observation.get("original_options") or {}) != dict(
-            payload.get("options") or {}
-        ):
-            raise AdminGatewayError(
-                "INVALID_SPEC_RESOLUTION_REQUEST", "原始规格不一致", False
-            )
+        task_version = observation.get("task_version")
+        original_options = observation.get("original_options")
        candidates = observation.get("candidates")
        snapshot_hash = observation.get("candidate_snapshot_hash")
        selected_color = observation.get("selected_color")
+        target_size = observation.get("target_size")
+        observed_at = observation.get("observed_at")
        if (
-            not isinstance(candidates, list)
+            not isinstance(task_version, int)
+            or isinstance(task_version, bool)
+            or task_version <= 0
+            or not MockAdminGateway._valid_spec_text(
+                observation.get("attempt_id")
+            )
+            or not MockAdminGateway._valid_spec_text(
+                observation.get("pdd_goods_id")
+            )
+            or not MockAdminGateway._valid_spec_text(selected_color)
+            or not MockAdminGateway._valid_spec_text(target_size)
+            or not isinstance(original_options, Mapping)
+            or not 1 <= len(original_options) <= 16
+            or any(
+                not MockAdminGateway._valid_spec_text(key)
+                or not MockAdminGateway._valid_spec_text(value)
+                for key, value in original_options.items()
+            )
+            or not isinstance(candidates, list)
            or not 1 <= len(candidates) <= 100
            or not isinstance(snapshot_hash, str)
-            or len(snapshot_hash) != 64
-            or not str(observation.get("attempt_id") or "").strip()
+            or re.fullmatch(r"[0-9a-f]{64}", snapshot_hash) is None
+            or not MockAdminGateway._valid_observed_at(observed_at)
        ):
            raise AdminGatewayError(
                "INVALID_SPEC_RESOLUTION_REQUEST", "规格候选结构无效", False
@@ -276,8 +299,9 @@ class MockAdminGateway(AdminGateway):
        if any(
            not isinstance(candidate, Mapping)
            or candidate.get("candidate_id") != f"c{index}"
-            or not isinstance(candidate.get("raw_text"), str)
-            or not candidate["raw_text"]
+            or not MockAdminGateway._valid_spec_text(
+                candidate.get("raw_text")
+            )
            or dict(candidate.get("options") or {})
            != {
                "color": selected_color,
@@ -319,7 +343,7 @@ class MockAdminGateway(AdminGateway):
        identity = "".join(
            MockAdminGateway._frame(str(value))
            for value in (
-                task.task_id,
+                task_id,
                observation["attempt_id"],
                snapshot_hash,
                "spec-resolution-v1",
@@ -333,6 +357,88 @@ class MockAdminGateway(AdminGateway):
                "SPEC_RESOLUTION_HASH_MISMATCH", "规格解析幂等键不一致", False
            )

+    def _validate_spec_resolution_target(
+        self,
+        task_id: str,
+        observation: Mapping[str, Any],
+    ) -> AdminTask:
+        item = self._find_task(task_id)
+        if item is None:
+            raise AdminGatewayError(
+                "TASK_NOT_FOUND", "规格解析任务不存在", False
+            )
+        task = item.task
+        if task.task_type.value != "purchase":
+            raise AdminGatewayError(
+                "TASK_NOT_PURCHASE", "当前任务不是采购任务", False
+            )
+        if observation.get("task_version") != task.version:
+            raise AdminGatewayError(
+                "TASK_VERSION_CONFLICT", "规格解析任务版本不一致", False
+            )
+        payload = task.payload
+        if observation.get("pdd_goods_id") != payload.get("goods_id"):
+            raise AdminGatewayError(
+                "PDD_GOODS_MISMATCH", "规格解析商品编号不一致", False
+            )
+        original_options = dict(observation.get("original_options") or {})
+        if (
+            original_options != dict(payload.get("options") or {})
+            or not MockAdminGateway._target_option_was_claimed(
+                original_options,
+                "color",
+                str(observation.get("selected_color") or ""),
+            )
+            or not MockAdminGateway._target_option_was_claimed(
+                original_options,
+                "size",
+                str(observation.get("target_size") or ""),
+            )
+        ):
+            raise AdminGatewayError(
+                "INVALID_SPEC_RESOLUTION_REQUEST", "原始规格不一致", False
+            )
+        if task_id not in self._claimed_task_ids:
+            raise AdminGatewayError(
+                "TASK_NOT_CLAIMED_BY_CLIENT",
+                "该 Client 从未领取过此任务",
+                False,
+            )
+        return task
+
+    @staticmethod
+    def _target_option_was_claimed(
+        options: Mapping[str, Any], preferred_key: str, target: str
+    ) -> bool:
+        if preferred_key in options:
+            return options[preferred_key] == target
+        return target in options.values()
+
+    @staticmethod
+    def _valid_spec_text(value: object) -> bool:
+        return (
+            isinstance(value, str)
+            and 1 <= len(value) <= 191
+            and not any(
+                unicodedata.category(character) == "Cc"
+                for character in value
+            )
+        )
+
+    @staticmethod
+    def _valid_observed_at(value: object) -> bool:
+        if not isinstance(value, str) or re.fullmatch(
+            r"\d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2}"
+            r"(?:\.\d+)?(?:Z|[+-]\d{2}:\d{2})",
+            value,
+        ) is None:
+            return False
+        try:
+            parsed = datetime.fromisoformat(value.replace("Z", "+00:00"))
+        except ValueError:
+            return False
+        return parsed.tzinfo is not None
+
    @staticmethod
    def _frame(value: str) -> str:
        return f"{len(value.encode('utf-8'))}:{value}"
@@ -2,6 +2,7 @@

 import hashlib
 import unittest
+from copy import deepcopy

 from src.admin_gateway import (
    AdminGateway,
@@ -17,6 +18,63 @@ from src.mock_admin_gateway import MockAdminGateway
 from src.task_models import TaskType


+def _frame(value: str) -> str:
+    return f"{len(value.encode('utf-8'))}:{value}"
+
+
+def _finalize_spec_observation(task_id: str, observation: dict) -> str:
+    candidates = observation["candidates"]
+    snapshot_values = [
+        "spec-resolution-v1",
+        observation["pdd_goods_id"],
+        observation["selected_color"],
+        str(len(candidates)),
+    ]
+    for candidate in candidates:
+        snapshot_values.extend(
+            (
+                candidate["candidate_id"],
+                candidate["raw_text"],
+                candidate["options"]["color"],
+                candidate["options"]["size"],
+            )
+        )
+    observation["candidate_snapshot_hash"] = hashlib.sha256(
+        "".join(_frame(value) for value in snapshot_values).encode("utf-8")
+    ).hexdigest()
+    identity_values = (
+        task_id,
+        observation["attempt_id"],
+        observation["candidate_snapshot_hash"],
+        "spec-resolution-v1",
+    )
+    return "spec-resolution-v1:" + hashlib.sha256(
+        "".join(_frame(value) for value in identity_values).encode("utf-8")
+    ).hexdigest()
+
+
+def _spec_observation(task_id: str = "PUR-SPEC") -> tuple[dict, str]:
+    observation = {
+        "schema_version": 1,
+        "task_version": 3,
+        "attempt_id": "attempt-001",
+        "pdd_goods_id": "737116531267",
+        "original_options": {"color": "黑色", "size": "60公斤"},
+        "selected_color": "黑色",
+        "target_size": "60公斤",
+        "candidates": [
+            {
+                "candidate_id": "c1",
+                "raw_text": "120斤",
+                "options": {"color": "黑色", "size": "120斤"},
+            }
+        ],
+        "candidate_snapshot_hash": "",
+        "observed_at": "2026-08-17T08:00:00Z",
+    }
+    return observation, _finalize_spec_observation(task_id, observation)
+
+
 class MockAdminGatewayContractTest(unittest.TestCase):
    def setUp(self):
        self.gateway = MockAdminGateway()
@@ -252,6 +310,115 @@ class MockAdminGatewayContractTest(unittest.TestCase):
        self.assertEqual(second, configured)
        self.assertEqual(self.gateway.spec_resolution_count, 1)

+    def test_spec_resolution_validation_matches_admin_contract(self):
+        task_id = "PUR-SPEC-VALIDATION"
+        task = AdminTask(
+            task_id=task_id,
+            task_type=TaskType.PURCHASE,
+            version=3,
+            priority=1,
+            payload={
+                "goods_id": "737116531267",
+                "options": {"color": "黑色", "size": "60公斤"},
+            },
+        )
+        self.gateway.enqueue_task(task, self.client.client_id)
+        self.gateway.claim_next(self.client, self.all_capabilities)
+        original, _key = _spec_observation(task_id)
+
+        cases = []
+
+        wrong_color = deepcopy(original)
+        wrong_color["selected_color"] = "白色"
+        wrong_color["candidates"][0]["options"]["color"] = "白色"
+        cases.append((
+            "已选颜色不是领取规格",
+            wrong_color,
+            _finalize_spec_observation(task_id, wrong_color),
+            "INVALID_SPEC_RESOLUTION_REQUEST",
+        ))
+
+        wrong_size = deepcopy(original)
+        wrong_size["target_size"] = "70公斤"
+        cases.append((
+            "目标尺码不是领取规格",
+            wrong_size,
+            _finalize_spec_observation(task_id, wrong_size),
+            "INVALID_SPEC_RESOLUTION_REQUEST",
+        ))
+
+        missing_timezone = deepcopy(original)
+        missing_timezone["observed_at"] = "2026-08-17T08:00:00"
+        cases.append((
+            "观测时间没有时区",
+            missing_timezone,
+            _finalize_spec_observation(task_id, missing_timezone),
+            "INVALID_SPEC_RESOLUTION_REQUEST",
+        ))
+
+        oversized_attempt = deepcopy(original)
+        oversized_attempt["attempt_id"] = "a" * 192
+        cases.append((
+            "执行尝试编号超长",
+            oversized_attempt,
+            _finalize_spec_observation(task_id, oversized_attempt),
+            "INVALID_SPEC_RESOLUTION_REQUEST",
+        ))
+
+        control_character = deepcopy(original)
+        control_character["target_size"] = "60\n公斤"
+        cases.append((
+            "字段包含控制字符",
+            control_character,
+            _finalize_spec_observation(task_id, control_character),
+            "INVALID_SPEC_RESOLUTION_REQUEST",
+        ))
+
+        oversized_body = deepcopy(original)
+        oversized_body["future_padding"] = "x" * (64 * 1024)
+        cases.append((
+            "请求体超过上限",
+            oversized_body,
+            _finalize_spec_observation(task_id, oversized_body),
+            "INVALID_BODY",
+        ))
+
+        for name, request, key, expected_code in cases:
+            with self.subTest(name=name):
+                with self.assertRaises(AdminGatewayError) as raised:
+                    self.gateway.resolve_purchase_spec(task_id, key, request)
+                self.assertEqual(raised.exception.code, expected_code)
+                self.assertFalse(raised.exception.retryable)
+
+    def test_spec_resolution_uses_admin_task_and_claim_error_codes(self):
+        observation, key = _spec_observation("PUR-MISSING")
+        with self.assertRaises(AdminGatewayError) as missing:
+            self.gateway.resolve_purchase_spec(
+                "PUR-MISSING", key, observation
+            )
+        self.assertEqual(missing.exception.code, "TASK_NOT_FOUND")
+
+        task_id = "PUR-UNCLAIMED"
+        self.gateway.enqueue_task(
+            AdminTask(
+                task_id=task_id,
+                task_type=TaskType.PURCHASE,
+                version=3,
+                priority=1,
+                payload={
+                    "goods_id": "737116531267",
+                    "options": {"color": "黑色", "size": "60公斤"},
+                },
+            ),
+            self.client.client_id,
+        )
+        observation, key = _spec_observation(task_id)
+        with self.assertRaises(AdminGatewayError) as unclaimed:
+            self.gateway.resolve_purchase_spec(task_id, key, observation)
+        self.assertEqual(
+            unclaimed.exception.code, "TASK_NOT_CLAIMED_BY_CLIENT"
+        )
+
    def test_same_idempotency_key_and_content_reuses_receipt(self):
        self.gateway.enqueue_task(
            self._task("TASK-001", TaskType.COLLECT), "client-001"