项目结构：

# encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 22:59 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : bead.py from dataclasses import dataclass @dataclass(frozen=False) class BeadItem: """ 多宝手串珠子实体 """ bead_id: str name: str material: str color_group: str # red/green/purple/gold unit_price: float stock: int # encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:00 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : jewelry.py from dataclasses import dataclass @dataclass(frozen=False) class JewelryItem: """ 成套首饰商品实体 """ sku_id: str name: str category: str # necklace / earring / bracelet / ring material: str # Au999 / 18K / S925 color: str style: str price: float stock: int has_gem: bool # encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:01 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : base_rule.py from abc import ABC, abstractmethod from typing import Any, List class BaseRule(ABC): """ 约束规则抽象基类 """ @abstractmethod def check(self, item: Any, path: List[Any], **kwargs) -> bool: """ 校验单个候选物料是否满足规则 :param item: 当前待选物料 :param path: 当前已选中集合 """ pass # encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:01 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : bracelet_rule.py from typing import Dict, List from .base_rule import BaseRule from Backtracking.dto import BeadItem class BraceletRule(BaseRule): """ 手串搭配约束规则 """ def __init__(self, max_single_color: int = 4): self.max_single_color = max_single_color def check(self, item: BeadItem, path: List[BeadItem], **kwargs) -> bool: """ :param item: :param path: :param kwargs: :return: """ # 1. 库存校验 used_count = path.count(item) if used_count >= item.stock: return False # 2. 色系均衡约束 color_cnt: Dict[str, int] = {} for b in path: color_cnt[b.color_group] = color_cnt.get(b.color_group, 0) + 1 if color_cnt.get(item.color_group, 0) >= self.max_single_color: return False return True # encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:04 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : scene_jewelry_rule.py from typing import Dict, List, Set from .base_rule import BaseRule from Backtracking.dto import JewelryItem class SceneJewelryRule(BaseRule): """ 场景化成套首饰约束规则 """ def __init__(self, scene_config: Dict): self.scene_config = scene_config def check(self, item: JewelryItem, path: List[JewelryItem], **kwargs) -> bool: allow_material: Set = self.scene_config["allow_material"] must_gem: bool = self.scene_config["must_gem"] # 库存 if item.stock <= 0: return False # 材质限制 if item.material not in allow_material: return False # 是否必须带宝石 if must_gem and not item.has_gem: return False return True @staticmethod def get_scene_config(scene_type: str) -> Dict: """ 场景配置中心，新增场景只在这里扩展 :param scene_type: :return: """ scene_map = { "wedding": { "allow_material": {"Au999", "18K"}, "must_gem": True }, "commute": { "allow_material": {"Au999", "S925", "18K"}, "must_gem": False }, "dinner": { "allow_material": {"18K"}, "must_gem": True } } return scene_map.get(scene_type, scene_map["commute"]) # encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:05 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : score_util.py from Backtracking.dto import BeadItem, JewelryItem def score_bracelet_scheme(scheme: list[BeadItem]) -> float: """ 手串方案评分：色系多样性优先 :param scheme: :return: """ color_set = {b.color_group for b in scheme} diversity = len(color_set) total_cost = sum(b.unit_price for b in scheme) return diversity * 10 - total_cost / 200 def score_jewelry_scheme(scheme: list[JewelryItem]) -> float: """ 成套首饰方案评分 :param scheme: :return: """ gem_cnt = sum(1 for i in scheme if i.has_gem) stock_score = sum(min(i.stock, 5) for i in scheme) return gem_cnt * 5 + stock_score # encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:07 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : backtrack_bracelet.py from typing import List from Backtracking.dto import BeadItem from Backtracking.rule.bracelet_rule import BraceletRule class BraceletBackTracker: """ """ def __init__(self, bead_pool: List[BeadItem], rule: BraceletRule): self.bead_pool = bead_pool self.rule = rule self.solutions: List[List[BeadItem]] = [] def backtrack(self, path: List[BeadItem], remain: int, total_cost: float, budget: float): """ :param path: :param remain: :param total_cost: :param budget: :return: """ if remain == 0: self.solutions.append(path.copy()) return if total_cost > budget: return for bead in self.bead_pool: if not self.rule.check(bead, path): continue path.append(bead) self.backtrack(path, remain - 1, total_cost + bead.unit_price, budget) path.pop() def run(self, target_count: int, budget: float) -> List[List[BeadItem]]: """ :param target_count: :param budget: :return: """ self.solutions.clear() self.backtrack([], target_count, 0.0, budget) return self.solutions # encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:08 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : backtrack_jewelry.py from typing import List, Set from Backtracking.dto import JewelryItem from Backtracking.rule.scene_jewelry_rule import SceneJewelryRule class JewelrySceneBackTracker: """ """ def __init__(self, goods_pool: List[JewelryItem], rule: SceneJewelryRule): self.goods_pool = goods_pool self.rule = rule self.solutions: List[List[JewelryItem]] = [] def backtrack( self, start_idx: int, selected: List[JewelryItem], total_price: float, budget: float, target_categories: Set[str] ): """ :param start_idx: :param selected: :param total_price: :param budget: :param target_categories: :return: """ selected_cats = {x.category for x in selected} if selected_cats == target_categories: self.solutions.append(selected.copy()) return if total_price > budget: return for i in range(start_idx, len(self.goods_pool)): item = self.goods_pool[i] if item.category in selected_cats: continue if not self.rule.check(item, selected): continue selected.append(item) self.backtrack(i + 1, selected, total_price + item.price, budget, target_categories) selected.pop() def run(self, budget: float, target_categories: Set[str]) -> List[List[JewelryItem]]: """ :param budget: :param target_categories: :return: """ self.solutions.clear() self.backtrack(0, [], 0.0, budget, target_categories) return self.solutions # encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:09 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : bracelet_service.py from typing import List from Backtracking.dto import BeadItem from Backtracking.core import BraceletBackTracker from Backtracking.rule.bracelet_rule import BraceletRule from Backtracking.common import score_bracelet_scheme class BraceletMatchService: """ 手串搭配业务服务层：封装算法调用、排序、截断 """ def __init__(self, bead_pool: List[BeadItem]): self.bead_pool = bead_pool def match( self, target_count: int, budget: float, max_color_limit: int = 4, top_n: int = 6 ) -> List[List[BeadItem]]: """ :param target_count: :param budget: :param max_color_limit: :param top_n: :return: """ rule = BraceletRule(max_single_color=max_color_limit) tracker = BraceletBackTracker(self.bead_pool, rule) schemes = tracker.run(target_count, budget) # 业务后处理：打分排序，只返回TopN schemes.sort(key=score_bracelet_scheme, reverse=True) return schemes[:top_n] # encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:10 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : jewelry_scene_service.py from typing import List, Set from Backtracking.dto import JewelryItem from Backtracking.core import JewelrySceneBackTracker from Backtracking.rule.scene_jewelry_rule import SceneJewelryRule from Backtracking.common import score_jewelry_scheme class JewelrySceneMatchService: """ 场景成套首饰业务服务 """ def __init__(self, goods_pool: List[JewelryItem]): self.goods_pool = goods_pool def match_by_scene( self, scene: str, budget: float, target_categories: Set[str], top_n: int = 8 ) -> List[List[JewelryItem]]: """ :param scene: :param budget: :param target_categories: :param top_n: :return: """ scene_conf = SceneJewelryRule.get_scene_config(scene) rule = SceneJewelryRule(scene_conf) tracker = JewelrySceneBackTracker(self.goods_pool, rule) schemes = tracker.run(budget, target_categories) schemes.sort(key=score_jewelry_scheme, reverse=True) return schemes[:top_n]

调用：

# encoding: utf-8 # 版权所有 2026 ©涂聚文有限公司™ ® # 许可信息查看：言語成了邀功盡責的功臣，還需要行爲每日來值班嗎 # 描述：Backtracking Algorithm # Author : geovindu,Geovin Du 涂聚文. # IDE : PyCharm 2024.3.6 python 3.11 # os : windows 10 # database : mysql 9.0 sql server 2019, postgreSQL 17.0 Oracle 21c Neo4j # Datetime : 2026/7/22 23:12 # User : geovindu # Product : PyCharm # Project : PyAlgorithms # File : BacktrackingBll.py from Backtracking.dto import BeadItem, JewelryItem from Backtracking.service import BraceletMatchService, JewelrySceneMatchService class BacktrackingBll(object): """ """ def test_bracelet_match(self): """ :return: """ bead_pool = [ BeadItem("B01", "南红圆珠", "南红", "red", 168, 4), BeadItem("B02", "和田玉圆珠", "和田玉", "green", 198, 5), BeadItem("B03", "紫水晶", "紫水晶", "purple", 128, 4), BeadItem("B04", "足金隔珠", "足金", "gold", 320, 3), ] svc = BraceletMatchService(bead_pool) result = svc.match(target_count=8, budget=2000) print("===== 多宝手串搭配方案 =====") for idx, scheme in enumerate(result, 1): total = sum(b.unit_price for b in scheme) names = [b.name for b in scheme] print(f"方案{idx} 总价:{total:.2f} 珠子:{names}") def test_jewelry_scene_match(self): """ :return: """ goods_pool = [ JewelryItem("N001", "碎钻项链", "necklace", "18K", "white", "luxury", 3299, 12, True), JewelryItem("N003", "素金项链", "necklace", "Au999", "yellow", "minimalist", 2199, 9, False), JewelryItem("E001", "白钻耳饰", "earring", "18K", "white", "luxury", 2199, 15, True), JewelryItem("E003", "素金耳饰", "earring", "Au999", "yellow", "minimalist", 1399, 11, False), ] svc = JewelrySceneMatchService(goods_pool) target_cats = {"necklace", "earring"} print("\n===== 婚嫁场景 =====") wedding = svc.match_by_scene("wedding", budget=8000, target_categories=target_cats) for item_set in wedding: print([x.name for x in item_set], "总价", sum(x.price for x in item_set)) print("\n===== 通勤场景 =====") commute = svc.match_by_scene("commute", budget=5000, target_categories=target_cats) for item_set in commute: print([x.name for x in item_set], "总价", sum(x.price for x in item_set)) def Demo(self): """ :return: """ self.test_bracelet_match() self.test_jewelry_scene_match()

输出：