Added reset teset in supply chain scenarios

maro/simulator/scenarios/supply_chain/datamodels/distribution.py

@@ -19,5 +19,8 @@ class DistributionDataModel(DataModelBase):
     def __init__(self) -> None:
         super(DistributionDataModel, self).__init__()
 
+    def initialize(self) -> None:
+        self.reset()
+
     def reset(self) -> None:
         super(DistributionDataModel, self).reset()

maro/simulator/scenarios/supply_chain/units/distribution.py

@@ -101,6 +101,8 @@ def initialize(self) -> None:
 
             # TODO: add vehicle patient setting if needed
 
+        self.data_model.initialize()
+
         for sku_id in self.facility.products.keys():
             self._unit_delay_order_penalty[sku_id] = self.facility.skus[sku_id].unit_delay_order_penalty
 

tests/data/supply_chain/case_04/test_case_04.csv

@@ -4,11 +4,23 @@ food_1,2021/1/2,43.1,33.39,20
 food_1,2021/1/3,43.2,33.39,30
 food_1,2021/1/4,43.3,33.39,40
 food_1,2021/1/5,43.4,33.39,50
+food_1,2021/1/6,43.4,33.39,60
+food_1,2021/1/7,43.4,33.39,70
+food_1,2021/1/8,43.4,33.39,80
+food_1,2021/1/9,43.4,33.39,90
+food_1,2021/1/10,43.4,33.39,100
+food_1,2021/1/11,43.4,33.39,110
 hobby_1,2021/1/1,28.32,21.79,100
 hobby_1,2021/1/2,28.32,21.79,200
 hobby_1,2021/1/3,28.32,21.79,300
 hobby_1,2021/1/4,28.32,21.79,400
 hobby_1,2021/1/5,28.32,21.79,500
+hobby_1,2021/1/6,28.32,21.79,600
+hobby_1,2021/1/7,28.32,21.79,700
+hobby_1,2021/1/8,28.32,21.79,800
+hobby_1,2021/1/9,28.32,21.79,900
+hobby_1,2021/1/10,28.32,21.79,1000
+hobby_1,2021/1/11,28.32,21.79,1100
 household_1,2022/5/14,17.35974446,14.4,146
 household_1,2022/5/15,17.29592816,14.4,164
 household_1,2022/5/16,17.43129521,14.4,127

tests/data/supply_chain/case_05/config.yml

@@ -146,6 +146,30 @@ world:
         unit_delay_order_penalty: 20
         unit_order_cost: 0
 
+    - name: "Supplier_SKU1"
+      definition_ref: "SupplierFacility"
+      skus:
+        sku1:
+          init_stock: 20
+          has_manufacture: True
+          max_manufacture_rate: 50
+          manufacture_leading_time: 1
+          unit_product_cost: 1
+          price: 10
+          unit_delay_order_penalty: 10
+          has_consumer: True
+        sku3:
+          init_stock: 80
+          unit_product_cost: 1
+          price: 10
+          unit_delay_order_penalty: 10
+      children:
+        storage: *small_storage
+        distribution: *normal_distribution
+      config:
+        unit_delay_order_penalty: 20
+        unit_order_cost: 0
+
     - name: "Warehouse_001"
       definition_ref: "WarehouseFacility"
       skus:
@@ -154,6 +178,7 @@ world:
           sub_storage_id: 1
           storage_upper_bound: 40
           price: 100
+          has_consumer: True
         sku2:
           init_stock: 12
           sub_storage_id: 1
@@ -257,6 +282,11 @@ world:
             cost: 0.6
 
     Warehouse_001:
+      sku1:
+        "Supplier_SKU1":
+          "train":
+            vlt: 3
+            cost: 1
       sku3:
         "Supplier_SKU3":
           "train":

tests/supply_chain/common.py

@@ -2,25 +2,93 @@
 # Licensed under the MIT license.
 
 import os
+from collections import defaultdict
+from typing import Dict, List, Tuple
 
 import numpy as np
 
 from maro.simulator import Env
 
 
-def build_env(case_name: str, durations: int):
+def build_env(case_name: str, durations: int) -> Env:
     case_folder = os.path.join("tests", "data", "supply_chain", case_name)
+    return Env(scenario="supply_chain", topology=case_folder, durations=durations)
 
-    env = Env(scenario="supply_chain", topology=case_folder, durations=durations)
 
-    return env
-
-
-def get_product_dict_from_storage(env: Env, frame_index: int, node_index: int):
+def get_product_dict_from_storage(env: Env, frame_index: int, node_index: int) -> Dict[int, int]:
     sku_id_list = env.snapshot_list["storage"][frame_index:node_index:"sku_id_list"].flatten().astype(np.int)
     product_quantity = env.snapshot_list["storage"][frame_index:node_index:"product_quantity"].flatten().astype(np.int)
 
-    return {sku_id: quantity for sku_id, quantity in zip(sku_id_list, product_quantity)}
+    return dict(zip(sku_id_list, product_quantity))
+
+
+def snapshot_query(env: Env, i: int) -> Tuple[
+    Dict[int, list], Dict[int, list], Dict[int, list], Dict[int, list], Dict[int, list], Dict[int, list]
+]:
+    consumer_nodes = env.snapshot_list["consumer"]
+    storage_nodes = env.snapshot_list["storage"]
+    seller_nodes = env.snapshot_list["seller"]
+    manufacture_nodes = env.snapshot_list["manufacture"]
+    distribution_nodes = env.snapshot_list["distribution"]
+
+    states_consumer: Dict[int, list] = {}
+    states_storage: Dict[int, list] = {}
+    states_seller: Dict[int, list] = {}
+    states_manufacture: Dict[int, list] = {}
+    states_distribution: Dict[int, list] = {}
+
+    env_metric = env.metrics
+
+    for idx in range(len(consumer_nodes)):
+        states_consumer[idx] = consumer_nodes[i:idx:consumer_features].flatten().astype(np.float)
+
+    for idx in range(len(storage_nodes)):
+        states_storage[idx] = list(storage_nodes[i:idx:storage_features].flatten().astype(np.float))
+        states_storage[idx].append(storage_nodes[i:idx:"product_id_list"].flatten().astype(np.int))
+        states_storage[idx].append(storage_nodes[i:idx:"product_quantity"].flatten().astype(np.float))
+        states_storage[idx].append(storage_nodes[i:idx:"remaining_space"].flatten().astype(np.float))
+
+    for idx in range(len(manufacture_nodes)):
+        states_manufacture[idx] = (
+            manufacture_nodes[i:idx:manufacture_features]
+            .flatten()
+            .astype(
+                np.float,
+            )
+        )
+
+    for idx in range(len(distribution_nodes)):
+        states_distribution[idx] = (
+            distribution_nodes[i:idx:distribution_features]
+            .flatten()
+            .astype(
+                np.float,
+            )
+        )
+
+    for idx in range(len(seller_nodes)):
+        states_seller[idx] = seller_nodes[i:idx:seller_features].flatten().astype(np.float)
+
+    return env_metric, states_consumer, states_storage, states_seller, states_manufacture, states_distribution
+
+
+def test_env_reset_snapshot_query(
+    env: Env,
+    action_1: object,
+    action_2: object,
+    expect_tick: int,
+    random_tick: list = None,
+) -> List[tuple]:
+    snapshots: List[tuple] = []  # List of (env_metric, states_consumer, ..., states_distribution)
+    for i in range(expect_tick):
+        snapshots.append(snapshot_query(env, i))
+        env.step(action_1)
+
+        if random_tick is not None:
+            if i in random_tick:
+                env.step(action_2)
+
+    return list(zip(*snapshots))
 
 
 SKU1_ID = 1
@@ -29,3 +97,40 @@ def get_product_dict_from_storage(env: Env, frame_index: int, node_index: int):
 SKU4_ID = 4
 FOOD_1_ID = 20
 HOBBY_1_ID = 30
+
+consumer_features = (
+    "id",
+    "facility_id",
+    "sku_id",
+    "order_base_cost",
+    "purchased",
+    "received",
+    "order_product_cost",
+    "latest_consumptions",
+    "in_transit_quantity",
+)
+
+storage_features = ("id", "facility_id")
+
+seller_features = (
+    "sold",
+    "demand",
+    "total_sold",
+    "id",
+    "total_demand",
+    "backlog_ratio",
+    "facility_id",
+    "product_unit_id",
+)
+
+manufacture_features = (
+    "id",
+    "facility_id",
+    "start_manufacture_quantity",
+    "sku_id",
+    "in_pipeline_quantity",
+    "finished_quantity",
+    "product_unit_id",
+)
+
+distribution_features = ("id", "facility_id", "pending_order_number", "pending_product_quantity")