From 5ae300389c0865758c2d5ff2bd61aa2f40bdbb07 Mon Sep 17 00:00:00 2001
From: Emelyanov535 <nouctanch777093@gmail.com>
Date: Sun, 13 Oct 2024 18:34:11 +0400
Subject: [PATCH] feature: wb parser

---
 .../.run/ParsingService [local].run.xml       | 14 +++
 parsing-service/build.gradle                  |  1 +
 .../parsingservice/enumeration/Category.java  |  5 ++
 .../enumeration/Marketplace.java              |  8 ++
 .../entity/PriceHistoryEntity.java            | 24 +++++-
 .../persistance/entity/ProductEntity.java     | 16 ++--
 .../repository/ProductPriceRepository.java    |  7 ++
 .../repository/ProductRepository.java         |  7 ++
 .../wildberries_parser/DebugRunner.java       | 19 +++++
 .../configuration/WbProperties.java           | 21 +++++
 .../configuration/WebClientConfig.java        | 18 ++++
 .../ProductInfoDto2ProductEntity.java         | 25 ++++++
 .../service/ParsingService.java               | 85 +++++++++++++++++++
 .../service/client/Client.java                |  7 ++
 .../service/client/ClientImpl.java            | 33 +++++++
 .../service/dto/ProductInfoDto.java           | 20 +++++
 .../src/main/resources/application.yml        | 14 ++-
 17 files changed, 316 insertions(+), 8 deletions(-)
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/enumeration/Category.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/enumeration/Marketplace.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/repository/ProductPriceRepository.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/repository/ProductRepository.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/DebugRunner.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/configuration/WbProperties.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/configuration/WebClientConfig.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/converter/ProductInfoDto2ProductEntity.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/ParsingService.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/client/Client.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/client/ClientImpl.java
 create mode 100644 parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/dto/ProductInfoDto.java

diff --git a/parsing-service/.run/ParsingService [local].run.xml b/parsing-service/.run/ParsingService [local].run.xml
index 1a5c151..6d4970e 100644
--- a/parsing-service/.run/ParsingService [local].run.xml	
+++ b/parsing-service/.run/ParsingService [local].run.xml	
@@ -13,4 +13,18 @@
       <option name="Make" enabled="true" />
     </method>
   </configuration>
+  <configuration default="false" name="ParsingService [local]" type="SpringBootApplicationConfigurationType" factoryName="Spring Boot">
+    <option name="ACTIVE_PROFILES" value="dev" />
+    <envs>
+      <env name="JDBC_PASSWORD" value="postgres" />
+      <env name="JDBC_USERNAME" value="postgres" />
+      <env name="JDBC_URL" value="localhost:5432/parsed_data" />
+      <env name="SERVER_PORT" value="8080" />
+    </envs>
+    <module name="parsing-service.main" />
+    <option name="SPRING_BOOT_MAIN_CLASS" value="ru.pricepulse.parsingservice.ParsingServiceApplication" />
+    <method v="2">
+      <option name="Make" enabled="true" />
+    </method>
+  </configuration>
 </component>
\ No newline at end of file
diff --git a/parsing-service/build.gradle b/parsing-service/build.gradle
index 6da600c..751ab0d 100644
--- a/parsing-service/build.gradle
+++ b/parsing-service/build.gradle
@@ -34,6 +34,7 @@ dependencies {
 	testImplementation 'org.springframework.boot:spring-boot-starter-test'
 	testImplementation 'org.springframework.kafka:spring-kafka-test'
 	testRuntimeOnly 'org.junit.platform:junit-platform-launcher'
+	implementation 'org.springframework.boot:spring-boot-starter-webflux'
 }
 
 tasks.named('test') {
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/enumeration/Category.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/enumeration/Category.java
new file mode 100644
index 0000000..a55bd5d
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/enumeration/Category.java
@@ -0,0 +1,5 @@
+package ru.pricepulse.parsingservice.enumeration;
+
+public enum Category {
+    LAPTOP
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/enumeration/Marketplace.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/enumeration/Marketplace.java
new file mode 100644
index 0000000..6c1e085
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/enumeration/Marketplace.java
@@ -0,0 +1,8 @@
+package ru.pricepulse.parsingservice.enumeration;
+
+public enum Marketplace {
+    WILDBERRIES,
+    OZON,
+    DNS
+}
+
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/entity/PriceHistoryEntity.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/entity/PriceHistoryEntity.java
index 530e66e..2cb925c 100644
--- a/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/entity/PriceHistoryEntity.java
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/entity/PriceHistoryEntity.java
@@ -1,18 +1,22 @@
 package ru.pricepulse.parsingservice.persistance.entity;
 
 import jakarta.persistence.*;
-import lombok.Getter;
-import lombok.Setter;
+import lombok.*;
 import org.hibernate.annotations.OnDelete;
 import org.hibernate.annotations.OnDeleteAction;
+import org.hibernate.proxy.HibernateProxy;
 
 import java.math.BigDecimal;
 import java.time.OffsetDateTime;
+import java.util.Objects;
 
 @Getter
 @Setter
 @Entity
 @Table(name = "price_history")
+@AllArgsConstructor
+@NoArgsConstructor
+@Builder
 public class PriceHistoryEntity {
     @Id
     @GeneratedValue(strategy = GenerationType.IDENTITY)
@@ -29,4 +33,20 @@ public class PriceHistoryEntity {
 
     @Column(name = "date", nullable = false)
     private OffsetDateTime date;
+
+    @Override
+    public final boolean equals(Object o) {
+        if (this == o) return true;
+        if (o == null) return false;
+        Class<?> oEffectiveClass = o instanceof HibernateProxy ? ((HibernateProxy) o).getHibernateLazyInitializer().getPersistentClass() : o.getClass();
+        Class<?> thisEffectiveClass = this instanceof HibernateProxy ? ((HibernateProxy) this).getHibernateLazyInitializer().getPersistentClass() : this.getClass();
+        if (thisEffectiveClass != oEffectiveClass) return false;
+        PriceHistoryEntity that = (PriceHistoryEntity) o;
+        return getId() != null && Objects.equals(getId(), that.getId());
+    }
+
+    @Override
+    public final int hashCode() {
+        return this instanceof HibernateProxy ? ((HibernateProxy) this).getHibernateLazyInitializer().getPersistentClass().hashCode() : getClass().hashCode();
+    }
 }
\ No newline at end of file
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/entity/ProductEntity.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/entity/ProductEntity.java
index 71d25e1..651760e 100644
--- a/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/entity/ProductEntity.java
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/entity/ProductEntity.java
@@ -1,19 +1,23 @@
 package ru.pricepulse.parsingservice.persistance.entity;
 
 import jakarta.persistence.*;
-import lombok.Getter;
-import lombok.Setter;
+import lombok.*;
 import org.hibernate.proxy.HibernateProxy;
+import ru.pricepulse.parsingservice.enumeration.Category;
+import ru.pricepulse.parsingservice.enumeration.Marketplace;
 
 import java.time.OffsetDateTime;
 import java.util.LinkedHashSet;
 import java.util.Objects;
 import java.util.Set;
 
+@Entity
 @Getter
 @Setter
-@Entity
 @Table(name = "product")
+@AllArgsConstructor
+@NoArgsConstructor
+@Builder
 public class ProductEntity {
     @Id
     @GeneratedValue(strategy = GenerationType.IDENTITY)
@@ -21,10 +25,12 @@ public class ProductEntity {
     private Long id;
 
     @Column(name = "marketplace", nullable = false, length = Integer.MAX_VALUE)
-    private String marketplace;
+    @Enumerated(EnumType.STRING)
+    private Marketplace marketplace;
 
     @Column(name = "category", nullable = false, length = Integer.MAX_VALUE)
-    private String category;
+    @Enumerated(EnumType.STRING)
+    private Category category;
 
     @Column(name = "brand", nullable = false, length = Integer.MAX_VALUE)
     private String brand;
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/repository/ProductPriceRepository.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/repository/ProductPriceRepository.java
new file mode 100644
index 0000000..4b965f5
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/repository/ProductPriceRepository.java
@@ -0,0 +1,7 @@
+package ru.pricepulse.parsingservice.persistance.repository;
+
+import org.springframework.data.jpa.repository.JpaRepository;
+import ru.pricepulse.parsingservice.persistance.entity.PriceHistoryEntity;
+
+public interface ProductPriceRepository extends JpaRepository<PriceHistoryEntity, Long> {
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/repository/ProductRepository.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/repository/ProductRepository.java
new file mode 100644
index 0000000..9fb095b
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/persistance/repository/ProductRepository.java
@@ -0,0 +1,7 @@
+package ru.pricepulse.parsingservice.persistance.repository;
+
+import org.springframework.data.jpa.repository.JpaRepository;
+import ru.pricepulse.parsingservice.persistance.entity.ProductEntity;
+
+public interface ProductRepository extends JpaRepository<ProductEntity, Long> {
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/DebugRunner.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/DebugRunner.java
new file mode 100644
index 0000000..976288f
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/DebugRunner.java
@@ -0,0 +1,19 @@
+package ru.pricepulse.parsingservice.wildberries_parser;
+
+import lombok.AllArgsConstructor;
+import org.springframework.boot.CommandLineRunner;
+import org.springframework.stereotype.Component;
+import ru.pricepulse.parsingservice.wildberries_parser.service.ParsingService;
+
+@Component
+@AllArgsConstructor
+public class DebugRunner implements CommandLineRunner {
+    private final ParsingService parsingService;
+
+    @Override
+    public void run(String... args){
+        System.out.println("Начинаем отладку...");
+        parsingService.parse();
+        System.out.println("Заканчиваем отладку...");
+    }
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/configuration/WbProperties.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/configuration/WbProperties.java
new file mode 100644
index 0000000..dc6b95a
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/configuration/WbProperties.java
@@ -0,0 +1,21 @@
+package ru.pricepulse.parsingservice.wildberries_parser.configuration;
+
+import lombok.Getter;
+import lombok.Setter;
+import org.springframework.boot.context.properties.ConfigurationProperties;
+import org.springframework.context.annotation.Configuration;
+
+@Configuration
+@ConfigurationProperties(prefix = "rest.wildberries")
+@Getter
+@Setter
+public class WbProperties {
+    private String baseUrl;
+    private String catalogUrl;
+    private String userAgent;
+    private String catalogWbUrl;
+    private int retryAttempts;
+    private long retryDelay;
+    private String laptopUrl;
+    private String shard;
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/configuration/WebClientConfig.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/configuration/WebClientConfig.java
new file mode 100644
index 0000000..944f9e1
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/configuration/WebClientConfig.java
@@ -0,0 +1,18 @@
+package ru.pricepulse.parsingservice.wildberries_parser.configuration;
+
+import org.springframework.context.annotation.Bean;
+import org.springframework.context.annotation.Configuration;
+import org.springframework.web.reactive.function.client.WebClient;
+
+@Configuration
+public class WebClientConfig {
+    @Bean
+    public WebClient webClient() {
+        return WebClient.builder()
+                .defaultHeader("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0)")
+                .codecs(configurer -> configurer
+                        .defaultCodecs()
+                        .maxInMemorySize(10 * 1024 * 1024))
+                .build();
+    }
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/converter/ProductInfoDto2ProductEntity.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/converter/ProductInfoDto2ProductEntity.java
new file mode 100644
index 0000000..6e4168e
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/converter/ProductInfoDto2ProductEntity.java
@@ -0,0 +1,25 @@
+package ru.pricepulse.parsingservice.wildberries_parser.converter;
+
+import org.springframework.core.convert.converter.Converter;
+import org.springframework.stereotype.Component;
+import ru.pricepulse.parsingservice.enumeration.Category;
+import ru.pricepulse.parsingservice.enumeration.Marketplace;
+import ru.pricepulse.parsingservice.persistance.entity.ProductEntity;
+import ru.pricepulse.parsingservice.wildberries_parser.service.dto.ProductInfoDto;
+
+import java.time.OffsetDateTime;
+
+
+@Component
+public class ProductInfoDto2ProductEntity implements Converter<ProductInfoDto, ProductEntity> {
+    @Override
+    public ProductEntity convert(ProductInfoDto source) {
+        return ProductEntity.builder()
+                .marketplace(Marketplace.WILDBERRIES)
+                .category(Category.LAPTOP)
+                .brand(source.getBrand())
+                .productName(source.getName())
+                .createdAt(OffsetDateTime.now())
+                .build();
+    }
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/ParsingService.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/ParsingService.java
new file mode 100644
index 0000000..8f33ae4
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/ParsingService.java
@@ -0,0 +1,85 @@
+package ru.pricepulse.parsingservice.wildberries_parser.service;
+
+import com.fasterxml.jackson.core.type.TypeReference;
+import com.fasterxml.jackson.databind.ObjectMapper;
+import jakarta.transaction.Transactional;
+import lombok.AllArgsConstructor;
+import org.springframework.core.convert.ConversionService;
+import org.springframework.stereotype.Service;
+import ru.pricepulse.parsingservice.persistance.entity.PriceHistoryEntity;
+import ru.pricepulse.parsingservice.persistance.entity.ProductEntity;
+import ru.pricepulse.parsingservice.persistance.repository.ProductPriceRepository;
+import ru.pricepulse.parsingservice.persistance.repository.ProductRepository;
+import ru.pricepulse.parsingservice.wildberries_parser.configuration.WbProperties;
+import ru.pricepulse.parsingservice.wildberries_parser.service.client.Client;
+import ru.pricepulse.parsingservice.wildberries_parser.service.dto.ProductInfoDto;
+
+import java.math.BigDecimal;
+import java.time.OffsetDateTime;
+import java.util.ArrayList;
+import java.util.List;
+import java.util.Map;
+
+@Service
+@AllArgsConstructor
+public class ParsingService {
+    private final Client client;
+    private final ObjectMapper objectMapper;
+    private final ConversionService conversionService;
+    private final ProductRepository productRepository;
+    private final WbProperties wbProperties;
+    private final ProductPriceRepository productPriceRepository;
+
+    @Transactional
+    public void parse() {
+        List<ProductEntity> productEntities = new ArrayList<>();
+        List<PriceHistoryEntity> priceHistories = new ArrayList<>();
+
+        final int elementsInPage = 100;
+        int page = 1;
+        Integer totalPages = null;
+
+        do {
+            var pageData = client.scrapPage(page, wbProperties.getShard(), wbProperties.getLaptopUrl());
+
+            if (totalPages == null) {
+                Map<String, Object> dataMap = (Map<String, Object>) pageData.get("data");
+                int totalElements = (int) dataMap.get("total");
+                totalPages = (int) Math.ceil((double) totalElements / elementsInPage);
+            }
+
+            List<ProductInfoDto> productInfoDtoList = convertMapObjectToListProductInfoDto(pageData);
+
+            productInfoDtoList.forEach(dto -> {
+
+                ProductEntity productEntity = conversionService.convert(dto, ProductEntity.class);
+
+                PriceHistoryEntity priceHistory = PriceHistoryEntity.builder()
+                        .product(productEntity)
+                        .price(BigDecimal.valueOf(dto.getSalePriceU() / 100.0))
+                        .date(OffsetDateTime.now())
+                        .build();
+
+                productEntities.add(productEntity);
+                priceHistories.add(priceHistory);
+            });
+
+            page++;
+        } while (page <= totalPages);
+
+        productRepository.saveAll(productEntities);
+        productPriceRepository.saveAll(priceHistories);
+    }
+
+    private List<ProductInfoDto> convertMapObjectToListProductInfoDto(Map<String, Object> map) {
+        Map<String, ArrayList<Object>> dataMap = (Map<String, ArrayList<Object>>) map.get("data");
+        return getProductInfoDtos(dataMap);
+    }
+
+    private List<ProductInfoDto> getProductInfoDtos(Map<String, ArrayList<Object>> dataMap) {
+        return objectMapper.convertValue(
+                dataMap.get("products"),
+                new TypeReference<>() {}
+        );
+    }
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/client/Client.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/client/Client.java
new file mode 100644
index 0000000..f98cbc8
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/client/Client.java
@@ -0,0 +1,7 @@
+package ru.pricepulse.parsingservice.wildberries_parser.service.client;
+
+import java.util.Map;
+
+public interface Client {
+    Map<String, Object> scrapPage(int page, String shard, String query);
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/client/ClientImpl.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/client/ClientImpl.java
new file mode 100644
index 0000000..d76fccf
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/client/ClientImpl.java
@@ -0,0 +1,33 @@
+package ru.pricepulse.parsingservice.wildberries_parser.service.client;
+
+import lombok.AllArgsConstructor;
+import org.springframework.core.ParameterizedTypeReference;
+import org.springframework.stereotype.Service;
+import org.springframework.web.reactive.function.client.WebClient;
+import ru.pricepulse.parsingservice.wildberries_parser.configuration.WbProperties;
+
+import java.util.Map;
+
+@AllArgsConstructor
+@Service
+public class ClientImpl implements Client {
+
+    private final WebClient webClient;
+    private final WbProperties wbProperties;
+
+    @Override
+    public Map<String, Object> scrapPage(int page, String shard, String query) {
+        String url = wbProperties.getCatalogWbUrl() +
+                shard +
+                query +
+                "?dest=-1257786&page=" + page + "&subject=2290";
+
+        return webClient.get()
+                .uri(url)
+                .retrieve()
+                .bodyToMono(new ParameterizedTypeReference<Map<String, Object>>() {
+                })
+                .retry(50)
+                .block();
+    }
+}
diff --git a/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/dto/ProductInfoDto.java b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/dto/ProductInfoDto.java
new file mode 100644
index 0000000..d66a36b
--- /dev/null
+++ b/parsing-service/src/main/java/ru/pricepulse/parsingservice/wildberries_parser/service/dto/ProductInfoDto.java
@@ -0,0 +1,20 @@
+package ru.pricepulse.parsingservice.wildberries_parser.service.dto;
+
+import lombok.AllArgsConstructor;
+import lombok.Builder;
+import lombok.Data;
+import lombok.NoArgsConstructor;
+
+@Data
+@AllArgsConstructor
+@NoArgsConstructor
+@Builder
+public class ProductInfoDto {
+    private Long id;
+    private String brand;
+    private String name;
+    private String supplier;
+    private Double supplierRating;
+    private Integer salePriceU;
+    private Integer reviewRating;
+}
diff --git a/parsing-service/src/main/resources/application.yml b/parsing-service/src/main/resources/application.yml
index 65278f5..060d516 100644
--- a/parsing-service/src/main/resources/application.yml
+++ b/parsing-service/src/main/resources/application.yml
@@ -13,4 +13,16 @@ spring:
     username: ${JDBC_USERNAME}
     password: ${JDBC_PASSWORD}
   liquibase:
-    change-log: classpath:/db/changelog/master.yml
\ No newline at end of file
+    change-log: classpath:/db/changelog/master.yml
+
+rest:
+  wildberries:
+    base-url: "https://static-basket-01.wbbasket.ru"
+    catalog-url: "/vol0/data/main-menu-ru-ru-v3.json"
+    user-agent: "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0)"
+    catalog-wb-url: "https://catalog.wb.ru/catalog/"
+    retry-attempts: 5
+    retry-delay: 1000
+    shard: "electronic15"
+    laptop-url: "/catalog"
+