Review search recommendation project design

1. BRD

  • Full name: Business Requirement Document
  • Definition: The business appealing party briefly expresses the business appeal specification that it wants to achieve

2. BRD sample recommended by review search

  • Business purpose: imitating the H5 page of Dianping app to provide offline services to users
  • Business scenario: Allow users to search for desired store services through the corresponding H5 page, and make store recommendations to users by guessing what you like
  • User: 1. Operation staff of the business. 2. C-end users who need services

Three, PRD

  • Full name: Product Requirement Document
  • Definition: The product manager formulates a clear product process, requirements, user behavior, and clear front-end and back-end functions that can be provided to R&D through the business requirements specification of the business party.

Fourth, review search recommendation PRD

Demand background

  • Imitating the needs of Dianping apps to provide users with offline search and recommendation services for stores

Noun definition

  • User: People who search for recommended pages using the imitation review app h5
  • Back-end operation: provide back-end web pages for back-end stores, merchants, and service category management
  • Operation: Use the operation background to configure the store, merchant, and service category personnel to operate the background function

Operation background function

  1. Operation login function
    Enter the user name and password to log in to the operation background
  2. Operate business management functions, including:
    business creation: business name, business score,
    business list query: business id, business name, business score,
    merchant enable and disable
  3. Service category management function, including:
    category creation: name, icon, sorting
    category list query: category id, name, icon, sorting
  4. Store management functions, including:
    store creation: name, per capita price, rating, address, label, business hours
    Store list query: store id, name, per capita price, score, address, label, business hours

H5 C terminal function

  1. User registration function:
    mobile phone number, password registration
  2. User login function:
    mobile phone number, password login
  3. Service category The
    service category displays icons and names in order of functions preset in the operation background
  4. Guess you like it:
    According to thousands of people, users’ preferences, the functions of the user service store are recommended, including store pictures, names, prices, ratings, addresses, labels, business hours, etc.
  5. Search:
    Search for service stores near the user based on bs and keywords, including store pictures, names, prices, ratings, addresses, tags, business hours, etc., and can be rooted

Five, project architecture design

Business model:

  • Make business models based on product requirements prd documents
    Business model
  • User: There are registration, login, and browsing interface behaviors
  • Merchants: there are settled, updated, evaluated, prohibited behaviors
  • Service type: create and update sorting behavior
  • Store: Create, locate, be searched and recommended behavior

Six, system module architecture

Diagram

System module architecture

mainly divided

  1. Back-end business module: used to build business function model, unified authentication access
  2. Back-end storage system: used for database, canal pipeline data synchronization, elasticsearch middleware construction
  3. Front page
  4. Search system: used for search access, semantic construction, index construction services
  5. Recommendation system: used for recommendation algorithm, recommended service access VI. Technology selection

Technology selection

  1. Back-end business: java (jdk1.8), SpringBoot framework (2.1.5)
  2. Back-end storage: mysql database (5.6), mybatis access
  3. Search system: elasticsearch distributed search engine (7.3.0), canal (1.1.3)
  4. Recommended system: spark mllib machine learning component (2.4.4)
  5. Front-end pages: html, CSS, JS, jquery

Guess you like

Origin blog.csdn.net/qq_36221788/article/details/109648213