spider/internal/model/task_detail.go

package model

import "time"

// TaskDetail stores a single operation within a task execution.
// Every search query, URL crawl, snippet extraction, and merchant callback is recorded,
// making the entire task fully reproducible and debuggable.
//
// Key fields for traceability:
//   - Depth: 0 = serper result, 1 = crawled from serper link, 2 = sub-page from depth-1, etc.
//   - ParentURL: which page led to this operation (trace the discovery chain)
//   - Input: raw content received (snippet text, page HTML summary, message text)
//   - Output: what was extracted/decided (usernames, classification, cleaned data)
type TaskDetail struct {
	ID        uint      `gorm:"primaryKey;autoIncrement" json:"id"`
	TaskID    uint      `gorm:"index;not null" json:"task_id"`
	Seq       int       `gorm:"not null" json:"seq"`
	Action    string    `gorm:"size:50;not null;index" json:"action"`
	URL       string    `gorm:"size:2048" json:"url"`
	ParentURL string    `gorm:"size:2048" json:"parent_url"`
	Depth     int       `gorm:"default:0" json:"depth"`
	Input     string    `gorm:"type:mediumtext" json:"input"`
	Output    string    `gorm:"type:mediumtext" json:"output"`
	Status    string    `gorm:"size:20;not null;default:'ok'" json:"status"`
	Duration  int       `gorm:"default:0" json:"duration_ms"`
	Extra     string    `gorm:"type:text" json:"extra"`
	CreatedAt time.Time `json:"created_at"`
}