Implement TED Scraper and Configuration Management
- Introduced a new TED scraper in `cmd/scraper` to handle downloading and parsing TED XML files. - Added configuration management for the scraper, including a new `Config` struct and YAML configuration file. - Created necessary handler, service, and repository layers for tender management, adhering to Clean Architecture principles. - Implemented comprehensive logging and error handling throughout the scraper functionality. - Updated API routes to include tender management operations, enhancing the overall system capabilities.
This commit is contained in:
@@ -0,0 +1,50 @@
|
||||
package main
|
||||
|
||||
import (
|
||||
"time"
|
||||
"tm/pkg/config"
|
||||
)
|
||||
|
||||
// Config defines the scraper application configuration
|
||||
type Config struct {
|
||||
Database config.DatabaseConfig `mapstructure:"database"`
|
||||
Logging config.LoggingConfig `mapstructure:"logging"`
|
||||
TED ScraperConfig `mapstructure:"ted"`
|
||||
}
|
||||
|
||||
type ScraperConfig struct {
|
||||
BaseURL string `mapstructure:"base_url"`
|
||||
Timeout time.Duration `mapstructure:"timeout"`
|
||||
MaxRetries int `mapstructure:"max_retries"`
|
||||
RetryDelay time.Duration `mapstructure:"retry_delay"`
|
||||
UserAgent string `mapstructure:"user_agent"`
|
||||
MaxConcurrency int `mapstructure:"max_concurrency"`
|
||||
DownloadDir string `mapstructure:"download_dir"`
|
||||
CleanupAfter time.Duration `mapstructure:"cleanup_after"`
|
||||
ScrapingInterval time.Duration `mapstructure:"scraping_interval"`
|
||||
}
|
||||
|
||||
// LoadConfig loads the configuration from file
|
||||
func LoadConfig(path string) (*Config, error) {
|
||||
cfg := &Config{
|
||||
// Set default values
|
||||
TED: ScraperConfig{
|
||||
BaseURL: "https://ted.europa.eu/packages/daily",
|
||||
Timeout: 30 * time.Second,
|
||||
MaxRetries: 3,
|
||||
RetryDelay: 5 * time.Second,
|
||||
UserAgent: "TenderManagement-TED-Scraper/1.0",
|
||||
MaxConcurrency: 3,
|
||||
DownloadDir: "/tmp/ted_downloads",
|
||||
CleanupAfter: 24 * time.Hour,
|
||||
ScrapingInterval: 12 * time.Hour,
|
||||
},
|
||||
}
|
||||
|
||||
result, err := config.LoadConfig(path, cfg)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
|
||||
return result, nil
|
||||
}
|
||||
Reference in New Issue
Block a user