276 lines
6.7 KiB
Go
276 lines
6.7 KiB
Go
package repo
|
|
|
|
import (
|
|
"bytes"
|
|
"database/sql"
|
|
"fmt"
|
|
"strconv"
|
|
"strings"
|
|
|
|
"github.com/jobs-scraper/internal/pkg/domain"
|
|
)
|
|
|
|
type JobRepository struct {
|
|
db *sql.DB
|
|
}
|
|
|
|
func NewJobRepository(db *sql.DB) *JobRepository {
|
|
return &JobRepository{db: db}
|
|
}
|
|
|
|
// JobExistsByLink checks if a job with the given link already exists in the database
|
|
func (r *JobRepository) JobExistsByLink(jobLink string) (bool, error) {
|
|
var exists bool
|
|
err := r.db.QueryRow("SELECT EXISTS(SELECT 1 FROM jobs WHERE job_link = $1)", jobLink).Scan(&exists)
|
|
if err != nil {
|
|
return false, fmt.Errorf("error checking job existence: %v", err)
|
|
}
|
|
return exists, nil
|
|
}
|
|
|
|
// SaveJob saves a single job and returns the generated ID
|
|
func (r *JobRepository) SaveJob(job domain.Job) (int64, error) {
|
|
sqlStatement := `
|
|
INSERT INTO jobs (title, company, company_link, location, job_link, job_timestamp, provider)
|
|
VALUES ($1, $2, $3, $4, $5, $6, $7)
|
|
ON CONFLICT (job_link) DO UPDATE SET
|
|
title = EXCLUDED.title,
|
|
company = EXCLUDED.company,
|
|
company_link = EXCLUDED.company_link,
|
|
location = EXCLUDED.location,
|
|
job_timestamp = EXCLUDED.job_timestamp,
|
|
provider = EXCLUDED.provider
|
|
RETURNING id
|
|
`
|
|
|
|
var id int64
|
|
err := r.db.QueryRow(sqlStatement,
|
|
job.Title, job.Company, job.CompanyLink, job.Location,
|
|
job.JobLink, job.JobPostTime, job.Provider,
|
|
).Scan(&id)
|
|
if err != nil {
|
|
return 0, fmt.Errorf("error inserting job: %v", err)
|
|
}
|
|
|
|
return id, nil
|
|
}
|
|
|
|
func (r *JobRepository) SaveJobs(jobs []domain.Job) error {
|
|
if len(jobs) == 0 {
|
|
return nil
|
|
}
|
|
|
|
// Deduplicate jobs by job_link to avoid duplicate errors
|
|
jobMap := make(map[string]domain.Job)
|
|
for _, job := range jobs {
|
|
jobMap[job.JobLink] = job
|
|
}
|
|
|
|
// Convert to slice of pointers for the bulk function
|
|
uniqueJobs := make([]*domain.Job, 0, len(jobMap))
|
|
for _, job := range jobMap {
|
|
jobCopy := job // Important: create a copy to avoid pointer issues
|
|
uniqueJobs = append(uniqueJobs, &jobCopy)
|
|
}
|
|
|
|
sqlTemplate := `
|
|
INSERT INTO jobs (title, company, company_link, location, job_link, job_timestamp, provider)
|
|
VALUES %s
|
|
ON CONFLICT (job_link) DO UPDATE SET
|
|
title = EXCLUDED.title,
|
|
company = EXCLUDED.company,
|
|
company_link = EXCLUDED.company_link,
|
|
location = EXCLUDED.location,
|
|
job_timestamp = EXCLUDED.job_timestamp,
|
|
provider = EXCLUDED.provider
|
|
`
|
|
|
|
sqlStatement, vals := prepareQueryCreateBulk(sqlTemplate, uniqueJobs)
|
|
|
|
_, err := r.db.Exec(sqlStatement, vals...)
|
|
if err != nil {
|
|
return fmt.Errorf("error inserting jobs: %v", err)
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
type JobFilter struct {
|
|
Location string
|
|
Keywords string
|
|
Provider domain.JobProvider
|
|
PageSize int
|
|
PageNumber int
|
|
}
|
|
|
|
func (r *JobRepository) GetAllJobs(filter JobFilter) ([]domain.Job, int, error) {
|
|
query := "SELECT id, title, company, company_link, location, job_link, job_timestamp FROM jobs"
|
|
countQuery := "SELECT COUNT(*) FROM jobs"
|
|
var conditions []string
|
|
var args []interface{}
|
|
paramIdx := 1
|
|
|
|
if filter.Location != "" {
|
|
conditions = append(conditions, fmt.Sprintf("location ILIKE $%d", paramIdx))
|
|
args = append(args, "%"+filter.Location+"%")
|
|
paramIdx++
|
|
}
|
|
|
|
if filter.Keywords != "" {
|
|
conditions = append(conditions, fmt.Sprintf("(title ILIKE $%d OR company ILIKE $%d)", paramIdx, paramIdx))
|
|
args = append(args, "%"+filter.Keywords+"%")
|
|
paramIdx++
|
|
}
|
|
|
|
if filter.Provider > 0 {
|
|
conditions = append(conditions, fmt.Sprintf("provider = $%d", paramIdx))
|
|
args = append(args, filter.Provider)
|
|
paramIdx++
|
|
}
|
|
|
|
whereClause := ""
|
|
if len(conditions) > 0 {
|
|
whereClause = " WHERE " + strings.Join(conditions, " AND ")
|
|
query += whereClause
|
|
countQuery += whereClause
|
|
}
|
|
|
|
var totalCount int
|
|
countArgs := make([]interface{}, len(args))
|
|
copy(countArgs, args)
|
|
if err := r.db.QueryRow(countQuery, countArgs...).Scan(&totalCount); err != nil {
|
|
return nil, 0, fmt.Errorf("error counting jobs: %v", err)
|
|
}
|
|
|
|
if filter.PageSize > 0 {
|
|
query += fmt.Sprintf(" LIMIT $%d", paramIdx)
|
|
args = append(args, filter.PageSize)
|
|
paramIdx++
|
|
|
|
if filter.PageNumber > 0 {
|
|
offset := (filter.PageNumber - 1) * filter.PageSize
|
|
query += fmt.Sprintf(" OFFSET $%d", paramIdx)
|
|
args = append(args, offset)
|
|
paramIdx++
|
|
}
|
|
}
|
|
|
|
rows, err := r.db.Query(query, args...)
|
|
if err != nil {
|
|
return nil, 0, fmt.Errorf("error querying jobs: %v", err)
|
|
}
|
|
defer rows.Close()
|
|
|
|
var jobs []domain.Job
|
|
for rows.Next() {
|
|
var job domain.Job
|
|
if err := rows.Scan(&job.ID, &job.Title, &job.Company, &job.CompanyLink, &job.Location, &job.JobLink, &job.JobPostTime); err != nil {
|
|
return nil, 0, fmt.Errorf("error scanning job row: %v", err)
|
|
}
|
|
jobs = append(jobs, job)
|
|
}
|
|
|
|
if err := rows.Err(); err != nil {
|
|
return nil, 0, fmt.Errorf("error iterating over job rows: %v", err)
|
|
}
|
|
|
|
return jobs, totalCount, nil
|
|
}
|
|
func (r *JobRepository) GetJobByID(id int) (*domain.Job, error) {
|
|
var job domain.Job
|
|
|
|
sqlStatement := `
|
|
SELECT id, title, company, company_link, location, job_link, job_timestamp
|
|
FROM jobs
|
|
WHERE id = $1
|
|
`
|
|
|
|
err := r.db.QueryRow(sqlStatement, id).Scan(
|
|
&job.ID,
|
|
&job.Title,
|
|
&job.Company,
|
|
&job.CompanyLink,
|
|
&job.Location,
|
|
&job.JobLink,
|
|
&job.JobPostTime,
|
|
)
|
|
|
|
if err == sql.ErrNoRows {
|
|
return nil, fmt.Errorf("job with ID %d not found", id)
|
|
}
|
|
|
|
if err != nil {
|
|
return nil, fmt.Errorf("error querying job: %v", err)
|
|
}
|
|
|
|
return &job, nil
|
|
}
|
|
|
|
func prepareQueryCreateBulk(s string, models []*domain.Job) (string, []interface{}) {
|
|
bf := bytes.Buffer{}
|
|
values := make([]interface{}, 0, len(models)*7)
|
|
|
|
for i, v := range models {
|
|
values = append(values, v.Title, v.Company,
|
|
v.CompanyLink, v.Location, v.JobLink, v.JobPostTime, v.Provider,
|
|
)
|
|
|
|
numFields := 7
|
|
n := i * numFields
|
|
|
|
bf.WriteString("(")
|
|
for j := 0; j < numFields; j++ {
|
|
bf.WriteString("$")
|
|
bf.WriteString(strconv.Itoa(n + j + 1))
|
|
if j < numFields-1 {
|
|
bf.WriteString(", ")
|
|
}
|
|
}
|
|
bf.WriteString(")")
|
|
if i < len(models)-1 {
|
|
bf.WriteString(", ")
|
|
}
|
|
}
|
|
|
|
return fmt.Sprintf(s, bf.String()), values
|
|
}
|
|
|
|
func (r *JobRepository) GetJobsByStatus(status domain.JobStatus) ([]domain.Job, error) {
|
|
query := `
|
|
SELECT id, title, company, company_link, location, job_link, job_timestamp
|
|
FROM jobs
|
|
WHERE status = $1
|
|
`
|
|
|
|
rows, err := r.db.Query(query, int(status))
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
defer rows.Close()
|
|
|
|
var jobs []domain.Job
|
|
for rows.Next() {
|
|
var job domain.Job
|
|
err := rows.Scan(
|
|
&job.ID,
|
|
&job.Title,
|
|
&job.Company,
|
|
&job.CompanyLink,
|
|
&job.Location,
|
|
&job.JobLink,
|
|
&job.JobPostTime,
|
|
)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
jobs = append(jobs, job)
|
|
}
|
|
|
|
return jobs, rows.Err()
|
|
}
|
|
|
|
func (r *JobRepository) UpdateJobStatus(jobID int64, status domain.JobStatus) error {
|
|
query := `UPDATE jobs SET status = $1 WHERE id = $2`
|
|
_, err := r.db.Exec(query, int(status), jobID)
|
|
return err
|
|
}
|