add parser

This commit is contained in:
2026-03-26 01:56:29 +07:00
parent 05a68caa87
commit ad7bc9f7dd
18 changed files with 899 additions and 10 deletions
+2 -2
View File
@@ -1,12 +1,12 @@
package app
import (
proto "pinned_message/proto"
"context"
proto "pinned_message/proto"
)
type Server struct {
proto.UnimplementedpinnedMessageServer
proto.UnimplementedPinnedMessageServer
}
func NewServer() *Server {
+37
View File
@@ -0,0 +1,37 @@
package config
import (
"os"
"path/filepath"
)
const (
ClientPort = ":8100"
FilePort = ":8120"
)
func GetScheduleFilepath() string {
return getFilepath("SCHEDULE_FILENAME", "data/schedule.json")
}
func getFilepath(env string, defaultFilepath string) string {
filepath := selectFilepath(env, defaultFilepath)
ensureDirExists(filepath)
return filepath
}
func selectFilepath(env string, defaultFilepath string) string {
filepath := os.Getenv(env)
if filepath != "" {
return filepath
}
return defaultFilepath
}
func ensureDirExists(filePath string) error {
dir := filepath.Dir(filePath)
if dir == "" || dir == "." || dir == "/" {
return nil
}
return os.MkdirAll(dir, 0755)
}
+32
View File
@@ -0,0 +1,32 @@
package models
import "time"
// День с событиями
type Day struct {
Date time.Time `json:"date"`
Performances []*DayPerformance `json:"performances"`
}
// Событие (мероприятия, праздник, поездка, событие)
type DayPerformance struct {
// Время сбора
TimeCollection string `json:"time_collection"`
// Время начала мероприятия
TimeStart string `json:"time_start"`
// Место
Place string `json:"place"`
// Наименование мероприятия
Name string `json:"name"`
// Номера
Numbers []*Number `json:"numbers"`
// Костюмы
Costumes string `json:"costumes"`
}
// Номер
type Number struct {
// Название
Name string `json:"name"`
}
@@ -0,0 +1,7 @@
package data_parser
import "context"
type IDataParser interface {
Parse(ctx context.Context, url string, v interface{}) error
}
+42
View File
@@ -0,0 +1,42 @@
package data_parser
import (
"context"
"fmt"
"net/http"
"regexp"
"github.com/gocarina/gocsv"
)
type parser struct{}
func NewGoogleTableScheduleParser() IDataParser {
return &parser{}
}
func (p *parser) Parse(_ context.Context, url string, v interface{}) error {
re := regexp.MustCompile(`/d/([a-zA-Z0-9-_]+)`)
matches := re.FindStringSubmatch(url)
if len(matches) < 2 {
return fmt.Errorf("Не удалось найти ID таблицы в ссылке")
}
sheetID := matches[1]
csvURL := fmt.Sprintf("https://docs.google.com/spreadsheets/d/%s/export?format=csv", sheetID)
resp, err := http.Get(csvURL)
if err != nil {
return fmt.Errorf("Ошибка при скачивании таблицы: %v", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return fmt.Errorf("Ошибка: статус код %d (убедитесь, что таблица публичная)", resp.StatusCode)
}
if err := gocsv.Unmarshal(resp.Body, v); err != nil {
return fmt.Errorf("Ошибка при парсинге CSV в структуру: %v", err)
}
return nil
}
@@ -0,0 +1,7 @@
package date_parser
import "time"
type IDateParser interface {
Parse(date string) (time.Time, error)
}
+53
View File
@@ -0,0 +1,53 @@
package date_parser
import (
"fmt"
"strconv"
"strings"
"time"
)
var (
ruMonths = map[string]time.Month{
"января": time.January,
"февраля": time.February,
"марта": time.March,
"апреля": time.April,
"мая": time.May,
"июня": time.June,
"июля": time.July,
"августа": time.August,
"сентября": time.September,
"октября": time.October,
"ноября": time.November,
"декабря": time.December,
}
)
type parser struct{}
func NewDateParser() IDateParser {
return &parser{}
}
func (p *parser) Parse(date string) (time.Time, error) {
parts := strings.Fields(date)
if len(parts) < 2 {
return time.Time{}, fmt.Errorf("Неверный формат даты, ожидалось 'День Месяц', получено: '%s'", date)
}
day, err := strconv.Atoi(parts[0])
if err != nil {
return time.Time{}, fmt.Errorf("Не удалось получить день: %v", err)
}
monthStr := strings.ToLower(parts[1])
month, ok := ruMonths[monthStr]
if !ok {
return time.Time{}, fmt.Errorf("Неизвестный месяц: %s", monthStr)
}
year := time.Now().Year()
return time.Date(year, month, day, 0, 0, 0, 0, time.Local), nil
}
@@ -0,0 +1,126 @@
package schedule_parser
import (
"context"
"log"
"pinned_message/internal/models"
"pinned_message/internal/modules/data_parser"
"pinned_message/internal/modules/date_parser"
"pinned_message/internal/services/schedule_storage"
"strings"
"time"
)
type performance struct {
Date string `csv:"ДАТА"`
Day string `csv:"день недели"`
Name string `csv:"НАЗВАНИЕ"`
Place string `csv:"МЕСТО"`
TimeCollection string `csv:"время СБОРА (для концерта)"`
TimeStart string `csv:"время НАЧАЛА"`
Numbers string `csv:"ЧТО ТАНЦУЕМ"`
Costumes string `csv:"КОСТЮМЫ"`
}
type ScheduleParser struct {
dataParser data_parser.IDataParser
dateParser date_parser.IDateParser
scheduleStorage schedule_storage.ScheduleStorage
}
func NewScheduleParser(
dataParser data_parser.IDataParser,
dateParser date_parser.IDateParser,
scheduleStorage schedule_storage.ScheduleStorage,
) *ScheduleParser {
return &ScheduleParser{
dataParser: dataParser,
dateParser: dateParser,
scheduleStorage: scheduleStorage,
}
}
func (p *ScheduleParser) Run(ctx context.Context) {
ticker := time.NewTicker(15 * time.Second) // TODO: set 1h
defer ticker.Stop()
sheetURL := "https://docs.google.com/spreadsheets/d/1v57bCAG764j1ULXDMb3amNFMzkkLmObKWsl5oE0Xq00/edit?gid=57461713#gid=57461713"
for {
select {
case <-ctx.Done():
return
case <-ticker.C:
days, err := p.parseSchedule(ctx, sheetURL)
if err != nil {
log.Printf("Error parse schedule: %s\n", sheetURL)
break
}
if err := p.scheduleStorage.SaveSchedule(days); err != nil {
log.Printf("Error save err: %s schedule: %s\n", err, sheetURL)
}
}
}
}
func (p *ScheduleParser) parseSchedule(ctx context.Context, sheetURL string) ([]*models.Day, error) {
var performances []performance
if err := p.dataParser.Parse(ctx, sheetURL, &performances); err != nil {
return nil, err
}
return p.mapSchedule(performances), nil
}
func (p *ScheduleParser) mapSchedule(performances []performance) []*models.Day {
days := []*models.Day{}
currentDay := &models.Day{}
for i, performance := range performances {
if performance.Name == "" || performance.Name == "-" {
continue
}
if performance.Date != "" {
if i > 0 {
days = append(days, currentDay)
currentDay = &models.Day{}
}
date, err := p.mapDate(performance.Date)
if err != nil {
panic(err)
}
currentDay.Date = date
}
currentDay.Performances = append(
currentDay.Performances,
&models.DayPerformance{
TimeCollection: performance.TimeCollection,
TimeStart: performance.TimeStart,
Place: performance.Place,
Name: performance.Name,
Numbers: p.mapNumbers(performance.Numbers),
Costumes: performance.Costumes,
},
)
}
days = append(days, currentDay)
return days
}
func (p *ScheduleParser) mapDate(date string) (time.Time, error) {
if date == "" {
return time.Time{}, nil
}
return p.dateParser.Parse(date)
}
func (p *ScheduleParser) mapNumbers(numbers string) []*models.Number {
names := strings.Split(numbers, ",")
res := make([]*models.Number, 0, len(names))
for _, name := range names {
res = append(
res,
&models.Number{
Name: name,
},
)
}
return res
}
@@ -0,0 +1,32 @@
package schedule_storage
import (
"encoding/json"
"log"
"os"
"pinned_message/internal/models"
)
type ScheduleStorage struct {
filepath string
}
func NewScheduleStorage(
filepath string,
) *ScheduleStorage {
return &ScheduleStorage{
filepath: filepath,
}
}
func (s *ScheduleStorage) SaveSchedule(days []*models.Day) error {
data, err := json.Marshal(days)
if err != nil {
return err
}
if err := os.WriteFile(s.filepath, data, 0x777); err != nil {
return err
}
log.Printf("save story to: %s", s.filepath)
return nil
}