working

TODO.md

@@ -9,12 +9,12 @@
 * ✓ Posting to telegram
 * ✓ Telegram rate limit: 20 msgs/min  https://core.telegram.org/bots/faq#my-bot-is-hitting-limits-how-do-i-avoid-this
     https://gist.github.com/zdebra/10f0e284c4672e99f0cb767298f20c11
-* First Run skip all but first n
-* Folder creation
+* ✓ First Run skip all but first n
+* ✓ Folder creation
+* ✓ Wrap content inside item
 * Change flags to os.Args[1] for config path
 * Cleanup
 * Make Rhash optional
-* Wrap content inside item
 
 ## Issues
 * ✓ Last record is not written fully to csv: *Fixed*. Had to Flush writer

app/app.go

@@ -73,16 +73,34 @@ func ProcessFeed(feed FeedCfg, scheduler Scheduler, dbDir string, tgram telegram
 
 	scheduler.Good(feed.Name)
 
-	filteredEntries, err := db.Filter(entries)
+	var records []Record
+	var newEntries []FeedEntry
+	if db.IsNewFeed() {
+		ftl := int(feed.FirstTimeLimit)
+		if feed.FirstTimeLimit == NoLimit || len(entries) <= ftl {
+			newEntries = entries
+		} else {
+			var filteredEntries []FeedEntry
+			newEntries, filteredEntries = entries[:ftl], entries[ftl:]
+			for _, entry := range filteredEntries {
+				records = append(records, Record{
+					Time:      time.Now(),
+					Status:    Filtered,
+					Filter:    "FirstTime",
+					FeedEntry: entry,
+				})
+			}
+		}
+	} else {
+		newEntries, err = db.Filter(entries)
 		if err != nil {
 			log.Error("failed to filter entries", "feed", feed.Name, "error", err)
 		}
+	}
 
-	var records []Record
-	for _, entry := range filteredEntries {
-		now := time.Now()
+	for _, entry := range newEntries {
 		r := Record{
-			Time:      now,
+			Time:      time.Now(),
 			FeedEntry: entry,
 		}
 		err := tgram.SendLink(entry.Link, feed.Channel, feed.Rhash)

app/config.go

@@ -7,6 +7,10 @@ import (
 	"gopkg.in/yaml.v3"
 )
 
+type FeedLimit int
+
+const NoLimit FeedLimit = -1
+
 type FeedCfg struct {
 	Name           string    `yaml:"name"`
 	Channel        string    `yaml:"channel"`
@@ -14,16 +18,18 @@ type FeedCfg struct {
 	Url            string    `yaml:"url"`
 	Cron           string    `yaml:"cron"`
 	Proxy          string    `yaml:"proxy"`
+	FTL            *int      `yaml:"first_time_limit"`
+	FirstTimeLimit FeedLimit `yaml:"-"`
 }
 
 type Config struct {
 	Proxy             string    `yaml:"proxy"`
 	TelegramProxy     string    `yaml:"telegram_proxy"`
-	//TODO: read from credential file
-	TelegramAuthToken string    `yaml:"telegram_auth_token"`
+	TelegramAuthToken string    `yaml:"telegram_auth_token"` //TODO: read from credential file
 	DataDir           string    `yaml:"data_dir"`
 	LastSuccessPath   string    `yaml:"last_loaded_path"`
 	DbDir             string    `yaml:"db_dir"`
+	FTL               *int      `yaml:"first_time_limit"`
 	Feeds             []FeedCfg `yaml:"feeds"`
 }
 
@@ -53,6 +59,17 @@ func ParseConfig(configPath string) (*Config, error) {
 		c.DbDir = path.Join(c.DataDir, "feed_data")
 	}
 
+	err = os.MkdirAll(c.DbDir, 0755)
+
+	if err != nil {
+		return nil, err
+	}
+
+	err = os.MkdirAll(c.DataDir, 0755)
+	if err != nil {
+		return nil, err
+	}
+
 	if c.Proxy != "" {
 		if c.TelegramProxy == "" {
 			c.TelegramProxy = c.Proxy
@@ -74,6 +91,13 @@ func ParseConfig(configPath string) (*Config, error) {
 		if feedCfg.Proxy == "NONE" {
 			feedCfg.Proxy = ""
 		}
+		if feedCfg.FTL != nil {
+			feedCfg.FirstTimeLimit = FeedLimit(*feedCfg.FTL)
+		} else if c.FTL != nil {
+			feedCfg.FirstTimeLimit = FeedLimit(*c.FTL)
+		} else {
+			feedCfg.FirstTimeLimit = NoLimit
+		}
 	}
 
 	return &c, nil

app/db.go

@@ -26,10 +26,12 @@ const TimeFormat string = "2006-01-02T15:04:05.999999999-07:00"
 type Record struct {
 	Time      time.Time
 	Status    Status
+	Filter    string
 	FeedEntry FeedEntry
 }
 
 type DB interface {
+	IsNewFeed() bool
 	Filter(entries []FeedEntry) ([]FeedEntry, error)
 	Save([]Record) error
 }
@@ -39,6 +41,10 @@ type db struct {
 	seenLinks map[string]struct{}
 }
 
+func (d *db) IsNewFeed() bool {
+	return len(d.seenLinks) == 0
+}
+
 func NewDB(storageDir, feedName string) (DB, error) {
 	dbPath := path.Join(storageDir, fmt.Sprintf("%s.csv", feedName))
 	db := db{dbPath: dbPath}
@@ -91,7 +97,7 @@ func (d *db) Save(records []Record) error {
 			"Date",
 			"Link",
 			"Status",
-			"FilteredBy",
+			"Filter",
 			"Content",
 		})
 	}
@@ -100,8 +106,8 @@ func (d *db) Save(records []Record) error {
 			r.Time.Format(TimeFormat),
 			r.FeedEntry.Link,
 			string(r.Status),
-			"-",
-			strings.ReplaceAll(r.FeedEntry.Content, "\n", " "),
+			r.Filter,
+			fmt.Sprintf("<item>%s</item>", strings.ReplaceAll(r.FeedEntry.Content, "\n", " ")),
 		})
 	}
 	csvw.Flush()

app/schedule.go

@@ -23,7 +23,7 @@ type scheduler struct {
 }
 
 func NewScheduler(filePath string) (Scheduler, error) {
-	s := scheduler{filePath: filePath}
+	s := scheduler{filePath: filePath, lastSuccessTime: map[string]time.Time{}}
 	data, err := os.ReadFile(filePath)
 	if err != nil {
 		if !errors.Is(err, os.ErrNotExist) {

exp/slice/main.go

@@ -0,0 +1,14 @@
+package main
+
+import "fmt"
+
+func main() {
+	var arr []int
+	for i := 0; i < 10; i++ {
+		arr = append(arr, i)
+	}
+	fmt.Println(arr)
+	fmt.Println(arr[:0], arr[0:])
+	fmt.Println(arr[:4], arr[4:])
+	fmt.Println(arr[:15], arr[15:])
+}

telegram/telegram.go

@@ -15,10 +15,6 @@ import (
 	"golang.org/x/time/rate"
 )
 
-func main() {
-	fmt.Println("vim-go")
-}
-
 type TelegramSender interface {
 	SendLink(link, channel, rhash string) error
 }