rss2hook/rss2hook.go

// rss2hook is a simple utility which will make HTTP POST
// requests to remote web-hooks when new items appear in an RSS feed.
//
// Steve
// poesty
//

package main

import (
	"bytes"
	"encoding/json"
	"flag"
	"io"
	"log"
	"net/http"
	"os"
	"os/signal"
	"reflect"
	"syscall"
	"time"

	"github.com/SlyMarbo/rss"
	"github.com/robfig/cron"
)

// RSSEntry describes a single RSS feed and the corresponding hook
type RSSEntry struct {
	// The URL of the RSS/Atom feed
	Url string `json:"url,omitempty"`
	// The retry count
	Retry int `json:"retry,omitempty"`
	// The hook end-point
	Hook string `json:"hook,omitempty"`
}

// Loaded contains the loaded feeds + hooks, as read from the specified
// configuration file
var Loaded []RSSEntry

// Timeout is the (global) timeout we use when loading remote RSS
// feeds.
var Timeout time.Duration

// loadConfig loads the named configuration file and populates our
// `Loaded` list of RSS-feeds & Webhook addresses
func loadConfig(filename string) {
	file, err := os.Open(filename)
	if err != nil {
		log.Fatalf("loadConfig: Error opening %s - %s\n", filename, err.Error())
	}
	defer file.Close()

	err = json.NewDecoder(file).Decode(&Loaded)
	if err != nil {
		log.Fatalf("loadConfig: Error decoding %s - %s\n", filename, err.Error())
	}

	log.Printf("Loaded %d feeds\n", len(Loaded))
}

// checkFeeds is our work-horse.
//
// For each available feed it looks for new entries, and when founds
// triggers `notify` upon the resulting entry
func checkFeeds(feeds []rss.Feed) {
	//
	// For each thing we're monitoring
	//
	for i, feed := range feeds {

		// skip empty feed, probably because:
		// 1. feed initialization failed
		// 2. retry count exceeded
		if reflect.ValueOf(feed).IsZero() {
			continue
		}

		// Fetch the feed-contents
		err := feed.Update()

		if err != nil {
			log.Printf("checkFeeds: Error fetching %s - %s\n",
				Loaded[i].Url, err.Error())

			if Loaded[i].Retry--; Loaded[i].Retry == 0 {
				log.Printf("checkFeeds: Temporally diable feed - %s\n", Loaded[i].Url)
				feeds[i] = rss.Feed{}
			}
			// log.Printf("checkFeeds: refresh time - %s\n", feed.Refresh.Format(time.RFC3339))

			continue
		}

		if feed.Unread != 0 {
			notify(i, feed.Items[0]) // tbh idk
			feed.Unread = 0
			feed.Items = nil // gc
		}
	}
}

// notify actually submits the specified item to the remote webhook.
// The RSS-item is submitted as a JSON-object.
func notify(i int, item *rss.Item) {

	// We'll post the item as a JSON object.
	// So first of all encode it.
	jsonValue, err := json.Marshal(item)
	if err != nil {
		log.Fatalf("notify: Failed to encode JSON:%s\n", err.Error())
	}

	//
	// Post to the specified hook URL.
	//
	res, err := http.Post(Loaded[i].Hook,
		"application/json",
		bytes.NewBuffer(jsonValue))

	if err != nil {
		log.Printf("notify: Failed to POST to %s - %s\n",
			Loaded[i].Hook, err.Error()) // TODO: retry?
		return
	}

	//
	// OK now we've submitted the post.
	//
	// We should retrieve the status-code + body, if the status-code
	// is "odd" then we'll show them.
	//
	defer res.Body.Close()
	_, err = io.ReadAll(res.Body)
	if err != nil {
		log.Printf("notify: Failed to read response from %s - %s\n",
			Loaded[i].Hook, err.Error())
		return
	}
	status := res.StatusCode

	if status >= 400 {
		log.Printf("notify: Warning - Status code %d from %s\n", status, Loaded[i].Hook)
	}
}

// main is our entry-point
func main() {

	// Parse the command-line flags
	config := flag.String("config", "config.json", "The path to the configuration-file to read")
	timeout := flag.Duration("timeout", 5*time.Second, "The timeout used for fetching the remote feeds")
	schedule := flag.String("schedule", "@every 5m", "The cron schedule for fetching the remote feeds")
	startup := flag.Bool("startup", false, "Run on startup (first check)")
	flag.Parse()

	// Setup the default timeout and TTL
	// Remember that we respect spec, like `ttl` field in RSS 2.0
	rss.DefaultRefreshInterval = 10 * time.Second
	rss.DefaultFetchFunc = func(url string) (*http.Response, error) {
		client := &http.Client{Timeout: *timeout}
		return client.Get(url)
	}

	//
	// Load the configuration file
	//
	loadConfig(*config)

	//
	// Show the things we're monitoring
	//
	// for _, ent := range Loaded {
	// 	log.Printf("Monitoring feed %s\nPosting to %s\n\n",
	// 		ent.Feed, ent.Hook)
	// }

	//
	// Make the initial scan of feeds immediately to avoid waiting too
	// long for the first time.
	//
	feeds := make([]rss.Feed, len(Loaded))
	for i, ent := range Loaded {
		go func(i int, ent RSSEntry) {
			feed, err := rss.Fetch(ent.Url)
			if err != nil {
				log.Printf("main: Error fetching %s - %s\n",
					ent.Url, err.Error())
				return
			}
			// TODO: URL validiy checks
			if !*startup {
				feed.Unread = 0
				feed.Items = nil // gc
			}
			feeds[i] = *feed
		}(i, ent)
	}

	//
	// Now repeat that every five minutes or in custom schedule.
	//
	c := cron.New()
	c.AddFunc(*schedule, func() { checkFeeds(feeds) })
	c.Start()

	//
	// Now we can loop waiting to be terminated via ctrl-c, etc.
	//
	sigs := make(chan os.Signal, 1)
	done := make(chan bool, 1)
	signal.Notify(sigs, syscall.SIGINT, syscall.SIGTERM)
	go func() {
		<-sigs
		done <- true
	}()
	<-done
}