hscloud/hswaw/site/feeds.go

package main

import (
	"context"
	"encoding/xml"
	"fmt"
	"html/template"
	"net/http"
	"sort"
	"time"

	"github.com/golang/glog"
)

// This implements 'Atom' feed parsing. Honestly, this was written without
// looking at any spec. If it ever breaks, you know why.

var (
	// feedURLs is a map from an atom feed name to its URL. All the following
	// feeds will be combined and rendered on the main page of the website.
	feedsURLs = map[string]string{
		"blog": "https://blog.hackerspace.pl/feed/atom/",
	}
)

// atomFeed is a retrieved atom feed.
type atomFeed struct {
	XMLName xml.Name     `xml:"feed"`
	Entries []*atomEntry `xml:"entry"`
}

// atomEntry is an entry (eg. blog post) from an atom feed. It contains fields
// directly from the XML, plus some additional parsed types and metadata.
type atomEntry struct {
	XMLName      xml.Name      `xml:"entry"`
	Author       string        `xml:"author>name"`
	Title        template.HTML `xml:"title"`
	Summary      template.HTML `xml:"summary"`
	UpdatedRaw   string        `xml:"updated"`
	PublishedRaw string        `xml:"published"`
	Link         struct {
		Href string `xml:"href,attr"`
	} `xml:"link"`

	// Updated is the updated time parsed from UpdatedRaw.
	Updated time.Time
	// UpdatedHuman is a human-friendly representation of Updated for web rendering.
	UpdatedHuman string
	// Published is the published time parsed from PublishedRaw.
	Published time.Time
	// Source is the name of the feed that this entry was retrieved from. Only
	// set after combining multiple feeds together (ie. when returned from
	// getFeeds).
	Source string
}

// getAtomFeed retrieves a single Atom feed from the given URL.
func getAtomFeed(ctx context.Context, url string) (*atomFeed, error) {
	r, err := http.NewRequestWithContext(ctx, "GET", url, nil)
	if err != nil {
		return nil, fmt.Errorf("NewRequest(%q): %w", url, err)
	}
	res, err := http.DefaultClient.Do(r)
	if err != nil {
		return nil, fmt.Errorf("Do(%q): %w", url, err)
	}
	defer res.Body.Close()

	var feed atomFeed
	d := xml.NewDecoder(res.Body)
	if err := d.Decode(&feed); err != nil {
		return nil, fmt.Errorf("Decode: %w", err)
	}

	for i, e := range feed.Entries {
		updated, err := time.Parse(time.RFC3339, e.UpdatedRaw)
		if err != nil {
			return nil, fmt.Errorf("entry %d: cannot parse updated date %q: %v", i, e.UpdatedRaw, err)
		}
		published, err := time.Parse(time.RFC3339, e.PublishedRaw)
		if err != nil {
			return nil, fmt.Errorf("entry %d: cannot parse published date %q: %v", i, e.PublishedRaw, err)
		}
		e.Updated = updated
		e.Published = published
		e.UpdatedHuman = e.Updated.Format("02-01-2006")
		if e.Author == "" {
			e.Author = "Anonymous"
		}
	}

	return &feed, nil
}

// feedWorker runs a worker which retrieves all atom feeds every minute and
// updates the services' feeds map with the retrieved data. On error, the feeds
// are not updated (whatever is already cached in the map will continue to be
// available) and the error is logged.
func (s *service) feedWorker(ctx context.Context) {
	okay := false
	get := func() {
		feeds := make(map[string]*atomFeed)

		prev := okay
		okay = true
		for name, url := range feedsURLs {
			feed, err := getAtomFeed(ctx, url)
			if err != nil {
				glog.Errorf("Getting feed %v failed: %v", feed, err)
				okay = false
				continue
			}
			feeds[name] = feed
		}

		// Log whenever the first fetch succeeds, or whenever the fetch
		// succeeds again (avoiding polluting logs with success messages).
		if !prev && okay {
			glog.Infof("Feeds okay.")
		}

		// Update cached feeds.
		s.feedsMu.Lock()
		s.feeds = feeds
		s.feedsMu.Unlock()
	}
	// Perform initial fetch.
	get()

	// ... and update every minute.
	t := time.NewTicker(time.Minute)
	defer t.Stop()

	for {
		select {
		case <-ctx.Done():
			return
		case <-t.C:
			get()
		}
	}
}

// getFeeds retrieves the currently cached feeds and combines them into a
// single reverse-chronological timeline, annotating each entries' Source field
// with the name of the feed from where it was retrieved.
func (s *service) getFeeds() []*atomEntry {
	s.feedsMu.RLock()
	feeds := s.feeds
	s.feedsMu.RUnlock()

	var res []*atomEntry
	for n, feed := range feeds {
		for _, entry := range feed.Entries {
			e := *entry
			e.Source = n
			res = append(res, &e)
		}
	}
	sort.Slice(res, func(i, j int) bool {
		return res[j].Published.Before(res[i].Published)
	})
	return res
}