feedget/mastodon.go

package main

import (
	"bytes"
	"context"
	"fmt"
	"io"
	"log"
	"net/http"
	"sort"
	"strings"
	"sync"
	"time"

	"github.com/PuerkitoBio/goquery"
	"golang.org/x/net/html"
	"golang.org/x/net/html/atom"
)

const maxResponseSize = 1e6 // 1MB

// https://docs.joinmastodon.org/spec/microformats/

// Root elements (h-*)
//
// h-feed
//     Represents a stream of entries. Attached to a profile's toots. Also
//     attached to the parent thread within detailed status views.
//
// h-entry
//     Represents episodic or date stamped online content. Attached to a status.
//
// URL properties (u-*)
//
// u-photo
//     Within h-card, represents the profile picture. Attached to the avatar image.
//
// u-uid
//     Within h-entry or h-cite, represents a universally unique identifier.
//     Attached to timestamp link.
//
// u-url
//     Within h-entry or h-cite, represents the status permalink. Attached to
//     timestamp link. Within h-card, represents the profile permalink.
//     Attached to display name link.
//
// Datetime properties (dt-*)
//
// dt-published
//     Within h-entry or h-cite, represents the date and time at which the
//     status was published. Attached to data element with value attribute.
//
// Element tree (e-*)
//
// e-content
//     Within h-entry or h-cite, represents the content of the status. Attached to status content.
//

// I learned after writing this that mastodon also has RSS feeds
// e.g. https://tiny.tilde.website/@magical.rss

type MastoSource struct {
	Items     []*MastoItem
	Title     string
	URL       string
	LastFetch time.Time
	Error     error

	LastStatusCode int
	LastStatus     string

	mu sync.Mutex
}

var _ Source = &MastoSource{}

type MastoFeed struct {
	Title string
	Items []*MastoItem
}

type MastoItem struct {
	Title           string `json:"title,omitempty"`
	Content         string `json:"content,omitempty"`
	Link            string `json:"link,omitempty"`
	PublishedString string `json:"published,omitempty"`
	Author          string `json:"author,omitempty"`
	IsBoost         bool   `json:"is_boost,omitempty"`
}

func NewMastoSource(url string) *MastoSource {
	return &MastoSource{
		URL: url,
	}
}

func (src *MastoSource) update(ctx context.Context) {
	src.mu.Lock()
	defer src.mu.Unlock()

	req, err := http.NewRequestWithContext(ctx, "GET", src.URL, nil)
	if err != nil {
		src.Error = fmt.Errorf("error fetching %q: %w", src.URL, err)
		log.Println(src.Error)
		return // return err?
	}
	req.Header.Set("User-Agent", UserAgent)
	// TODO: If-Modified-Since, Etag

	resp, err := http.DefaultClient.Do(req)
	if err != nil {
		err := fmt.Errorf("error fetching %q: %w", src.URL, err)
		log.Println(err)
		src.Error = err
		return // return err?
	}

	if resp != nil && resp.Body != nil {
		defer func() {
			err := resp.Body.Close()
			if err != nil {
				log.Printf("error closing response body for %q: %v", src.URL, err)
			}
		}()
	}

	src.LastStatusCode = resp.StatusCode
	src.LastStatus = resp.Status
	if resp.StatusCode != 200 {
		src.Error = fmt.Errorf("error fetching %q: status %s", src.URL, resp.Status)
		log.Println(src.Error)
		return
	}

	// TODO: cache body
	body := MaxBytesReader(resp.Body, maxResponseSize)

	feed, err := parseMicroformats(body)
	if err != nil {
		err := fmt.Errorf("error parsing %q: %w", src.URL, err)
		log.Println(err)
		src.Error = err
		return // return err?
	}

	items := feed.Items
	sort.Slice(items, func(i, j int) bool {
		return items[i].PublishedString >= items[j].PublishedString
	})
	src.Title = feed.Title
	src.Items = items
	src.LastFetch = time.Now()
	src.Error = nil
}

func parseMicroformats(r io.Reader) (*MastoFeed, error) {
	doc, err := goquery.NewDocumentFromReader(r)
	if err != nil {
		return nil, err
	}
	feed := new(MastoFeed)
	if doc.Find(".h-feed").Length() == 0 {
		return nil, fmt.Errorf("no feed content found")
	}
	feed.Title = doc.Find(".h-feed > .p-name").First().AttrOr("value", "")
	doc.Find(".h-feed").Find(".h-entry, .h-cite").Each(func(i int, elem *goquery.Selection) {
		cw := strings.TrimSpace(text(elem.Find(".p-summary").First()))
		// TODO: move this logic to GetItems
		if cw != "" {
			cw = "[" + cw + "] "
		}
		feed.Items = append(feed.Items, &MastoItem{
			Title:           "",
			Content:         cw + text(elem.Find(".e-content").First()),
			Link:            elem.Find("a.u-url.u-uid").AttrOr("href", ""),
			Author:          text(elem.Find(".p-author .p-name").First()),
			PublishedString: elem.Find("data.dt-published").AttrOr("value", ""),
			IsBoost:         elem.HasClass("h-cite"),
		})
	})
	return feed, nil
}

// Text gets the combined text contents of each element in the set of matched
// elements, including their descendants.
func text(s *goquery.Selection) string {
	var buf bytes.Buffer

	// Slightly optimized vs calling Each: no single selection object created
	var f func(*html.Node)
	f = func(n *html.Node) {
		if n.Type == html.TextNode {
			// Keep newlines and spaces, like jQuery
			buf.WriteString(n.Data)
		} else if n.Type == html.ElementNode && n.DataAtom == atom.Br {
			//buf.WriteString("\n")
			buf.WriteString(" ")
		} else if n.Type == html.ElementNode && n.DataAtom == atom.P && n.PrevSibling != nil {
			//buf.WriteString("\n\n")
			buf.WriteString(" ")
		} else if n.Type == html.ElementNode && (n.DataAtom == atom.Script || n.DataAtom == atom.Style || n.DataAtom == atom.Template) {
			// nothing
		}
		if n.FirstChild != nil {
			for c := n.FirstChild; c != nil; c = c.NextSibling {
				f(c)
			}
		}
	}
	for _, n := range s.Nodes {
		f(n)
	}

	return buf.String()
}

func (src *MastoSource) GetTitle() string { return src.Title }
func (src *MastoSource) GetError() error  { return src.Error }

func (src *MastoSource) GetItems() (items []Item) {
	for _, x := range src.Items {
		text := x.Content
		if x.IsBoost {
			text = "RT @" + x.Author + ": " + text
		}
		d, _ := time.Parse(time.RFC3339, x.PublishedString)
		items = append(items, Item{
			Date: d,
			Link: x.Link,
			Text: text,
		})
	}
	return
}