minstrel/internal/indexer/indexer.go

// Package indexer is what every search source answers with, and the Torznab
// client (M502 #5518). Ported from Fabled Librarian's internal/indexer.
//
// A client here holds ONE endpoint and knows nothing about the others. The
// fan-out across sources, the merge and what a source being down means belong
// to internal/acquire, because they are decisions about a search rather than
// about one source.
//
// Prowlarr is not served from here: it has no aggregate Torznab feed (`all`
// 404s like any id it does not have, measured by Fabled Librarian, #4304), so
// asking it everything at once means its native API, which internal/prowlarr
// speaks.
package indexer

import (
	"errors"
	"fmt"
	"io"
	"net/http"
	"strings"
	"time"
)

// ErrUnconfigured is a source with no address (or, for Prowlarr, no key): not
// a failure, an instance that does not search there.
var ErrUnconfigured = errors.New("indexer is not configured")

// ErrUnreachable is a search that could not be made: no network, a timeout, a
// malformed answer, a status outside 2xx.
var ErrUnreachable = errors.New("indexer unreachable")

// ErrAuth is a key the source refused. Separate from ErrUnreachable because
// it needs a different sentence and a different action: a refused key is a
// field in Settings, an outage is something to wait out.
var ErrAuth = errors.New("the API key was refused")

// UserAgent names Minstrel in the operator's own Prowlarr and indexer logs.
const UserAgent = "Minstrel (self-hosted music server; release search)"

// MusicCategories are the Newznab categories a music search asks for: 3000
// Audio, with its MP3 (3010) and Lossless (3040) children alongside, because
// an indexer that files music under a child answers only for the child.
// 3030 (Audiobook) is deliberately absent.
var MusicCategories = []int{3000, 3010, 3040}

// Query is one search: the words, and which categories to ask in. Empty
// Categories asks every category.
type Query struct {
	Text       string
	Categories []int
}

// Result is one search result, as the source described it.
type Result struct {
	Title string
	// InfoHash is the torrent's identity, lowercased; empty when the source
	// gave none.
	InfoHash string
	// Magnet is a magnet URI when the source published one.
	Magnet string
	// DownloadURL is a .torrent URL when it did not. It must be fetched
	// through the source that offered it (it holds the tracker credentials),
	// never handed to the download client.
	DownloadURL string
	SizeBytes   int64
	Category    int
	// Indexer is the tracker BEHIND the source, as the source named it.
	Indexer string
	// Source is the acquisition_services code that was asked: where a
	// .torrent fetch must be routed back to.
	Source    string
	Published time.Time
	// Seeders and Peers as reported. A feed that has not sampled a swarm
	// reports a nominal 1, so these are not health unless SwarmKnown.
	Seeders int
	Peers   int
	// SwarmKnown says Seeders was measured, which only FabledExchange's own
	// search can say.
	SwarmKnown bool
	// FilesCount is how many files the torrent holds, 0 when not said.
	FilesCount int
}

// Send sends req through c, stopping at a redirect to a magnet (returned as a
// *MagnetLink), and reads at most 8 MB of the answer. A call that never got
// an answer is ErrUnreachable; the status is the caller's to read.
func Send(c *http.Client, req *http.Request) (*http.Response, []byte, error) {
	resp, err := StopAtMagnet(c).Do(req)
	if m, ok := AsMagnet(err); ok {
		return nil, nil, &MagnetLink{URI: m}
	}
	if err != nil {
		return nil, nil, fmt.Errorf("%w: %v", ErrUnreachable, err)
	}
	defer func() { _ = resp.Body.Close() }()
	body, err := io.ReadAll(io.LimitReader(resp.Body, 8<<20))
	if err != nil {
		return nil, nil, fmt.Errorf("%w: read body: %v", ErrUnreachable, err)
	}
	return resp, body, nil
}

// StatusError is what a non-2xx status means for a search source: a 429 is
// RateLimited naming who, anything else ErrUnreachable. Nil for a 2xx.
func StatusError(resp *http.Response, who string) error {
	switch {
	case resp.StatusCode == http.StatusTooManyRequests:
		return &RateLimited{Who: who, RetryAfter: RetryAfter(resp.Header, time.Now())}
	case resp.StatusCode < 200 || resp.StatusCode >= 300:
		return fmt.Errorf("%w: HTTP %d", ErrUnreachable, resp.StatusCode)
	}
	return nil
}

// NormalizeHash is a torrent's identity in its one canonical spelling. Every
// comparison of infohashes goes through it, so a hash differing only in case
// is never two releases.
func NormalizeHash(s string) string {
	return strings.ToLower(strings.TrimSpace(s))
}

// IsInfoHash says whether s is shaped like a v1 (40) or v2 (64) hex infohash.
func IsInfoHash(s string) bool {
	s = NormalizeHash(s)
	if len(s) != 40 && len(s) != 64 {
		return false
	}
	for _, r := range s {
		if (r < '0' || r > '9') && (r < 'a' || r > 'f') {
			return false
		}
	}
	return true
}

// MagnetHash pulls the infohash out of a magnet link, or returns "". It does
// not validate it.
func MagnetHash(magnet string) string {
	const prefix = "urn:btih:"
	at := strings.Index(magnet, prefix)
	if at < 0 {
		return ""
	}
	hash := magnet[at+len(prefix):]
	if cut := strings.IndexAny(hash, "&/?"); cut >= 0 {
		hash = hash[:cut]
	}
	return strings.ToLower(hash)
}

// FirstLine is an error's first line, for a sentence on the settings page.
func FirstLine(s string) string {
	if i := strings.IndexByte(s, '\n'); i >= 0 {
		return s[:i]
	}
	return s
}