Add validation package (#19)

Closes #7

Added a simple validation package we can expand on. For now we can use it to validate links that get sent to the bot and make sure that:
1. It's an actual url
2. It's a url for a service we support

Also added a make file to run tests and other tasks we might need in the future.

Added a github action to run tests.

Reviewed-on: #19
Co-authored-by: Javier Feliz <me@javierfeliz.com>
Co-committed-by: Javier Feliz <me@javierfeliz.com>

.gitea/workflows/tests.yml

@@ -0,0 +1,21 @@
+name: Run Go Tests
+
+on:
+  pull_request:
+    branches: [master]
+
+jobs:
+  test:
+    name: Run Tests
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout Code
+        uses: actions/checkout@v4
+
+      - name: Set up Go
+        uses: actions/setup-go@v4
+        with:
+          go-version: '1.22'
+
+      - name: Run Tests
+        run: make test

.gitignore

@@ -1,6 +1,7 @@
 testfile.wav
 play.opus
 output.opus
+*.of
 .env
 *.wav
 *.opus

Makefile

@@ -0,0 +1,3 @@
+test:
+	@echo "Running tests for main and all packages"
+	go test ./...

README.md

@@ -6,7 +6,7 @@ Currently in development.
 - `ffmpeg` make sure libopus is included
 - `yt-dlp`
 
-## Current testing steps
+## Running locally 
 
 Copy .env.example to .env
 
@@ -14,4 +14,8 @@ Populate the discord bot keys
 
 `go run .`
 
-type `!` in any channel in the discord while you're in a voice channel.
+type `!` in any channel in the discord while you're in a voice channel.
+
+## Running tests locally
+
+`make test`

main.go

@@ -2,7 +2,6 @@ package main
 
 import (
 	"fmt"
-	"io"
 	"log"
 	"os"
 	"os/exec"
@@ -10,6 +9,7 @@ import (
 	"syscall"
 	"time"
 
+	"git.thegrind.dev/thegrind/papibot/pkg/opusframes"
 	dg "github.com/bwmarrin/discordgo"
 	"github.com/javif89/dotenv"
 )
@@ -108,79 +108,28 @@ func playCommand(s *dg.Session, i *dg.InteractionCreate) {
 }
 
 func playOnVoiceChannel(voiceChannel *dg.VoiceConnection) {
-	log.Println("Starting ffmpeg stream")
-	// I got the original implementation from:
-	// https://github.com/nhooyr/botatouille/blob/7e1cd9d5a8d517fd43fd11599b2a62bf832a5c96/cmd/botatouille/music/music.go#L62-L104
-	// after hours of searching.
-	ffmpeg := exec.Command(
-		"ffmpeg",
-		"-i", "vid.webm",
-		"-hide_banner",
-		"-loglevel", "quiet",
-		"-f", "data",
-		"-map", "0:a",
-		"-ar", "48k",
-		"-ac", "2",
-		"-acodec", "libopus",
-		"-b:a", "128k",
-		"pipe:1")
-
-	ffmpegOut, err := ffmpeg.StdoutPipe()
+	log.Println("Decoding")
+	start := time.Now()
+	frames, err := opusframes.Decode("vid.of")
 	if err != nil {
-		log.Fatal(err)
+		log.Println("Decoding error: ", err)
+		return
 	}
+	duration := time.Since(start)
+	log.Printf("Decoding took: %s", duration)
 
-	err = ffmpeg.Start()
-	if err != nil {
-		log.Fatal(err)
-	}
-
-	packets := [][]byte{}
-
-	startTime := time.Now()
-
-	bytes_sent := 0
-	for {
-		// I read in the RFC that frames will not be bigger than this size
-		p := make([]byte, 960)
-		n, err := ffmpegOut.Read(p)
-		bytes_sent = bytes_sent + n
-
-		if err != nil {
-			log.Printf("Bytes: %d", n)
-			if err == io.EOF {
-				log.Println("Done streaming")
-				break
-			}
-			log.Fatal(err)
-		}
-
-		packets = append(packets, p[:n])
-	}
-
-	elapsedTime := time.Since(startTime)
-
-	log.Printf("bytes sent = %d", bytes_sent)
-	log.Printf("Took %s seconds to run", elapsedTime)
+	// for _, f := range frames {
+	// 	log.Printf("Got frame. Size: %d", len(f))
+	// }
 
 	voiceChannel.Speaking(true)
 	time.Sleep(time.Second * 2)
 	log.Println("Playing sound")
 
-	startTime = time.Now()
-	bytes_sent = 0
-	for _, p := range packets {
+	for _, p := range frames {
 		log.Printf("Sending packet: %d bytes", len(p))
-		bytes_sent += len(p)
-
 		voiceChannel.OpusSend <- p
 	}
-
-	elapsedTime = time.Since(startTime)
-	log.Printf("Packets = %d", len(packets))
-	log.Printf("Network bytes sent = %d", bytes_sent)
-	log.Printf("Took %s seconds to run", elapsedTime)
-	log.Println("Ended stream")
 }
 
 func handleSlashCommand(s *dg.Session, i *dg.InteractionCreate) {
@@ -233,5 +182,10 @@ func downloadVideo(url string) error {
 	err := cmd.Run()
 	log.Println("Downloaded")
 
+	start := time.Now()
+	opusframes.Encode("vid.webm", "vid.of")
+	duration := time.Since(start)
+	log.Printf("Encoding took: %s", duration)
+
 	return err
 }

pkg/opusframes/opusframes.go

@@ -0,0 +1,130 @@
+package opusframes
+
+// Package to encode/decode .of files.
+// The .of file format will allow us to cache/preprocess
+// files from the queue and save it in a format that we
+// can quickly read and send to discord.
+// The format spect is very simple:
+// frame size: 2 bytes
+// frame data: Up to 1275 bytes
+// data will be stored in little endian format
+import (
+	"bytes"
+	"encoding/binary"
+	"errors"
+	"io"
+	"log"
+	"os"
+	"os/exec"
+)
+
+type ErrFFMPEG error
+type ErrEncoding error 
+
+// TODO: Allow streaming frames as they are encoded instead
+// of having to wait for the whole file. 
+// We can probably do this by taking an io.Reader as input
+// instead of just a file path. Then we return an io.Writer
+// and allow the user to either stream each encoded frame
+// or save it to a file. We could also make the Decode
+// function take an io.Reader so these functions can
+// just be piped into each other.
+
+// TODO: Can we pack the frames into chunks of 960? I wonder
+// if that would make playback smoother or just fuck the
+// audio. We can try it at some point.
+func Encode(input, output string) error {
+	ffmpeg := exec.Command(
+		"ffmpeg",
+		"-i", input,
+		"-hide_banner",
+		"-loglevel", "quiet",
+		"-f", "data",
+		"-map", "0:a",
+		"-ar", "48k",
+		"-ac", "2",
+		"-acodec", "libopus",
+		"-b:a", "128k",
+		"pipe:1")
+
+	ffmpegOut, err := ffmpeg.StdoutPipe()
+	if err != nil {
+		return ErrFFMPEG(err)
+	}
+
+	err = ffmpeg.Start()
+	if err != nil {
+		return ErrFFMPEG(err)
+	}
+
+	var fileBuffer bytes.Buffer
+
+	for {
+		p := make([]byte, 960)
+		n, err := ffmpegOut.Read(p)
+
+		binary.Write(&fileBuffer, binary.LittleEndian, uint16(n))
+		binary.Write(&fileBuffer, binary.LittleEndian, p[:n])
+
+		log.Printf("Frame of: %d", n)
+
+		if err != nil {
+			if err == io.EOF {
+				break
+			}
+			
+			return ErrEncoding(err)
+		}
+	}
+
+	f, err := os.OpenFile(output, os.O_CREATE|os.O_RDWR|os.O_TRUNC, 0644)
+	if err != nil {
+		log.Println(err)
+	}
+	defer f.Close()
+
+	f.Write(fileBuffer.Bytes())
+
+	return nil
+}
+
+var ErrMalformedFile error = errors.New("malformed file")
+
+// TODO: Allow streaming frames as they are decoded instead
+// of having to decode the whole file.
+func Decode(input string) ([][]byte, error){
+	f, err := os.Open(input)
+	if err != nil {
+		return nil, err
+	}
+	defer f.Close()
+
+	// We read the first 2 bytes for the length
+	// then n bytes after that until we're done
+	frames := [][]byte{}
+	for {
+		var frameSize uint16
+		err = binary.Read(f, binary.LittleEndian, &frameSize)
+
+		if err != nil {
+			if err == io.EOF {
+				return frames, nil
+			} 
+				
+			return nil, err
+		}
+
+		frame := make([]byte, frameSize)
+		n, err := io.ReadFull(f, frame)
+
+		if n != int(frameSize) {
+			return nil, ErrMalformedFile	
+		}
+
+		if err != nil {
+			return nil, err 
+		}
+
+		frames = append(frames, frame)
+	}
+}

pkg/validation/validation.go

@@ -0,0 +1,54 @@
+package validation
+
+import (
+	"errors"
+	"fmt"
+	neturl "net/url"
+	"strings"
+)
+
+// Errors
+var (
+	ErrNotAUrl = errors.New("not a URL")
+	ErrIncorrectProtocol = errors.New("incorrect url protocol")
+	ErrServiceUnsupported = errors.New("not a URL")
+)
+
+// Music hosts that we support
+var musicHosts = []string{
+	"youtube.com",
+	"www.youtube.com",
+}
+
+func IsUrl(url string) (bool, *neturl.URL) {
+	// If a URL has no scheme, this will fail.
+	// So we'll add one if not present
+	if !strings.Contains(url, "://") {
+		url = fmt.Sprintf("https://%s", url)
+	}
+
+	parsed, err := neturl.ParseRequestURI(url) 
+
+	return (err == nil), parsed
+}
+
+func IsMusicUrl(url string) (bool, error) {
+	isUrl, parsed := IsUrl(url)
+
+	if !isUrl {
+		return false, ErrNotAUrl 
+	}
+
+	// If we have a scheme and it's not http/https we fail
+	if parsed.Scheme != "" && !strings.Contains(parsed.Scheme, "http") {
+		return false, ErrIncorrectProtocol
+	}
+
+	for _, host := range musicHosts {
+		if host == parsed.Host {
+			return true, nil
+		}
+	}
+
+	return false, ErrServiceUnsupported
+}

pkg/validation/validation_test.go

@@ -0,0 +1,85 @@
+package validation
+
+import (
+	"testing"
+)
+
+func TestIsUrl(t *testing.T) {
+	is, _ := IsUrl("definitely not a url")
+
+	if is {
+		t.Error("Non-url text detected as URL")
+	}
+
+	is, _ = IsUrl("https://example.com")
+
+	if !is {
+		t.Error("URL not detected as URL")
+	}
+}
+
+func TestSchemeHandling(t *testing.T) {
+	// No scheme but valid url
+	is, _ := IsUrl("youtube.com")
+
+	if !is {
+		t.Error("URL without scheme came back as not a url")
+	}
+
+	// Preserve scheme
+	is, parsed := IsUrl("ftp://youtube.com")
+
+	if !is {
+		t.Error("URL without scheme came back as not a url")
+	}
+
+	if parsed.Scheme != "ftp" {
+		t.Error("URL scheme was replaced incorrectly")
+	}
+}
+
+func TestSupportedMusicUrls(t *testing.T) {
+	// Test actual music url
+	for _, url := range musicHosts {
+		is, _ := IsMusicUrl(url)
+
+		if !is {
+			t.Error("Supported service was detected as unsupported")
+		}
+	}
+}
+
+func TestIsMusicUrlErrors(t *testing.T) {
+	// Not a URL
+	is, err := IsMusicUrl("not a url")
+
+	if is {
+		t.Error("Non-URL was detected as url")
+	}
+
+	if err != ErrNotAUrl {
+		t.Error("Incorrect error returned for Non-url link")
+	}
+
+	// Incorrect protocol
+	is, err = IsMusicUrl("ssh://youtube.com")
+
+	if is {
+		t.Error("Incorrect protocol was not caught")
+	}
+
+	if err != ErrIncorrectProtocol {
+		t.Error("Incorrect error returned for incorrect protocol")
+	}
+
+	// Unsupported service
+	is, err = IsMusicUrl("https://www.deezer.com/")
+
+	if is {
+		t.Error("Unsupported service was not caught")
+	}
+
+	if err != ErrServiceUnsupported {
+		t.Error("Unsupported service did not return the correct error")
+	}
+}