From 8002dd0912e8988816cbb6cdec867861a6d4e5f4 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E5=BC=A0=E6=96=8C?= <259278618@qq.com>
Date: Mon, 13 Jul 2026 09:35:11 +0800
Subject: [PATCH] 1

---
 Dockerfile                          |   2 +-
 go.mod                              |   1 +
 go.sum                              |   4 +
 shortdrama/service/drama_service.go | 251 +++++++++++++++++++++++++---
 4 files changed, 232 insertions(+), 26 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index caf83b3..561f976 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -13,7 +13,7 @@ RUN go build -ldflags="-s -w" -o main ./main.go
 
 FROM alpine:3.19
 RUN sed -i 's/dl-cdn.alpinelinux.org/mirrors.aliyun.com/g' /etc/apk/repositories \
-  && apk add --no-cache ca-certificates tzdata ffmpeg
+  && apk add --no-cache ca-certificates tzdata
 ENV TZ=Asia/Shanghai
 RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone
 WORKDIR /app
diff --git a/go.mod b/go.mod
index 50695dd..32af510 100644
--- a/go.mod
+++ b/go.mod
@@ -3,6 +3,7 @@ module video-factory
 go 1.26.1
 
 require (
+	github.com/Eyevinn/mp4ff v0.53.0
 	github.com/gogf/gf/contrib/drivers/sqlite/v2 v2.10.2
 	github.com/gogf/gf/v2 v2.10.2
 )
diff --git a/go.sum b/go.sum
index 57341e6..56dc234 100644
--- a/go.sum
+++ b/go.sum
@@ -1,5 +1,7 @@
 github.com/BurntSushi/toml v1.6.0 h1:dRaEfpa2VI55EwlIW72hMRHdWouJeRF7TPYhI+AUQjk=
 github.com/BurntSushi/toml v1.6.0/go.mod h1:ukJfTF/6rtPPRCnwkur4qwRxa8vTRFBF0uk2lLoLwho=
+github.com/Eyevinn/mp4ff v0.53.0 h1:aK4OF9gFwjrqvHRf0znEMCUVupgRFR1LB8OVpw9z1RA=
+github.com/Eyevinn/mp4ff v0.53.0/go.mod h1:AhC+bOI7GSZmzuN4zFY9U76qMedbHI+8BdQXWrC9+8U=
 github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
 github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
 github.com/clbanning/mxj/v2 v2.7.0 h1:WA/La7UGCanFe5NpHF0Q3DNtnCsVoxbPKuyBNHWRyME=
@@ -25,6 +27,8 @@ github.com/go-logr/logr v1.4.3 h1:CjnDlHq8ikf6E492q6eKboGOC0T8CDaOvkHCIg8idEI=
 github.com/go-logr/logr v1.4.3/go.mod h1:9T104GzyrTigFIr8wt5mBrctHMim0Nb2HLGrmQ40KvY=
 github.com/go-logr/stdr v1.2.2 h1:hSWxHoqTgW2S2qGc0LTAI563KZ5YKYRhT3MFKZMbjag=
 github.com/go-logr/stdr v1.2.2/go.mod h1:mMo/vtBO5dYbehREoey6XUKy/eSumjCCveDpRre4VKE=
+github.com/go-test/deep v1.1.0 h1:WOcxcdHcvdgThNXjw0t76K42FXTU7HpNQWHpA2HHNlg=
+github.com/go-test/deep v1.1.0/go.mod h1:5C2ZWiW0ErCdrYzpqxLbTX7MG14M9iiw8DgHncVwcsE=
 github.com/goccy/go-json v0.10.6 h1:p8HrPJzOakx/mn/bQtjgNjdTcN+/S6FcG2CTtQOrHVU=
 github.com/goccy/go-json v0.10.6/go.mod h1:oq7eo15ShAhp70Anwd5lgX2pLfOS3QCiwU/PULtXL6M=
 github.com/gogf/gf/contrib/drivers/sqlite/v2 v2.10.2 h1:KLS68SWS2W749x7e+eCCOO3UD2Sbw+bIbLEPR8o1FXw=
diff --git a/shortdrama/service/drama_service.go b/shortdrama/service/drama_service.go
index d693e40..4a59b55 100644
--- a/shortdrama/service/drama_service.go
+++ b/shortdrama/service/drama_service.go
@@ -9,7 +9,6 @@ import (
 	"math"
 	"net/http"
 	"os"
-	"os/exec"
 	"path/filepath"
 	"strings"
 	"time"
@@ -21,6 +20,7 @@ import (
 	"video-factory/shortdrama/model/dto"
 	"video-factory/shortdrama/model/entity"
 
+	"github.com/Eyevinn/mp4ff/mp4"
 	"github.com/gogf/gf/v2/database/gdb"
 	"github.com/gogf/gf/v2/frame/g"
 	"github.com/gogf/gf/v2/os/gcache"
@@ -1401,33 +1401,236 @@ func (s *dramaService) waitForSegmentVideo(ctx context.Context, d *entity.Drama,
 	return fmt.Errorf("串行等待超时: 任务 %d 第%d段视频在 %d 次轮询后仍未就绪", task.Id, task.SegmentIdx+1, maxAttempts)
 }
 
-func (s *dramaService) extractLastFrame(videoPath, outputPath string) error {
-	cmd := exec.Command("ffmpeg", "-y", "-sseof", "-1", "-i", videoPath,
-		"-frames:v", "1", "-q:v", "2", outputPath)
-	return cmd.Run()
+func (s *dramaService) concatVideos(inputs []string, output string) error {
+	if len(inputs) < 2 {
+		return fmt.Errorf("需要至少2个输入文件才能合并")
+	}
+
+	// 1. 读取所有 MP4 文件到内存
+	var files []*mp4.File
+	for _, path := range inputs {
+		f, err := mp4.ReadMP4File(path)
+		if err != nil {
+			return fmt.Errorf("读取 %s 失败: %w", path, err)
+		}
+		files = append(files, f)
+	}
+
+	// 2. 计算累积 mdat 负载大小和各文件原始基址
+	cumPayload := make([]uint64, len(files))
+	totalPayload := uint64(0)
+	origBase := make([]uint64, len(files))
+	for i, f := range files {
+		cumPayload[i] = totalPayload
+		totalPayload += uint64(len(f.Mdat.Data))
+		origBase[i] = f.Ftyp.Size() + f.Moov.Size() + f.Mdat.HeaderSize()
+	}
+
+	// 3. 以第一个文件为蓝本构建输出结构
+	outFile := mp4.NewFile()
+	outFile.Ftyp = files[0].Ftyp
+	outFile.Moov = files[0].Moov
+
+	outFile.Mdat = &mp4.MdatBox{}
+	outFile.Mdat.SetData(make([]byte, totalPayload))
+	pos := 0
+	for _, f := range files {
+		copy(outFile.Mdat.Data[pos:], f.Mdat.Data)
+		pos += len(f.Mdat.Data)
+	}
+
+	// 4. 按轨道类型初始化合并状态
+	type mergeState struct {
+		stbl      *mp4.StblBox
+		chunkBase uint32 // 当前已合并的 chunk 数量（stsc 偏移用）
+	}
+
+	states := make(map[string]*mergeState)
+	for _, trak := range files[0].Moov.Traks {
+		hdlr := trak.Mdia.Hdlr.HandlerType
+		if _, ok := states[hdlr]; ok {
+			continue
+		}
+		stbl := trak.Mdia.Minf.Stbl
+		nChunks := chunkCount(stbl)
+		states[hdlr] = &mergeState{stbl: stbl, chunkBase: nChunks}
+	}
+
+	// 5. 合并后续文件
+	for fi := 1; fi < len(files); fi++ {
+		f := files[fi]
+		for _, srcTrak := range f.Moov.Traks {
+			hdlr := srcTrak.Mdia.Hdlr.HandlerType
+			st, ok := states[hdlr]
+			if !ok {
+				continue
+			}
+			srcStbl := srcTrak.Mdia.Minf.Stbl
+			nChunks := chunkCount(srcStbl)
+
+			// stco/co64：偏移调整 = 第一个文件的基址差 + 累积 mdat 偏移
+			adjust := int64(origBase[0]-origBase[fi]) + int64(cumPayload[fi])
+			mergeChunkOffsets(st.stbl, srcStbl, adjust)
+
+			// stsz：合并采样大小表
+			if err := mergeSampleSizes(st.stbl.Stsz, srcStbl.Stsz); err != nil {
+				return err
+			}
+
+			// stts：合并时域采样表
+			mergeTimeToSample(st.stbl.Stts, srcStbl.Stts)
+
+			// stsc：合并样块映射表，FirstChunk 需平移
+			if srcStbl.Stsc != nil {
+				for i, e := range srcStbl.Stsc.Entries {
+					sid := srcStbl.Stsc.GetSampleDescriptionID(i + 1)
+					if err := st.stbl.Stsc.AddEntry(e.FirstChunk+st.chunkBase, e.SamplesPerChunk, sid); err != nil {
+						return fmt.Errorf("stsc 合并失败: %w", err)
+					}
+				}
+			}
+
+			st.chunkBase += nChunks
+		}
+	}
+
+	// 6. 修正所有 chunk 偏移（合并后 moov 大小变化）
+	combinedBase := outFile.Ftyp.Size() + outFile.Moov.Size() + outFile.Mdat.HeaderSize()
+	finalAdjust := int64(combinedBase - origBase[0])
+	for _, st := range states {
+		adjustChunkOffsets(st.stbl, finalAdjust)
+	}
+
+	// 7. 更新 mvhd 总时长
+	totalDur := uint64(0)
+	for _, f := range files {
+		totalDur += f.Moov.Mvhd.Duration
+	}
+	outFile.Moov.Mvhd.Duration = totalDur
+
+	// 更新每个 tkhd 时长
+	for _, trak := range outFile.Moov.Traks {
+		hdlr := trak.Mdia.Hdlr.HandlerType
+		totalDur = 0
+		for _, f := range files {
+			for _, t := range f.Moov.Traks {
+				if t.Mdia.Hdlr.HandlerType == hdlr {
+					totalDur += t.Tkhd.Duration
+					break
+				}
+			}
+		}
+		trak.Tkhd.Duration = totalDur
+	}
+
+	// 8. 重建顶级 Children（确保 mdat 指向合并后的数据）
+	outFile.Children = []mp4.Box{outFile.Ftyp, outFile.Moov, outFile.Mdat}
+
+	// 9. 写入输出文件
+	return mp4.WriteToFile(outFile, output)
 }
 
-func (s *dramaService) concatVideos(inputs []string, output string) error {
-	listPath := output + ".list"
-	f, err := os.Create(listPath)
-	if err != nil {
-		return err
+// chunkCount 返回 stbl 中的 chunk 数量
+func chunkCount(stbl *mp4.StblBox) uint32 {
+	if stbl.Stco != nil {
+		return uint32(len(stbl.Stco.ChunkOffset))
 	}
-	for _, p := range inputs {
-		_, _ = f.WriteString(fmt.Sprintf("file '%s'\n", p))
+	if stbl.Co64 != nil {
+		return uint32(len(stbl.Co64.ChunkOffset))
 	}
-	_ = f.Close()
-	defer os.Remove(listPath)
+	return 0
+}
 
-	cmd := exec.Command("ffmpeg", "-y", "-f", "concat", "-safe", "0",
-		"-i", listPath, "-c", "copy", output)
-	out, err := cmd.CombinedOutput()
-	if err != nil {
-		return fmt.Errorf("ffmpeg 合并失败 (%v): %s", err, string(out))
+// mergeChunkOffsets 将 src 的 stco/co64 追加到 base，每个偏移加上 adjust
+func mergeChunkOffsets(base, src *mp4.StblBox, adjust int64) {
+	if src.Stco != nil {
+		if base.Stco == nil {
+			base.Stco = &mp4.StcoBox{}
+		}
+		for _, off := range src.Stco.ChunkOffset {
+			base.Stco.ChunkOffset = append(base.Stco.ChunkOffset, uint32(int64(off)+adjust))
+		}
+	}
+	if src.Co64 != nil {
+		if base.Co64 == nil {
+			base.Co64 = &mp4.Co64Box{}
+		}
+		for _, off := range src.Co64.ChunkOffset {
+			base.Co64.ChunkOffset = append(base.Co64.ChunkOffset, uint64(int64(off)+adjust))
+		}
+	}
+}
+
+// adjustChunkOffsets 对 stbl 中所有 stco/co64 偏移统一加上 adjust
+func adjustChunkOffsets(stbl *mp4.StblBox, adjust int64) {
+	if stbl.Stco != nil {
+		for i := range stbl.Stco.ChunkOffset {
+			stbl.Stco.ChunkOffset[i] = uint32(int64(stbl.Stco.ChunkOffset[i]) + adjust)
+		}
+	}
+	if stbl.Co64 != nil {
+		for i := range stbl.Co64.ChunkOffset {
+			stbl.Co64.ChunkOffset[i] = uint64(int64(stbl.Co64.ChunkOffset[i]) + adjust)
+		}
+	}
+}
+
+// mergeSampleSizes 合并 stsz 采样大小表
+func mergeSampleSizes(base, src *mp4.StszBox) error {
+	if src == nil {
+		return nil
+	}
+	if base.SampleUniformSize > 0 {
+		if src.SampleUniformSize == base.SampleUniformSize {
+			base.SampleNumber += src.SampleNumber
+		} else {
+			sizes := make([]uint32, base.SampleNumber)
+			for i := range sizes {
+				sizes[i] = base.SampleUniformSize
+			}
+			if src.SampleUniformSize > 0 {
+				for i := uint32(0); i < src.SampleNumber; i++ {
+					sizes = append(sizes, src.SampleUniformSize)
+				}
+			} else {
+				sizes = append(sizes, src.SampleSize...)
+			}
+			base.SampleUniformSize = 0
+			base.SampleSize = sizes
+			base.SampleNumber = uint32(len(sizes))
+		}
+	} else {
+		if src.SampleUniformSize > 0 {
+			for i := uint32(0); i < src.SampleNumber; i++ {
+				base.SampleSize = append(base.SampleSize, src.SampleUniformSize)
+			}
+		} else {
+			base.SampleSize = append(base.SampleSize, src.SampleSize...)
+		}
+		base.SampleNumber = uint32(len(base.SampleSize))
 	}
 	return nil
 }
 
+// mergeTimeToSample 合并 stts 时域采样表（相邻相同 delta 的条目合并）
+func mergeTimeToSample(base, src *mp4.SttsBox) {
+	if src == nil || len(src.SampleCount) == 0 {
+		return
+	}
+	if len(base.SampleCount) > 0 {
+		lastDelta := base.SampleTimeDelta[len(base.SampleTimeDelta)-1]
+		firstDelta := src.SampleTimeDelta[0]
+		if lastDelta == firstDelta {
+			base.SampleCount[len(base.SampleCount)-1] += src.SampleCount[0]
+			base.SampleCount = append(base.SampleCount, src.SampleCount[1:]...)
+			base.SampleTimeDelta = append(base.SampleTimeDelta, src.SampleTimeDelta[1:]...)
+			return
+		}
+	}
+	base.SampleCount = append(base.SampleCount, src.SampleCount...)
+	base.SampleTimeDelta = append(base.SampleTimeDelta, src.SampleTimeDelta...)
+}
+
 func (s *dramaService) downloadFile(ctx context.Context, url, dest string) error {
 	resp, err := http.Get(url)
 	if err != nil {
@@ -1466,14 +1669,12 @@ func (s *dramaService) downloadFile(ctx context.Context, url, dest string) error
 }
 
 func (s *dramaService) probeVideo(path string) error {
-	cmd := exec.Command("ffprobe", "-v", "error", "-show_entries",
-		"format=duration", "-of", "default=noprint_wrappers=1:nokey=1", path)
-	out, err := cmd.CombinedOutput()
+	f, err := mp4.ReadMP4File(path)
 	if err != nil {
-		return fmt.Errorf("ffprobe 检查失败 (%v): %s", err, string(out))
+		return fmt.Errorf("视频文件无效: %w", err)
 	}
-	if len(out) == 0 {
-		return fmt.Errorf("无法获取视频时长")
+	if f.Moov == nil {
+		return fmt.Errorf("视频文件缺少 moov box")
 	}
 	return nil
 }