對于筆者這個openharmony高校開發(fā)者來說，開源的樂趣在于折騰。在此記錄下折騰OpenHarmony標(biāo)準(zhǔn)系統(tǒng)開機動畫的過程，ohos的開機動畫從3.1beta之后經(jīng)歷過一次變更。下面展示一下變更前和變更后的開機動畫。

OpenHarmony3.1beta版本使用的舊開機動畫

• 這個開機動畫和HarmonyOS的有異曲同工之妙。

從OpenHarmony3.1release版本開始使用的新開機動畫

帶有開機聲效的視頻如文章開頭。

新版Logo設(shè)計者是劉石老師，劉石老師在《新發(fā)布的 OpenHarmony Logo 竟有這么多的故事！》一文中分享了新版logo背后的故事

兩個版本開機動畫實現(xiàn)方式不一樣

3.2beta3開機動畫模塊源碼分析

• graphic圖形子系統(tǒng)包含了開機動畫模塊，開機動畫模塊在ohos3.2beta3源碼下foundation/graphic/graphic_2d/frameworks/bootanimation，開機動畫模塊bootanimation源碼結(jié)構(gòu)

├── BUILD.gn
├── data
│   ├── bootanimation_tool
│   │   ├── README.md
│   │   ├── raw_maker.py
│   │   └── raw_player.py
│   ├── bootpic.zip   # 包括了開機動畫的所有圖片幀和json播放配置文件
│   ├── bootsound.wav # 開機聲效
│   └── generate_raw.sh
├── include           # 開機動畫模塊的頭文件
│   ├── boot_animation.h
│   ├── log.h
│   └── util.h
└── src               # 開機動畫的源文件
    ├── boot_animation.cpp
    ├── main.cpp
    └── util.cpp

• bootpic.zip打開后內(nèi)容如下
  

  • OH_bootAni compressed內(nèi)容如下，其中包括了150張開機動畫的圖片幀。
    
  • config.json內(nèi)容如下，在這個文件中設(shè)置開機動畫的播放幀率，范圍為30~60幀，此處視頻幀率被設(shè)置為30幀。OH_bootAni compressed文件夾內(nèi)有150張圖片，所以開機動畫的播放時間為5秒。

{
	"Remark": "FrameRate Support 30, 60 frame rate configuration",
	"FrameRate": 30
}

開機動畫源文件bootpic.zip和bootsound.wav等作為配置文件打包至開發(fā)板/system/etc/init目錄下

閱讀bootanimation目錄下BUILD.gn可以知道bootpic.zip和bootsound.wav等作為配置文件打包至開放板/system/etc/init目錄下。

## Install data/*.jpg to /system/etc/init/ {{{
ohos_prebuilt_etc("bootanimation_pics") {
  source = "data/bootpic.zip"     ## bootpic.zip在data目錄下
  relative_install_dir = "init"    
  part_name = "graphic_standard"  ## 部件名
  subsystem_name = "graphic"      ## 子系統(tǒng)名
}

ohos_prebuilt_etc("bootanimation_sounds") {
  source = "data/bootsound.wav"   ## bootsound.wav在data目錄下
  relative_install_dir = "init"
  part_name = "graphic_standard"
  subsystem_name = "graphic"
}
## Install data/*.jpg to /system/etc/init/ }}}

bootpic.zip的解壓縮依賴三方庫zlib

閱讀bootanimation目錄下BUILD.gn可以知道bootpic.zip的解壓縮依賴三方庫zlib

config.json文件在util.h中被聲明為C ++ 的std::string類型

• 閱讀bootanimation目錄下BUILD.gn可以知道config.json的解析依賴三方庫cJSON
  

開機動畫服務(wù)啟動配置graphic.cfg分別啟動了bootanimation進程

開機動畫服務(wù)啟動配置graphic.cfg在ohosbeta3源碼./foundation/graphic/graphic_2d/graphic.cfg目錄，分別啟動了bootanimation和render_service進程。

.cfg只是一個為開發(fā)及使用方便而"發(fā)明"的一個后綴名。所以，這種文件沒有固定的格式，其實也并不能算作是一種文件類型。

{  
    "jobs" : [{
            "name" : "init",
            "cmds" : [
                "chmod 666 /dev/mali0",
                "chown system graphics /dev/mali0"
            ]
        }, {
            "name": "services:restartrender_service",
            "cmds": [
                "reset foundation",
                "reset bootanimation",
                "reset gralloc_host",
                "reset hwc_host"
            ]
        }
    ],
    "services" : [{
            "name" : "render_service",  # 渲染服務(wù)端
            "path" : ["/system/bin/render_service"],
            "critical" : [1, 5, 60],
            "importance" : -20,
            "uid" : "system",
            "gid" : ["system", "shell", "uhid", "root"],
            "caps" : ["SYS_NICE"],
            "secon" : "u:r:render_service:s0",
            "jobs" : {
                "on-restart" : "services:restartrender_service"
            },
            "once" : 0
        }, {
            "name" : "bootanimation",  # 開機啟動進程
            "path" : ["/system/bin/bootanimation"],
            "bootevents": "bootevent.bootanimation.started",
            "importance" : -20,
            "once" : 1,
            "uid" : "graphics",
            "gid" : ["graphics", "system", "shell", "uhid", "root"],
            "secon" : "u:r:bootanimation:s0"
        }
    ]
}

• 可以用hdc_std工具將pc端與開發(fā)板連接進入shell界面運行ps -ef命令，查看進程信息。

# ps -ef
UID            PID  PPID C STIME TTY          TIME CMD
····
system         565     1 19 09:12:27 ?    00:00:01 bootanimation
system         566     1 6 09:12:27 ?     00:00:00 render_service
····

• 查看beta3源碼foundation/graphic/graphic_2d/BUILD.gn,可以知道graphic.cfg會被打包到開發(fā)板/system/etc/init/目錄下
  
• 用hdc_std工具連接開發(fā)板后，服務(wù)啟動配置graphic.cfg在開發(fā)板systemetcinit目錄也能找到

find . -name graphic.cfg

boot_animation.cpp中將開發(fā)板目錄/system/etc/init下的bootpic.zip和bootsound.wav聲明為C ++ 的std::string類型

using namespace OHOS;
static const std::string BOOT_PIC_ZIP = "/system/etc/init/bootpic.zip";
static const std::string BOOT_SOUND_URI = "file://system/etc/init/bootsound.wav";

• C ++中的std :: string類

  • C ++在其定義中具有一種將字符序列表示為class對象的方式。此類稱為std ::字符串。
  • 它是一個容器類

• static const既是只讀的,又是只在當(dāng)前模塊中可見的。

• const 就是只讀的意思,只在聲明中使用;
• static 一般有2個作用，規(guī)定作用域和存儲方式。
  • 對于局部變量, static規(guī)定其為靜態(tài)存儲方式, 每次調(diào)用的初始值為上一次調(diào)用的值,調(diào)用結(jié)束后存儲空間不釋放
  • 對于全局變量, 如果以文件劃分作用域的話,此變量只在當(dāng)前文件可見; 對于static函數(shù)也是在當(dāng)前模塊內(nèi)函數(shù)可見。

• using namespace OHOS;表示使用using指令使用OHOS空間，其在boot_animation.h中定義boot_animation.CPP中要用到boot_animation.h中OHOS空間中的函數(shù)和變量。

namespace OHOS {
class BootAnimation {
public:
    void Init(int32_t width, int32_t height, const std::shared_ptr& handler,
        std::shared_ptr& runner);
    void Draw();
    void CheckExitAnimation();
    void PlaySound();
    bool CheckFrameRateValid(int32_t ratevalue);
    ~BootAnimation();
private:
    void OnVsync();
    void OnDraw(SkCanvas* canvas, int32_t curNo);
    void InitBootWindow();
    void InitRsSurface();
    void InitPicCoordinates();
    int32_t windowWidth_;
    int32_t windowHeight_;
    sptr window_;
    sptr scene_;
    std::unique_ptr framePtr_;
    std::shared_ptr rsSurface_;
    OHOS::Rosen::RenderContext* rc_;
    int32_t freq_ = 30;
    int32_t realHeight_ = 0;
    int32_t realWidth_ = 0;
    int32_t pointX_ = 0;
    int32_t pointY_ = 0;
    int32_t picCurNo_ = -1;
    int32_t imgVecSize_ = 0;
    std::shared_ptr receiver_ = nullptr;
    std::shared_ptr soundPlayer_ = nullptr;
    ImageStructVec imageVector_;
    std::shared_ptr mainHandler_ = nullptr;
    std::shared_ptr runner_ = nullptr;
    bool setBootEvent_ = false;
};
} // namespace OHOS

3.1Beta開機動畫模塊源碼分析

• 開機動畫模塊在社區(qū)gitee倉庫OpenHarmony / graphic_graphic_2d
  • 找到標(biāo)簽欄，找到3.1Beta分支。下載下來分析。
    
• bootanimation開機模塊目錄結(jié)構(gòu)如下：

├─data
│    ├── bootanimation-480x960.raw
│    └── generate_raw.sh
├─include
│       ├── raw_parser.h
│       └── util.h
└─src
│    ├── raw_parser.cpp
│    ├── main.cpp
│    └── util.cpp
└─BUILD.gn

開機動畫文件bootanimation.raw作為配置文件打包至開發(fā)板/system/etc目錄下

## Install data/bootanimation-480x960.raw to /system/etc/bootanimation-480x960.raw {{{
ohos_prebuilt_etc("bootanimation-480x960.raw") {
  source = "data/bootanimation-480x960.raw"
  part_name = "graphic_standard"
  subsystem_name = "graphic"
}
## Install data/bootanimation-480x960.raw to /system/etc/bootanimation-480x960.raw }}}

RAW是未經(jīng)處理、也未經(jīng)壓縮的格式，可以把RAW概念化為“原始圖像編碼數(shù)據(jù)”或更形象的稱為“數(shù)字底片”。

3.2beta3開機動畫制作

• 1.開機動畫在3.2beta版本的源文件為bootpic.zip和bootpic.wav,需要利用視頻剪輯軟件制作好一段自己的開機動畫。
• 2.然后利用特殊軟件工具將制作好的視頻變成150張圖片，然后將圖片編號再排好序。
• 3.最后打包為bootpic.zip壓縮包。利用hdc_std工具將自制的bootpic.zip導(dǎo)入開發(fā)板/system/etc/init目錄下。
• 4.重啟開發(fā)板就可以了。

筆者使用的剪映軟件（對筆者來說剪映夠用）制作開機動畫。然后在
https://www.img2go.com/zh/convert-to-image網(wǎng)址將視頻轉(zhuǎn)換成圖片。

最后打包成bootpic.zip（需要和原生的bootpic.zip目錄機構(gòu)一樣），hdc_std工具使用命令如下：

C:UsersjjhDesktop3.2beta1SDKohos-sdkwindowstoolchains-windows-3.2.2.5-Beta1toolchains>hdc_std shell
# cd system/etc/init # 進入system/etc/init目錄
# mount -o remount,rw / # 將系統(tǒng)變成可讀寫
# ls
access_token.cfg                     locationsa.cfg
accessibility.cfg                    media_service.cfg
accountmgr.cfg                       memmgrservice.cfg
appspawn.cfg                         misc.cfg
audio_policy.cfg                     mmi_uinput.rc
batterystats.cfg                     msdp_musl.cfg
bgtaskmgr_service.cfg                multimodalinput.cfg
bluetooth_service.cfg                netmanager_base.cfg
bootpic.zip                          netsysnative.cfg
bootsound.wav                        nwebspawn.cfg
bytrace.cfg                          param_watcher.cfg
camera_service.cfg                   pasteboardservice.cfg
config.txt                           pinauth_sa_profile.cfg
console.cfg                          pulseaudio.cfg
dcamera.cfg                          resource_schedule_service.cfg
device_usage_statistics_service.cfg  samgr_standard_musl.cfg
deviceauth_service.cfg               screenlockservice.cfg
deviceinfoservice.cfg                sensors_musl.cfg
dhardware.cfg                        softbus_server_musl.cfg
distributed_data.cfg                 storage_daemon.cfg
distributedbms.cfg                   storage_manager.cfg
distributedfile.cfg                  telephony.cfg
distributedsched_musl.cfg            thermal.cfg
downloadservice.cfg                  thermal_protector.cfg
dscreen.cfg                          timeservice.cfg
dslm_service.cfg                     token_sync.cfg
edm.cfg                              udevd.rc
faceauth_sa_profile.cfg              ueventd.cfg
faultloggerd.cfg                     ui_service.cfg
fms_service.cfg                      updater_sa.cfg
foundation.cfg                       usb_service.cfg
graphic.cfg                          useriam.cfg
hidumper_service.cfg                 wallpaperservice.cfg
hilogd.cfg                           watchdog.cfg
hiview.cfg                           weston.cfg
huks_service.cfg                     wifi_hal_service.cfg
init.reboot.cfg                      wifi_standard.cfg
inputmethodservice.cfg               work_scheduler_service.cfg
installs.cfg
# rm bootpic.zip  # 刪除原有的開機動畫
# exit            # 退出開發(fā)板系統(tǒng)

C:UsersjjhDesktop3.2beta1SDKohos-sdkwindowstoolchains-windows-3.2.2.5-Beta1toolchains>hdc_std file send bootpic.zip /system/etc/init/
FileTransfer finish, File count = 1, Size:2776406 time:389ms rate:7137.29kB/s # 將自制的開機動畫發(fā)送至/system/etc/init目錄下

3.1beta版本開機動畫制作

準(zhǔn)備好一段開機動畫的mp4格式動畫，利用以下raw_maker.py腳本,內(nèi)容如下：

# !/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
* Copyright (c) 2022 Shenzhen Kaihong Digital Industry Development Co., Ltd.
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
"""

import struct
import zlib
import os
import argparse
import re
import pip


def lost_module(module_name):
    print("""
need %s module, try install first:

    pip install %s""" % (module_name, module_name))
    exit()


try:
    import cv2
except ImportError:
    pip.main(["install", "opencv-python", "-i", "https://pypi.tuna.tsinghua.edu.cn/simple"])
    try:
        import cv2
    except ImportError:
        cv2 = None
        lost_module("opencv-python")

try:
    from PIL import Image
except ImportError:
    pip.main(["install", "pillow"])
    try:
        from PIL import Image
    except ImportError:
        Image = None
        lost_module("pillow")

try:
    import numpy as np
except ImportError:
    pip.main(["install", "numpy"])
    try:
        import numpy as np
    except ImportError:
        np = None
        lost_module("numpy")


class RawMaker:
    """
    Make a boot video by a MP4 file or some .img files:
    """

    def __init__(self, args):
        self._mp4 = args.mp4
        self._image = args.image
        self._out = args.out
        self._display = [int(i) for i in re.split(r'[xX* ]+', args.display.strip())]
        self._rotate = args.rotate
        self._flip = args.flip
        self._fnp = 0
        self._vdo = None
        self._image_files = []

    def _iter_img(self):
        if self._mp4:
            success, frame = self._vdo.read()
            if success:
                image = Image.fromarray(frame)
                return success, image
            else:
                return False, None
        else:
            if self._fnp >= len(self._image_files):
                return False, None
            image = Image.open(os.path.join(self._image, self._image_files[self._fnp]))
            self._fnp += 1
            return True, image

    def make(self):
        frame_count, width, height = 0, 0, 0
        if self._mp4:
            if not os.path.exists(self._mp4):
                print("mp4 file %s is not exist" % self._mp4)
                exit()
            self._vdo = cv2.VideoCapture(self._mp4)
            fps = int(self._vdo.get(cv2.CAP_PROP_FPS))
            w = int(self._vdo.get(cv2.CAP_PROP_FRAME_WIDTH))
            h = int(self._vdo.get(cv2.CAP_PROP_FRAME_HEIGHT))
            frame_count = int(self._vdo.get(cv2.CAP_PROP_FRAME_COUNT))
            if fps != 30:
                print("video fps :", fps, ", width :", w, ", height :", h, ", frame count :", frame_count)
            if frame_count <= 0:
                exit()
        elif self._image:
            for fn in os.listdir(self._image):
                self._image_files.append(fn)
            frame_count = len(self._image_files)
            if frame_count <= 0:
                exit()
            self._image_files.sort()
        else:
            exit()

        output_bytes = bytearray(b"RAW.diff")
        offset = 8
        screen_old_bytes = None
        num = 0
        while True:
            ret, img = self._iter_img()
            if not ret:
                break
            num += 1
            img = img.convert("RGBA")
            if self._flip:
                img = img.transpose(Image.FLIP_LEFT_RIGHT)
            if self._rotate == 90:
                img = img.transpose(Image.ROTATE_90)
            elif self._rotate == 180:
                img = img.transpose(Image.ROTATE_180)
            elif self._rotate == 270:
                img = img.transpose(Image.ROTATE_270)
            if self._display[0] != 0:
                img = img.resize((self._display[0], self._display[1]))
            img = np.array(img)
            height, width = img.shape[0], img.shape[1]
            img[img < 20] = 0
            img = img.reshape(-1)
            screen_now_bytes = img.tobytes()
            if screen_old_bytes is None:
                screen_old_bytes = screen_now_bytes
                start_pos = 0
                end_pos = width * height * 4
            else:
                start_pos, end_pos = 3, 6
                for i in range(width * height * 4):
                    if screen_now_bytes[i] != screen_old_bytes[i]:
                        start_pos = i
                        break
                for i in range(width * height * 4 - 1, start_pos, -1):
                    if screen_now_bytes[i] != screen_old_bytes[i]:
                        end_pos = i + 1
                        break
                screen_old_bytes = screen_now_bytes
            print("r|%s%s|" % ("=" * int(num / frame_count * 30), " " * (30 - int(num / frame_count * 30))), 
			        "%.2f%%" % (num / frame_count * 100), end="", flush=True)
            if start_pos == 3 or end_pos == 6:
                output_bytes[offset:offset + 16] = struct.pack("IIII", 0, 0, 0, 0)
                offset += 16
                continue
            compressed_bytes = zlib.compress(screen_old_bytes[start_pos:end_pos])
            raw_len = end_pos - start_pos
            new_len = len(compressed_bytes)
            output_bytes[offset:offset + 16] = struct.pack("IIII", 2, start_pos, raw_len, new_len)
            offset += 16
            output_bytes[offset:offset + new_len] = compressed_bytes
            offset += new_len
            while new_len % 4 != 0:
                new_len += 1
                output_bytes[offset:offset + 1] = b''
                offset += 1

        if not os.path.exists(self._out):
            os.makedirs(self._out)
        with open(os.path.join(self._out, "bootanimation-%dx%d.raw" % (width, height)), "wb") as fp:
            fp.write(output_bytes)
        print("nGenerate successfully!")


def parse_option():
    parser = argparse.ArgumentParser(description="Make a boot video by a MP4 file or some .img files",
                                     usage="python raw_maker.py (-m <*.mp4> | -i ) [-o ] "
                                           "[-d ] [-r ] [-f]n"
                                           "  eg.: python raw_maker.py -i ./source/png -o ./out -d 640x480n"
                                           "       python raw_maker.py -m ./animation.mp4 -o ./out -d 640x480")
    exclusive_group = parser.add_mutually_exclusive_group(required=True)
    exclusive_group.add_argument("-m", "--mp4", metavar="<*.mp4>", help="The input <*.mp4> file")
    exclusive_group.add_argument("-i", "--image", metavar="",
                                 help="The  where image files are stored")
    parser.add_argument("-o", "--out", metavar="", default=".",
                        help="Place generated .raw files into the ")
    parser.add_argument("-d", "--display", metavar="", default="0x0",
                        help="Set the boot video display  and zoom the image, e.g.:640x480")
    parser.add_argument("-r", "--rotate", metavar="", type=int, help="Rotate video , e.g.:90 180 270")
    parser.add_argument("-f", "--flip", action="store_true", help="Flip the video", )
    return parser.parse_args()


if __name__ == "__main__":
    raw_maker = RawMaker(parse_option())
    raw_maker.make()

在當(dāng)前文件夾打開cmd窗口執(zhí)行以下命令

python raw_maker.py -m ./cat.mp4 -d 720x1280

此時會生成cat.raw文件，分辨率為720x1280

播放cat.raw準(zhǔn)備好raw_player.py腳本，內(nèi)容如下：

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
* Copyright (c) 2022 Shenzhen Kaihong Digital Industry Development Co., Ltd.
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
"""

import zlib
import struct
import time
import numpy as np
import cv2
import sys
import re
import argparse


class RawPlayer:
    """
        Play a boot video file
    """

    def __init__(self, args):
        self._raw = args.raw
        cv2.namedWindow("play", cv2.WINDOW_AUTOSIZE)
        pass

    def play(self):
        screen_size = re.findall("bootanimation-([0-9]+)x([0-9]+).raw", self._raw)
        if len(screen_size) != 1:
            exit()
        width, height = int(screen_size[0][0]), int(screen_size[0][1])
        with open(sys.argv[-1], "rb") as fp:
            data = fp.read()
        off = 8
        img = None
        while off < len(data):
            data_type, offset, length, data_length = struct.unpack("IIII", data[off:off + 16])
            off += 16
            if data_type == 0:
                time.sleep(0.03)
                continue

            out = zlib.decompress(data[off:off + data_length])

            if img is None:
                img = np.copy(np.frombuffer(out, dtype=np.uint8))
            else:
                temp_img = np.frombuffer(out, dtype=np.uint8)
                img[offset:offset + length] = temp_img
            reshape_img = img.reshape((height, width, 4))
            cv2.imshow("play", reshape_img)
            if cv2.waitKey(30) & 0xff == 27 or cv2.getWindowProperty("play", cv2.WND_PROP_VISIBLE) < 1:
                break
            while data_length % 4 != 0:
                data_length += 1
            off += data_length


def parse_option():
    parser = argparse.ArgumentParser(description="Play a boot video file",
                                     usage="python raw_player.py [-h] <*.raw>n"
                                           "  eg.: python raw_player.py ./bootanimation-640x480.raw")
    parser.add_argument("raw", metavar="<*.raw>", help="file <*.raw> to play")
    return parser.parse_args()


if __name__ == "__main__":
    raw_player = RawPlayer(parse_option())
    raw_player.play()

在當(dāng)前文件夾打開cmd窗口，執(zhí)行以下命令;

python raw_player.py bootanimation-720x1280.raw

效果如下：