我在使用PaddleOCR进行arm64打包的时候发现 依赖包 lmdb是否不支持aarch64(arm64)

[WilliamGrant]

开始执行我的docker 虚拟化部署镜像的构建，因为我要在arm64的板子上做多应用场景，结果发现lmdb每次都是这个库报错，我感觉这个库可能不支持arm架构。所以，诸位帮忙瞅瞅有什么办法能够解决，或者说不用这个库的版本。
安装以来如下

opencv-python-headless==4.10.0.84
https://download.pytorch.org/whl/cpu/torch-2.4.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl#sha256=7c159e89d4ecf08403f9d1373d554422240b9b1146a0a19129069dc357a72b2b
https://download.pytorch.org/whl/cpu/torchvision-0.19.0-cp310-cp310-linux_aarch64.whl#sha256=d467d434005fd05a227a2ba7af4c591bb67e6d4a97bbd06eda8da83f43e9fd07
paddleocr==2.9.1
paddlepaddle==3.0.0b2
ultralytics==8.3.28

命令行提示如下：

#9 [5/6] RUN pip install --no-cache-dir -r requirements.txt
#9 14.16 Collecting torch==2.4.0
#9 15.33   Downloading https://download.pytorch.org/whl/cpu/torch-2.4.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (89.8 MB)
#9 27.05      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 89.8/89.8 MB 8.2 MB/s eta 0:00:00
#9 29.48 Collecting torchvision==0.19.0
#9 30.18   Downloading https://download.pytorch.org/whl/cpu/torchvision-0.19.0-cp310-cp310-linux_aarch64.whl (14.1 MB)
#9 32.06      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 14.1/14.1 MB 8.5 MB/s eta 0:00:00
#9 35.02 Collecting opencv-python-headless==4.10.0.84
#9 36.34   Downloading opencv_python_headless-4.10.0.84-cp37-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (29.3 MB)
#9 40.96      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 29.3/29.3 MB 8.5 MB/s eta 0:00:00
#9 41.74 Collecting paddleocr==2.9.1
#9 42.04   Downloading paddleocr-2.9.1-py3-none-any.whl (544 kB)
#9 42.12      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 544.7/544.7 kB 9.7 MB/s eta 0:00:00
#9 42.88 Collecting paddlepaddle==3.0.0b2
#9 43.17   Downloading paddlepaddle-3.0.0b2-cp310-cp310-manylinux2014_aarch64.whl (86.9 MB)
#9 65.05      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 86.9/86.9 MB 4.1 MB/s eta 0:00:00
#9 69.03 Collecting ultralytics==8.3.28
#9 69.33   Downloading ultralytics-8.3.28-py3-none-any.whl (881 kB)
#9 69.54      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 881.2/881.2 kB 4.7 MB/s eta 0:00:00
#9 73.56 Collecting numpy>=1.21.2
#9 73.86   Downloading numpy-2.1.3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (13.9 MB)
#9 77.02      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 13.9/13.9 MB 4.4 MB/s eta 0:00:00
#9 78.39 Collecting albumentations==1.4.10
#9 78.69   Downloading albumentations-1.4.10-py3-none-any.whl (161 kB)
#9 78.72      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 161.9/161.9 kB 10.2 MB/s eta 0:00:00
#9 80.36 Collecting opencv-python
#9 80.66   Downloading opencv_python-4.10.0.84-cp37-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (41.7 MB)
#9 88.16      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 41.7/41.7 MB 7.5 MB/s eta 0:00:00
#9 89.18 Collecting fire>=0.3.0
#9 89.47   Downloading fire-0.7.0.tar.gz (87 kB)
#9 89.50      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 87.2/87.2 kB 10.3 MB/s eta 0:00:00
#9 89.64   Preparing metadata (setup.py): started
#9 94.39   Preparing metadata (setup.py): finished with status 'done'
#9 95.38 Collecting tqdm
#9 95.68   Downloading tqdm-4.67.0-py3-none-any.whl (78 kB)
#9 95.70      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 78.6/78.6 kB 10.6 MB/s eta 0:00:00
#9 98.07 Collecting fonttools>=4.24.0
#9 98.37   Downloading fonttools-4.55.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (4.6 MB)
#9 99.47      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 4.6/4.6 MB 4.2 MB/s eta 0:00:00
#9 100.3 Collecting lmdb
#9 100.6   Downloading lmdb-1.5.1.tar.gz (881 kB)
#9 100.8      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 881.2/881.2 kB 4.3 MB/s eta 0:00:00
#9 101.1   Preparing metadata (setup.py): started
#9 103.0   Preparing metadata (setup.py): finished with status 'error'
#9 103.1   error: subprocess-exited-with-error
#9 103.1   
#9 103.1   × python setup.py egg_info did not run successfully.
#9 103.1   │ exit code: 1
#9 103.1   ╰─> [8 lines of output]
#9 103.1       sh: 1: patch: not found
#9 103.1       Traceback (most recent call last):
#9 103.1         File "<string>", line 2, in <module>
#9 103.1         File "<pip-setuptools-caller>", line 34, in <module>
#9 103.1         File "/tmp/pip-install-do0kq16z/lmdb_cbda668ae95a4f00929997395edf2c65/setup.py", line 119, in <module>
#9 103.1           raise Exception('Applying patch failed')
#9 103.1       Exception: Applying patch failed
#9 103.1       py-lmdb: Using bundled liblmdb with py-lmdb patches; override with LMDB_FORCE_SYSTEM=1 or LMDB_PURE=1.
#9 103.1       [end of output]
#9 103.1   
#9 103.1   note: This error originates from a subprocess, and is likely not a problem with pip.
#9 103.1 error: metadata-generation-failed
#9 103.1 
#9 103.1 × Encountered error while generating package metadata.
#9 103.1 ╰─> See above for output.
#9 103.1 
#9 103.1 note: This is an issue with the package mentioned above, not pip.
#9 103.1 hint: See above for details.
#9 105.0 
#9 105.0 [notice] A new release of pip is available: 23.0.1 -> 24.3.1
#9 105.0 [notice] To update, run: pip install --upgrade pip
#9 ERROR: process "/dev/.buildkit_qemu_emulator /bin/sh -c pip install --no-cache-dir -r requirements.txt" did not complete successfully: exit code: 1
------
 > [5/6] RUN pip install --no-cache-dir -r requirements.txt:
103.1 error: metadata-generation-failed
103.1 
103.1 × Encountered error while generating package metadata.
103.1 ╰─> See above for output.

[GreatV]

lmdb一般情况下不需要，试试打包的时候移除。

[WilliamGrant]

你指的打包的时候移除，是指在pip install paddleocr[no-lmdb]吗？还是将paddleOCR项目在本地打包的时候移除项目（paddleOCR）中的对lmdb的引用，然后再用于我的业务项目中？我还是个py新人，有些机制还不是很懂，还望多多指点。

[GreatV]

本地打包的时候移除