安装调试kotti_ai：AI+互联网企业级部署应用软件包@riscv+OpenKylin

先上结论：riscv+OpenKylin可以安装pyramid+kotti+kotti_ai

但是paddle_serving_client无法安装，项目的AI实现部分需要改用其它方法，比如onnx。最终onnx也没有装成，只好用飞桨自己的推理。

安装kotti

pip install kotti

安装kotti和pyramid相关库

pip install kotti_image kotti_tinymce pyramid_debugtoolbar plone

安装kotti_ai

python 3 setup.py develop

最终bcrypt过不去，放弃image部分。最终发现没有bcrypt kotti都过不去。最后是通过apt安装bcrypt：apt install python3-bcrypt

尝试编译安装rust（失败）

先安装需要的包：

apt install curl
apt install ninja-build

git clone https://github.com/rust-lang/rust

进入rust目录，然后编译,可以先提前设置环境变量：export MAX_JOBS=64

不过我感觉没有，还是16个编译线程，估计算能云就是给了16个？

cd rust
./configure
make -j 64 
make install

估计需要4-5小时。而且自动识别了riscv！

[2172/2183] 304/335 结果到了最后报错，大约是需要llvm，但是安装llvm之后还是报错。

下载二进制文件安装rust

最终是下载编译好的软件包：Release Rust RISC-V Binaries · msizanoen1/rust-binaries · GitHub

下载之后tar -zvf 解包，然后执行install.sh文件即可。

安装llvm

那就开装

git clone --depth 1 https://github.com/llvm/llvm-project llvm
cd llvm/llvm
mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
# cmake --build . -j64
make -j64
make install

最终放弃rust。

尝试编译安装bcrypt

官网文档：GitHub - pyca/bcrypt: Modern(-ish) password hashing for your software and your servers

Installation

To install bcrypt, simply:

$ pip install bcrypt

Note that bcrypt should build very easily on Linux provided you have a C compiler and a Rust compiler (the minimum supported Rust version is 1.56.0).

For Debian and Ubuntu, the following command will ensure that the required dependencies are installed:

$ sudo apt-get install build-essential cargo

下载源码：

git clone https://github.com/pyca/bcrypt

源码安装：python3 setup.py install

结果需要rust，这里卡住过不去了。

rust无法装，所以bcypt无法装，kotti也就无法装.... 一度想放弃在riscv OpenKylin使用kotti和kotti_ai

尝试使用apt安装bcrypt

apt install python3-bcrypt

安装成功

运行飞桨例子

代码参见以102flowers数据集为例训练ResNet50模型_flowers102数据集-CSDN博客

报错，说是numpy那里有问题。

运行mnist的例子，可以过：飞桨AI框架安装和使用示例_飞浆本地安装-CSDN博客

再次重新安装kotti和kotti_ai

安装kotti

pip install kotti过了。

安装 pip install kotti_tinymce

安装sqlalchemy低版本：pip install sqlalchemy==1.4.42

安装：pip install kotti_image

安装pyramid 1.9（不是必须，安装pyramid1.10也行）

安装kotti_ai

pip install -e ".[testing]" 不行，说没安上，去dist 使用pip安装.egg文件，还是报错。

使用python setup.py develop 报错需要pyramid<2.0 ,但其实已经安装pyramid1.9或1.10了。

把这个目录删除/usr/lib/python3.8/site-packages/pyramid-2.0.2-py3.8.egg

python setup.py develop ok

不ok，要用这个方法安装：

 python3 setup.py develop  --user

怀疑是因为使用了root账户，python环境有点乱了。

最终这样就算把kotti_ai安装好了。当然会碰到paddle_serving_client报错，这个以后再说。

结论：

riscv+OpenKylin下成功安好了kotti+kotti_ai

我们就可以进行下一步AI开发工作了！

调试

安装kotti的时候提示：Failed to build bcrypt

Successfully built kotti js.bootstrap js.jquery js.jqueryui pyramid_deform zope.interface docopt filedepot html2text js.angular js.fineuploader js.html5shiv js.jquery_form js.jquery_tablednd js.jquery_timepicker_addon js.jqueryui_tagit pyramid_beaker pyramid_chameleon beaker js.jquery-maskedinput js.jquery-maskmoney js.jquery-sortable js.modernizr js.select2 js.tinymce zope.hookable zope.i18nmessageid
Failed to build bcrypt
ERROR: Could not build wheels for bcrypt, which is required to install pyproject.toml-based projects

按照提示安装bcrypt

安装bcrypt的时候报错：error: can't find Rust compiler

按照提示安装rust

安装rust报错：libbzip2

Building HTSlib requires libbzip2 development files to be installed on the
      build machine; you may need to ensure a package such as libbz2-dev (on Debian
      or Ubuntu Linux) or bzip2-devel (on RPM-based Linux distributions or Cygwin)
      is installed.

      Either configure with --disable-bz2 (which will make some CRAM files
      produced elsewhere unreadable) or resolve this error to build HTSlib.
      config.mk:2: *** Resolve configure error first. Stop.
      # pysam: htslib configure options: None

先不装rust bcrypt了。

编译安装rust报错：Couldn't find required command: ninja (or ninja-build)

You should install ninja as described at
<https://github.com/ninja-build/ninja/wiki/Pre-built-Ninja-packages>,
or set `ninja = false` in the `[llvm]` section of `config.toml`.
Alternatively, set `download-ci-llvm = true` in that `[llvm]` section
to download LLVM rather than building it.

安装ninja-build

apt install ninja-build

编译安装rust编译到100%报错：

64gc-unknown-linux-gnu/stage0-rustc/riscv64gc-unknown-linux-gnu/release/deps/librustc_driver-aba2a9eed125dd79.so" "-shared" "-Wl,-z,relro,-z,now" "-Wl,-O1" "-nodefaultlibs" "-Wl,-z,origin" "-Wl,-rpath,$ORIGIN/../lib"
= note: /usr/bin/ld: cannot find -lWl,--push-state,--as-needed,-latomic,--pop-state: No such file or directory
collect2: error: ld returned 1 exit status

error: could not compile `rustc_driver` (lib) due to 1 previous error
Build completed unsuccessfully in 0:00:20
make: *** [Makefile:19: all] Error 1

这里暂时过不去了。

最终从这里下载编译好的包。Release Rust RISC-V Binaries · msizanoen1/rust-binaries · GitHub

安装bcrypt报错：error: Rust 1.41.0-nightly does not match extension requirement >=1.63.0

二进制版本太低，需要1.63以上，再搞不定就真的不搞了。

好像这里真的搞不定了。

最后用apt install python3-bcrypt解决

运行飞桨例子报错

代码见这里：以102flowers数据集为例训练ResNet50模型_flowers102数据集-CSDN博客

python3 flower102.py
Traceback (most recent call last):
File "flower102.py", line 11, in <module>
    train_dataset = Flowers(mode='train', transform=transform)
File "/usr/local/lib/python3.8/dist-packages/paddle/vision/datasets/flowers.py", line 166, in __init__
    scio = try_import('scipy.io')
File "/usr/local/lib/python3.8/dist-packages/paddle/utils/lazy_import.py", line 32, in try_import
    mod = importlib.import_module(module_name)
File "/usr/lib/python3.8/importlib/__init__.py", line 127, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
File "<frozen importlib._bootstrap>", line 1014, in _gcd_import
File "<frozen importlib._bootstrap>", line 991, in _find_and_load
File "<frozen importlib._bootstrap>", line 975, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
File "<frozen importlib._bootstrap_external>", line 783, in exec_module
File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
File "/usr/lib/python3/dist-packages/scipy/io/__init__.py", line 97, in <module>
    from .matlab import loadmat, savemat, whosmat, byteordercodes
File "/usr/lib/python3/dist-packages/scipy/io/matlab/__init__.py", line 13, in <module>
    from .mio import loadmat, savemat, whosmat
File "/usr/lib/python3/dist-packages/scipy/io/matlab/mio.py", line 11, in <module>
    from .mio4 import MatFile4Reader, MatFile4Writer
File "/usr/lib/python3/dist-packages/scipy/io/matlab/mio4.py", line 11, in <module>
    import scipy.sparse
File "/usr/lib/python3/dist-packages/scipy/sparse/__init__.py", line 229, in <module>
    from .base import *
File "/usr/lib/python3/dist-packages/scipy/sparse/base.py", line 8, in <module>
    from .sputils import (isdense, isscalarlike, isintlike,
File "/usr/lib/python3/dist-packages/scipy/sparse/sputils.py", line 16, in <module>
    supported_dtypes = [np.typeDict[x] for x in supported_dtypes]
File "/usr/lib/python3/dist-packages/scipy/sparse/sputils.py", line 16, in <listcomp>
    supported_dtypes = [np.typeDict[x] for x in supported_dtypes]
File "/usr/local/lib/python3.8/dist-packages/numpy/__init__.py", line 320, in __getattr__
    raise AttributeError("module {!r} has no attribute "
AttributeError: module 'numpy' has no attribute 'typeDict'

先不管这个了，至少mnist的例子可以过。

安装完kotti后报错 module 'sqlalchemy.ext.baked' has no attribute 'bake_lazy_loaders'

>>> import kotti
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/local/lib/python3.8/dist-packages/kotti/__init__.py", line 18, in <module>
from kotti.sqla import Base as KottiBase
File "/usr/local/lib/python3.8/dist-packages/kotti/sqla.py", line 17, in <module>
baked.bake_lazy_loaders()
AttributeError: module 'sqlalchemy.ext.baked' has no attribute 'bake_lazy_loaders'

看sqlalchemy版本：pip show sqlalchemy

安装低版本： pip install sqlalchemy==1.4.42

好了，过了。

安装kotti_ai报错pyramid 2.0.2 is installed but pyramid<2,>=1.9 is required by {'Kotti'}

Installed /usr/lib/python3.8/site-packages/kotti_ai-0.0.2.dev0-py3.8.egg
Processing dependencies for kotti-ai==0.0.2.dev0
error: pyramid 2.0.2 is installed but pyramid<2,>=1.9 is required by {'Kotti'}

pip install pyramid==1.9解决，好吧，还是报这个错，就跟pyramid2已经深入骨子里一样。

在kotti_ai目录执行：