paddleocr，windows pip 安装巨坑 lanms 库

安装 lanms 最佳参考： paddleocr，windows pip 安装巨坑 lanms 库

防丢失，直接抄录的一份

---

paddleocr最后几个库一个比一个难装，特别是 lanms 库，巨难装，拒绝任何花里胡哨，十分钟，三步内解决问题。

pip 下载报错

Keyring is skipped due to an exception: 'keyring.backends'
Collecting lanms
  Using cached lanms-1.0.2.tar.gz (973 kB)
    ERROR: Command errored out with exit status 1:
     command: 'C:\Users\TensorFlow\anaconda3\python.exe' -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'C:\\Users\\TENSOR~1\\AppData\\Local\\Temp\\pip-install-46p1kew5\\lanms\\setup.py'"'"'; __file__='"'"'C:\\Users\\TENSOR~1\\AppData\\Local\\Temp\\pip-install-46p1kew5\\lanms\\setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' egg_info --egg-base 'C:\Users\TENSOR~1\AppData\Local\Temp\pip-install-46p1kew5\lanms\pip-egg-info'
         cwd: C:\Users\TENSOR~1\AppData\Local\Temp\pip-install-46p1kew5\lanms\
    Complete output (11 lines):
    Traceback (most recent call last):
      File "", line 1, in 
      File "C:\Users\TENSOR~1\AppData\Local\Temp\pip-install-46p1kew5\lanms\setup.py", line 27, in 
        if subprocess.call(['make', '--always-make','-C', BASE_DIR]) != 0:
      File "C:\Users\TensorFlow\anaconda3\lib\subprocess.py", line 339, in call
        with Popen(*popenargs, **kwargs) as p:
      File "C:\Users\TensorFlow\anaconda3\lib\subprocess.py", line 800, in __init__
        restore_signals, start_new_session)
      File "C:\Users\TensorFlow\anaconda3\lib\subprocess.py", line 1207, in _execute_child
        startupinfo)
    FileNotFoundError: [WinError 2] 系统找不到指定的文件。
    ----------------------------------------
ERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output.

是以上这样或者是，下载好了一部分后面直接报错（无法复原了）。

第一步 上 github 下载 C 语言版本自行编译

https://github.com/SakuraRiven/LANMS

上面是网址。

 内容是这个样子，下载后安装他说的使用 python setup.py bdist_wheel 进行编译。

注意：报错

(venv3) D:\study\配料表分析\lanms-master>python setup.py bdist_wheel
Traceback (most recent call last):
  File "setup.py", line 27, in 
    if subprocess.call(['make', '--always-make','-C', BASE_DIR]) != 0:
  File "C:\Users\TensorFlow\anaconda3\lib\subprocess.py", line 339, in call
    with Popen(*popenargs, **kwargs) as p:
  File "C:\Users\TensorFlow\anaconda3\lib\subprocess.py", line 800, in __init__
    restore_signals, start_new_session)
  File "C:\Users\TensorFlow\anaconda3\lib\subprocess.py", line 1207, in _execute_child
    startupinfo)
FileNotFoundError: [WinError 2] 系统找不到指定的文件。

if subprocess.call(['make', '--always-make','-C', BASE_DIR]) != 0:
   raise RuntimeError('Cannot compile lanms in the directory: {}'.format(BASE_DIR))

我也不晓得这是缺了个什么文件，只是个判断而已注释这两行代码就好了。

第二步 pip 安装 whl 文件

运行完毕后，在dist文件夹下就能找到 whl 格式的库文件，放到环境下安装即可。

pip install lanms-1.0.2-cp37-cp37m-win_amd64.whl

到这里就算是安装完毕了，如果报错看看是不是第三步的样子。

第三步 修改库源代码

安装后运行可能还会报错，但是已经和之前不一样了，报错：

Traceback (most recent call last):
  File "D:/study/配料表分析/PaddleOCR-release-2.6/test.py", line 1, in 
    from paddleocr import PaddleOCR
  File "D:\study\配料表分析\PaddleOCR-release-2.6\paddleocr.py", line 34, in 
    from tools.infer import predict_system
  File "D:\study\配料表分析\PaddleOCR-release-2.6\tools\infer\predict_system.py", line 33, in 
    import tools.infer.predict_det as predict_det
  File "D:\study\配料表分析\PaddleOCR-release-2.6\tools\infer\predict_det.py", line 31, in 
    from ppocr.data import create_operators, transform
  File "D:\study\配料表分析\PaddleOCR-release-2.6\ppocr\data\__init__.py", line 35, in 
    from ppocr.data.imaug import transform, create_operators
  File "D:\study\配料表分析\PaddleOCR-release-2.6\ppocr\data\imaug\__init__.py", line 48, in 
    from .drrg_targets import DRRGTargets
  File "D:\study\配料表分析\PaddleOCR-release-2.6\ppocr\data\imaug\drrg_targets.py", line 21, in 
    from lanms import merge_quadrangle_n9 as la_nms
  File "C:\Users\TensorFlow\anaconda3\envs\AI\lib\site-packages\lanms\__init__.py", line 2, in 
    from .adaptor import merge_quadrangle_n9 as nms_impl
ModuleNotFoundError: No module named 'lanms.adaptor'

定位到报错的源码，读源码可以看到其实这里是个递归，不知道什么原因函数的调用出了些问题。
此处改动源码请保存好原来的源码
直接替换掉 __init __.py 文件下的代码。

import numpy as np


def merge_quadrangle_n9(polys, thres=0.3, precision=10000):
    if len(polys) == 0:
        return np.array([], dtype='float32')
    p = polys.copy()
    p[:, :8] *= precision
    ret = np.array(merge_quadrangle_n9(p, thres), dtype='float32')
    ret[:, :8] /= precision
    return ret

---

后面两项直接安装： 

# 安装时注意关闭魔法
pip install lmdb
pip install tqdm