We Keep Coding

Converting a Python project to DLL or decreasing the size and imports

I have a python project for OCR MRZ detection with 2 modules 1 is for ID which uses EasyOcr,pythorch and other one is for Passport documents which uses Pytesseract and tensorflow.
I need to prepare this project for deployment I have tried some methods but none of them was practible for deployment process.
I have tried pyinstaller with couple of configurations with --onefile
option the setup is great but it takes too long to unpack the exe
when executed.
I have then tried --onedir option the delay was gone but now
installation package was too complicated and size was too
large(1.8GB).
I have tried to "compile" python code by using Cython but even with a
helloworld.py sample app I couldn't manage to make this one work I
got couple of errors during gcc compiling the last error I got due to
msvcp package which i have installed but still got the error.
And as last I have used Nuitka to get a dll-like file to import this
in C# and use it like a package, I have successfully created a test
.pyd file from a helloworld.py but i couldn't import it in C# as i
planned.
What I need is to prepare this project as a simpler and low-sized application which is hard to reverse-engineered for source codes ready for deployment. For passport OCR I can switch development to C# but for ID I couldn't find any alternative OCR library to get the MRZ information so at least I need to use ID OCR module from my Python project.
Any help would be appreciated,
Thanks

import matplotlib failed while deploying my model in AWS sagemaker

I have deployed my AWS model successfully.
but while testing i am getting runtime Error: "import matplotlib.pyplot as plt" . I think it is due to pytorch framework version i used(framework_version=1.2.0). I am facing the same issue when i use higher versions as well.
PyTorchModel(model_data=model_artifact,
role = role,
framework_version=1.2.0,
entry_point='predict.py',
predictor_cls=ImagePredictor)
I have other issue when i use version=1.0.0. i.e i am not able to import libraries from sub directories and deployment itself is failing.
Eg: i have some code files in "Code" directory.
from Code.CTModel import NetWork ---> **this line will fail as "No module named Code" when i use version=1.0.0**
Ultimately i want to how to use/import libraries which are written under sub-directories.

It sounds like you want to inject some additional code libraries into the SageMaker PyTorch serving container. You might have to dig into the source code for how the PyTorch serving container is built to further customize it: https://github.com/aws/sagemaker-pytorch-inference-toolkit, or build your own image.
Digging into that source code a bit, I see that the container has enabled the importing of arbitrary code, but only when "multi-model mode" is enabled. Can you verify that the code exists under a directory "code" in your model directory and that "multi-model mode" is enabled?
def initialize(self, context):
# Adding the 'code' directory path to sys.path to allow importing user modules when multi-model mode is enabled.
if (not self._initialized) and ENABLE_MULTI_MODEL:
code_dir = os.path.join(context.system_properties.get("model_dir"), 'code')
sys.path.append(code_dir)
self._initialized = True
Reference: https://github.com/aws/sagemaker-pytorch-inference-toolkit/blob/c4e7abc49aeebc2f9b6035337548a90e4330113d/src/sagemaker_pytorch_serving_container/handler_service.py#L47
If this all seems complicated to you (it is), you might want to look into some standardized formats for serializing your PyTorch model such as https://onnx.ai/. I'd love to learn more about what you're trying to do here sometime if you reach out to me at contact#modelzoo.dev. I'm beta-testing a platform that enables deployment in a single line of code and would love to test it out here.

Let me make my query little bit high level: I have predict.py, jupyter notebook , Code(Direcotry),Evoludation(directory) and other .py files in source_dir.
--Code
--ResNet.py
--Densenet.py
--DataLoader.py
--Evaluation
--Evaluation.py
--predict.py
--CT_Code.ipynb
When i execute the predict file from jupyter notebook in my local system, all the modules are imported properly and everything is working fine. But when i am deploying same thing in sagemaker notebook facing issues as mentioned in my question.(Not able to import libraries from Code directory and some basic modules like imageio,PIL, Matplotlib)

What is a pytorch snapshot and a Dlib model? (Running Hopenet)

I am trying to make Hopenet run on my computer, using the code on github
I have installed all required libraries. This is my forked code, with test_hopenet.py updated to python 3.
I installed all required libraries, and using pip install "pillow<7" because of some old requirements.
python code/test_on_video_dlib.py --snapshot PATH_OF_SNAPSHOT --face_model PATH_OF_DLIB_MODEL --video PATH_OF_VIDEO --output_string STRING_TO_APPEND_TO_OUTPUT --n_frames N_OF_FRAMES_TO_PROCESS --fps FPS_OF_SOURCE_VIDEO
It looks like I am missing some basic understanding which is supposed to be obvious:
What is a SNAPSHOT? Where do I get one?
What is a Dlib Model? Where do I get the right one?
Again - I just want to make this code run, but I can't understand the instructions.

For the dlib model you can use:
https://github.com/davisking/dlib-models/mmod_human_face_detector.dat.bz2
For the snapshot path you can use hopenet_robust_alpha1.pkl downloadable from the link "300W-LP, alpha 1, robust to image quality" under Pretrained Models section of Nathanial Ruiz's README.md

Error in calling a Python script from a Basic function

I am new in Python and Basic. I am trying to replicate the IMPORTHTML function from Google Sheets in LibreOffice (LO) Calc. In a nutshell, I want to create a GetHtmTable( Url, Table Index) Basic function in Calc which will call a Python script to do the heavy work.
So based on Villeroy's great example, I implemented in LO 5.1.6.2. the Basic SOUNDEX function which calls the Python script sheetFunctions.py to get familiar with the process. My environment is Linux Mint 18, I use Python 3, I imported all kind of libraries such as Uno, PIP etc.. I use PycharmProjects as a Python editor.
I see clearly under the LO Calc menu tools->macro->organize macros->python the sheetFunctions.py Python script, which indeed is in the folder /usr/lib/libreoffice/share/Scripts/python.
Whenever, I run the SOUNDEX Basic function I see the following error message:
BASIC runtime error. An exception occurred Type:
com.sun.star.script.provider.ScriptFrameworkErrorException Message:
: an
error occurred during file opening
/usr/lib/libreoffice/program/pythonscript.py:429 in function
getModuleByUrl() [lastRead = self.sfa.getDateTimeModified( url )]
/usr/lib/libreoffice/program/pythonscript.py:993 in function
getScript() [mod = self.provCtx.getModuleByUrl( fileUri )]
I tried to debug the SOUNDEX basic function and found out the blocking point is when the program runs getScript("vnd.sun.star.script:sheetFunctions.py$soundex?language=Python&location=user").
I've been trying for days now to overcome this error, unsuccessfully I must confess.
I wonder if I need to bring some extra extensions in the Basic environment or a missing add-in in the Linux/Python one?
I changed the location=user by location=document and got stuck again. I added recently libreoffice-script-provider-python thanks to the command sudo apt-get install libreoffice-script-provider-python but this did not help. I also embedded in the Calc document the Python script but same this did not solve the issue.

The location name does not match. The standard place for self-written scripts is under the user directory. This is location=user, for example ~/.config/libreoffice/4/user/Scripts/python.
Then there is location=share, which refers to the path in your question. These parameters are described under Python Script in the URI Specification.
See also my answer to this question. Be sure to try the APSO extension if you haven't yet. Especially, APSO helps when using location=document, because embedding requires several steps including editing manifest.xml.

Got errors, while running exe file built with pyinstaller and Google Cloud API integration in python

I am working one file python project.
I integrated google-cloud-API for realtime speech streaming and recognition.
It works with python aaa.py command well.
Now I need windows build file(.exe), so I used pyinstaller program and I got aaa.exe file successfully.
But I got this error while running speech streaming by using Google cloud API.
[Errno 2] No such file or directory:
'D:\AI\ai\dist\AAA\google\cloud\gapic\speech\v1\speech_client_config.json'
So I copied this speech_client_config.json file in needed path, after that I got below error again.
Exception in 'grpc._cython.cygrpc.ssl_roots_override_callback'
ignored E0511 01:13:14.320000000 3108
src/core/lib/security/security_connector/security _connector.cc:1170]
assertion failed: pem_root_certs != nullptr
Then, I can not find solution to get working version with google-cloud API.
I am using python version 2.7.14
I need your friendly help.
Thanks.

I had the same problem. If you are willing to distribute roots.pem with your executable (just search for the file - it should be buried deep within the installation directory of grpcio), I had luck fixing this by setting GRPC_DEFAULT_SSL_ROOTS_FILE_PATH environment variable to the full path of this roots.pem file.

Update 2021
To anyone who is experiencing this issue. I got it working thanks to these amazing people. See the full conversation on this github issue.
Here is the link
Step 1
Credits to #cbenhagen & #rising-stark on this github link.
A PyInstaller hook called hook-grpc.py looking like this would do the trick:
Create a python file named hook-grpc.py with this code.
from PyInstaller.utils.hooks import collect_data_files
datas = collect_data_files('grpc')
Step 2
Put the hook-grpc.py file in your \site-packages\PyInstaller\hooks directory of the python environment you are running on. So basically you can find it at
C:\Users\yourusername\AppData\Local\Programs\Python\Python37\Lib\site-packages\PyInstaller\hooks
Note:
Just change the yourusername and Python37 to your
respective username and python version you are using.
For Anaconda users it might be different. Check this site
to find the anaconda python environment path you are using.
Step 3
Once you've done that you can now convert your .py python program to .exe using pyinstaller and it should work.

This looks to me like a SSL credentials mistake. I think you are not being allowed to GC. Check this code snippet and this documentation.