Tutorial about serverless functions

[nmanovic]

Fixed #2926
Fixed #2270
Fixed #2259
Fixed #2185

Motivation and context

Users and customers have problems with running automatic and semi-automatic annotation in CVAT using serverless function. They don't know how to integrate own DL models. The tutorial answers on many typical questions which I saw as github issues or requests in gitter.

How has this been tested?

• Clone CVAT repo and CD to cvat
• Checkout the branch nm/serverless_tutorial
• Go to site directory
• Run hugo server -D command
• Open http://localhost:1313/docs/manual/advanced/serverless-tutorial/ page

Checklist

• I submit my changes into the develop branch
• I have added description of my changes into CHANGELOG file
• I have updated the documentation accordingly
• I have added tests to cover my changes
• I have linked related issues (read github docs)
• I have increased versions of npm packages if it is necessary (cvat-canvas,
  cvat-core, cvat-data and cvat-ui)

License

• I submit my code changes under the same MIT License that covers the project.
  Feel free to contact the maintainers if that's a concern.
• I have updated the license header for each file (see an example below)

# Copyright (C) 2021 Intel Corporation
#
# SPDX-License-Identifier: MIT

[coveralls]

Coverage increased (+0.02%) to 74.094% when pulling 3e378ff on nm/serverless_tutorial into 38b7740 on develop.

[sevaraMS]

When it will be ready???. I need tutorial

[sevaraMS]

I tried use onepanel, but it is not suitable for windows

[sevaraMS]

with openvino not worked))))))))))))))))

[sevaraMS]

Do we have another way to add our own trained model for auto annotation????

[TOsmanov]

I encountered a problem when running nuctl from windows 10, I managed to deployment models only using ubuntu. I took the following actions:

1. Installation ubuntu 18.04 from Microsoft store

2. In settings docker desktop in tab resourcrces wsl integration enable integration for ubuntu-18.04

   

3. Install curl command:

   sudo apt install curl

4. Installation nuctl on ubuntu, with use command from manuals on nuclio CLI:

   curl -s https://api.github.com/repos/nuclio/nuclio/releases/latest \
               | grep -i "browser_download_url.*nuctl.*$(uname)" \
               | cut -d : -f 2,3 \
               | tr -d \" \
               | wget -O nuctl -qi - && chmod +x nuctl

5. Installation git on ubuntu:

   sudo apt install git

6. Cloning to the repository and moving:

   git clone https://github.com/openvinotoolkit/cvat.git --branch nm/serverless_tutorial
   cd cvat
   

7. Executing commands from serverless tutorial

There was a problem with the deployment of models
(similar to this #2828) I solved it by adding it to the function.yaml file of the each model, attribute port: 32001

triggers:
myHttpTrigger:
    maxWorkers: 1
    kind: 'http'
    workerAvailabilityTimeoutMilliseconds: 10000
    attributes:
    port: 32001 <----- add this ( different ports for different models)
    maxRequestBodySize: 33554432 # 32MB

May be worth adding information that Windows 10 users are better off deploying models through ubuntu, or adding this information to section "Troubleshooting"?

[nmanovic]

I encountered a problem when running nuctl from windows 10, I managed to deployment models only using ubuntu. I took the following actions:

1. Installation ubuntu 18.04 from Microsoft store
2. In settings docker desktop in tab resourcrces wsl integration enable integration for ubuntu-18.04
   
3. Install curl command:

   sudo apt install curl

4. Installation nuctl on ubuntu, with use command from manuals on nuclio CLI:

   curl -s https://api.github.com/repos/nuclio/nuclio/releases/latest \
               | grep -i "browser_download_url.*nuctl.*$(uname)" \
               | cut -d : -f 2,3 \
               | tr -d \" \
               | wget -O nuctl -qi - && chmod +x nuctl

5. Installation git on ubuntu:

   sudo apt install git

6. Cloning to the repository and moving:

   git clone https://github.com/openvinotoolkit/cvat.git --branch nm/serverless_tutorial
   cd cvat
   

7. Executing commands from serverless tutorial

There was a problem with the deployment of models
(similar to this #2828) I solved it by adding it to the function.yaml file of the each model, attribute port: 32001

triggers:
myHttpTrigger:
    maxWorkers: 1
    kind: 'http'
    workerAvailabilityTimeoutMilliseconds: 10000
    attributes:
    port: 32001 <----- add this ( different ports for different models)
    maxRequestBodySize: 33554432 # 32MB

May be worth adding information that Windows 10 users are better off deploying models through ubuntu, or adding this information to section "Troubleshooting"?

Thanks for these instructions. I will add the information into the tutorial.

[nmanovic]

@dmitriikhurtin , could you please help me to understand the issue? In the PR I see a couple of errors from pylint linter. For example,

    {
        "type": "error",
        "module": "main",
        "obj": "init_context",
        "line": 19,
        "column": 12,
        "path": "serverless/openvino/omz/intel/semantic-segmentation-adas-0001/nuclio/main.py",
        "symbol": "too-many-function-args",
        "message": "Too many positional arguments for constructor call",
        "message-id": "E1121"
    }

If I check the file locally, the linter doesn't complain. How should I run the linter locally to reproduce the problem?

$ echo "Pylint version: "`pylint --version | head -1`
Pylint version: pylint 2.7.0
$ echo "The files will be checked: "`echo ${changed_files_pylint}`
The files will be checked: /home/nmanovic/Workspace/cvat/serverless/openvino/omz/intel/semantic-segmentation-adas-0001/nuclio/main.py
$ pylint --verbose ${changed_files_pylint} --output-format json
Using config file /home/nmanovic/Workspace/cvat/.pylintrc
[]