Best practices in coding for beginners
Download as PPTX, PDF4 likes682 views
Dr. Miguel Fierro's presentation covers best practices in Python programming for beginners, emphasizing its popularity in AI and ease of use. Key points include the importance of using Python 3.x, test-driven development, and adhering to software development principles like modularity and DRY (Don't Repeat Yourself). The document also discusses how to create a library, structure projects, and automate code styling using various IDEs.
![15
def sigmoid(x):
"""Sigmoid function. Output between 0-1
Args:
x (np.array): An array.
Returns:
result (np.array): Sigmoid of the array.
Examples:
>>> x = np.array([[1,1,-3],[0.5,-1,1]])
>>> sigmoid(x)
array([[0.73, 0.73, 0.04],
[0.62, 0.26, 0.73]])
"""
return 1. / (1 + np.exp(-x))
Documentation: Python docstrings
Python
Google style Numpy style
def sigmoid(x):
"""Sigmoid function. Output between 0-1
Parameters
-----------
x : np.array
An array.
Returns
-------
result : np.array
Sigmoid of the array.
Examples
--------
>>> x = np.array([[1,1,-3],[0.5,-1,1]])
>>> sigmoid(x)
array([[0.73, 0.73, 0.04],
[0.62, 0.26, 0.73]])
"""
return 1. / (1 + np.exp(-x))
Dr. Miguel Fierro - @miguelgfierro](https://image.slidesharecdn.com/miguelfierropythonbestpractices-180612145740/85/Best-practices-in-coding-for-beginners-15-320.jpg)
Best practices in coding for beginners
- 1. Dr. Miguel Fierro @miguelgfierro Best practices in coding for beginners
- 3. 3 Agenda The why, the what and the how of Python Development strategies Software development principles What’s next?
- 4. 4 Agenda The why, the what and the how of Python Development strategies Software development principles What’s next?
- 5. Why Python? Most popular language According to PYPL, Python surpassed Java in 2018 as the most popular programming language. The language of AI It has became the language of AI, machine learning and deep learning. Most of the top AI libraries use python. Easier As Python typically involves less code, it also takes less time to complete a job. This also means less money. Dr. Miguel Fierro - @miguelgfierro
- 6. 6 The final 2.x version 2.7 release came out in mid-2010 Python 2.7 will only receive necessary security updates until 2020 Python 3.5 was released in 2015 and Python 3.6 in 2016. All recent standard library improvements are available in Python 3.x by default Python 2.x Python 3.x Python 3.x is the present and future of the language Dr. Miguel Fierro - @miguelgfierro
- 7. 7 $ source activate my_env (my_env) $ pip install –r requirements.txt bash Python environments # file: requirements.txt Django==1.7.10 markdown2==2.3.0 Pillow==4.2.1 numpy>=1.13.3 opencv>=3.3.0 pyodbc>=4.0.21 requests>=2.18.4 beautifulsoup4>=4.6.0 pandas>=0.22.0 scikit-image>=0.13.1 scikit-learn>=0.19.1 scipy>=0.19.0 matplotlib>=1.5.3 file Python environments encapsulate development Dr. Miguel Fierro - @miguelgfierro
- 8. 8 $ conda create –n my_env –f conda.yaml $ source activate my_env (my_env) $ bash Conda environments # file: conda.yaml channels: - conda-forge - defaults dependencies: - python==3.6 - numpy==1.13.3 - flask==1.0.2 - cherrypy==12.0.1 - ipykernel==4.6.1 - pip: - pandas==0.22.0 - scipy==1.1.0 - scikit-learn==0.19.1 - jupyter==1.0.0 file Conda environments make development easier Dr. Miguel Fierro - @miguelgfierro
- 9. Typical project structure 9 docs robotics tests third_party tools .coveragerc .gitignore .travis.yml LICENSE.md README.md setup.py Example of library: robotics
- 10. 10 Agenda The why, the what and the how of Python Development strategies Software development principles What’s next?
- 11. 11 Jupyter notebooks Dr. Miguel Fierro - @miguelgfierro
- 12. Test Driven Development Develop the tests first 1) Add a test Write a test that defines a function or improvements of a function. It makes the developer focus on the requirements before writing the code, a subtle but important difference. 2) Run the test to check it fails Execute all the tests and check whether they pass or fail. The new test should fail for the expected reason. 3) Write the code Write some code that causes the test to pass. The new code written at this stage is not perfect and will be improved later. At this point, the only purpose of the written code is to pass the test. 5) Refactor code Clean the code, remove duplication, avoid magic numbers (constants without a variable) and apply other software development techniques. 4) Run the test to check it passes The new test should pass. Now the programmer can be confident that the new code meets the test requirements. 6) Repeat For each new functionality of the code, start by adding a new test Dr. Miguel Fierro - @miguelgfierro
- 13. 13 def test_predict_accuracy (self): acc = predict() self.assertGreaterEqual(acc, 0) self.assertLessEqual(acc, 1) Type of tests Python def test_predict_accuracy (self): acc = predict() self.assertGreaterEqual(acc, 0.65) def test_predict_accuracy (self): acc = predict() self.assertEqual(acc, 0.654345675) Unit tests Integration tests Utility tests It tests the correct behavior It tests that whether the result is acceptable It shows an example of the behavior Dr. Miguel Fierro - @miguelgfierro
- 14. 14 How to create a library Dr. Miguel Fierro - @miguelgfierro
- 15. 15 def sigmoid(x): """Sigmoid function. Output between 0-1 Args: x (np.array): An array. Returns: result (np.array): Sigmoid of the array. Examples: >>> x = np.array([[1,1,-3],[0.5,-1,1]]) >>> sigmoid(x) array([[0.73, 0.73, 0.04], [0.62, 0.26, 0.73]]) """ return 1. / (1 + np.exp(-x)) Documentation: Python docstrings Python Google style Numpy style def sigmoid(x): """Sigmoid function. Output between 0-1 Parameters ----------- x : np.array An array. Returns ------- result : np.array Sigmoid of the array. Examples -------- >>> x = np.array([[1,1,-3],[0.5,-1,1]]) >>> sigmoid(x) array([[0.73, 0.73, 0.04], [0.62, 0.26, 0.73]]) """ return 1. / (1 + np.exp(-x)) Dr. Miguel Fierro - @miguelgfierro
- 16. 16 Code style: PEP8 Python's official style guide, for formatting code correctly to maximize its readability. Created in 2001 by Guido van Rossum et. al. !DILEMMA: How to automatize the code style? Use auto-format in your favorite IDE
- 17. PyCharm Free and payed version Visual Studio Heavy weight IDE with many functionalities VS Code Ubuntu default Atom Results visualization No matter which IDE you prefer, it will save you time and make the development easier. Dr. Miguel Fierro - @miguelgfierro
- 18. 18 Agenda The why, the what and the how of Python Development strategies Software development principles What’s next?
- 19. Software development principles Single responsibility DRY (Don’t Repeat Yourself) MVP (Minimum Viable Product) Plan for reuse Key concepts for creating great software Dr. Miguel Fierro - @miguelgfierro
- 20. Modularity Single responsibility Each class or module should be responsible for a single part of the software, in other words: “A class should have only one reason to change”. Part of a broader set of principles called SOLID. A class does only one thing, a function does only one thing Dr. Miguel Fierro - @miguelgfierro
- 21. DRY If a piece of code can be generalized and reuse, do it! Extract a new function or a private method. Reducing complexity Don’t Repeat Yourself … please Dr. Miguel Fierro - @miguelgfierro
- 22. Fail fast MVP The MVP is the minimum product that collects the maximum amount of validated learning about customers with the least effort. Based on the Lean Startup methodology. Minimum Viable Product Dr. Miguel Fierro - @miguelgfierro
- 23. Generalize your code to be reusable Generalization Plan for reuse Whenever you create a piece of software, design it in a way so it can be reusable in other projects. It will save you a lot of time. Dr. Miguel Fierro - @miguelgfierro
- 24. 24 Agenda The why, the what and the how of Python Development strategies Software development principles What’s next?
- 25. 25 Create your own codebase Dr. Miguel Fierro - @miguelgfierro
- 26. 26 Start contributing to opensource Dr. Miguel Fierro - @miguelgfierro
- 27. Dr. Miguel Fierro @miguelgfierro Best practices in coding for beginners