After a simple c&p from GH repo when streamlit run simple_plot.py

File "C:\Users\myuser\AppData\Local\Programs\Python\Python312\Lib\site-packages\py_wanderer\plotter.py", line 21, in plot_maze_with_paths
cmap = cm.get_cmap("tab20")
^^^^^^^^^^^
AttributeError: module 'matplotlib.cm' has no attribute 'get_cmap'

I dont normally use poetry so could be wrong here.

Suggestion

I like the idea of using two examples, each to showcase the advantages of either using functions or classes. I think an important "feature" of classes is that you can create multiple instances of something. Can you add a line about that? I think the bank accounts example actually shows that quite clearly, since you can have multiple bank accounts, so classes work great there.

https://github.com/ArjanCodes/examples/blob/main/2023/apitesting/main_split.py#L54
The item parameter should be annotated as ItemUpdate.

https://github.com/ArjanCodes/examples/blob/main/2023/apitesting/operations.py#L59
Missing ItemUpdate as a parameter and in for statement.

hi Arjan,

In you Streamlit/openai app you may benefit from using more params the openAI API offers.
For example, because the first question ("which programming language?") expects a single word, you can use max_tokens to enforce a single-word answer. Likewise, logit_bias might help. Also, for a nicer UX you can stream the model's completion.

Happy to share an example,
Best,

Clem

Can you make sure your code examples don't have type errors if you switch on strict type checking in vs code?

This is always recommended, even if there are no external dependencies, mostly because it points to the correct Python version used. It also points to the idea that every code should have an isolated environment to run, and poetry is the default package to manage it.

https://github.com/ArjanCodes/wip/blob/30ac17044f1e1b2592e330ee14301fa3eecea9b0/funclass/classes.py#L18-L25

Suggestion:

What is the best practice for creating custom exception objects? I see you added a 'msg' parameter, but I think exceptions already support this, so there's no need for a new instance variable, and perhaps even the str is not needed.

In 2024/tuesday_tips/fail_fast, I tried running the tests, and two of the three tests fail.

===================================================== short test summary info =====================================================
FAILED tests/test_main.py::test_successful_employee_update - sqlite3.OperationalError: no such table: employees
FAILED tests/test_main.py::test_update_non_existing_employee - sqlite3.OperationalError: no such table: employees
=================================================== 2 failed, 1 passed in 1.26s ===================================================

How to reproduce:

cd 2024/tuesday_tips/fail_fast
poetry install
poetry run pytest

Is that the correct procedure to run the tests? Is some other command needed to initialize the database? README.md was empty, so the above commands were my guess at how to do it.

Thanks!

The dataset is indeed quite large. Can we use a subset of the data so that no special unzipping action is needed? I prefer to have an example that runs out of the box.

To run the code, there is a step to do: unzip the file, because GitHub has a limitation on uploading files greater than 100Mb. Grabbing a small sample of the dataset, and being able to upload it directly, would resolve this issue.

Suggestion:

Can you make sure your code examples don't have type errors if you switch on strict type checking in vs code?