Max stat points (general strategy creation UX thread)

[rand0musername]

We should definitely document the health+points+2strength+2defence+2special == 100 somewhere. Just a reminder so we don't forget. Could cause random disqualifications.

[rand0musername]

In general, let's enable them to focus on writing the core logic with clear instructions and understandable examples. Issue #4 is also relevant here.

[rand0musername]

Unrelated to this issue but maybe we should use some reflection and import os magic to fetch all python files in a folder called strategies/ from run.py. This way the flow is:

1. The competitor clones this repo.
2. The competitor copies one of the files from strategies/, renames it to '__strategy.py' and renames the class inside to 'Strategy'.
3. The competitor rewrites the logic and it all seamlessly works with run.py, no other code changes are needed.
4. For submission, the competitor sends only their python file.
5. We put all received python files in our strategies/ folder (we could do strategies/dayN or something and publish all of them later instead of overwriting).
6. We run the tournament, no other changes are needed. Maybe now we need a way to exclude example strategies but we could hack that up in the import os magic code.

If you already have better ideas and a submission pipeline set in stone disregard this, just rambling.

[Dreian]

We should definitely document the health+points+2strength+2defence+2special == 100 somewhere. Just a reminder so we don't forget. Could cause random disqualifications.

Agreed. Already have a draft doc. Should discuss whether there should be a "level up" system that raises this max.

[Dreian]

In general, let's enable them to focus on writing the core logic with clear instructions and understandable examples. Issue #4 is also relevant here.

Will document.

[Dreian]

Unrelated to this issue but maybe we should use some reflection and import os magic to fetch all python files in a folder called strategies/ from run.py. This way the flow is:

1. The competitor clones this repo.

2. The competitor copies one of the files from `strategies/`, renames it to '__strategy.py' and renames the class inside to 'Strategy'.

3. The competitor rewrites the logic and it all seamlessly works with `run.py`, no other code changes are needed.

4. For submission, the competitor sends only their python file.

5. We put all received python files in our `strategies/` folder (we could do `strategies/dayN` or something and publish all of them later instead of overwriting).

6. We run the tournament, no other changes are needed. Maybe now we need a way to exclude example strategies but we could hack that up in the `import os` magic code.

If you already have better ideas and a submission pipeline set in stone disregard this, just rambling.

The pipeline does make sense and is what I had in mind. Someone will need to look at Python reflection capabilities. I'd keep the random strategies as our bots that also participate and provide easy victories.

[rand0musername]

Easy victories make sense 👍
Regarding Python reflection, maybe something like this to find all python files in strategies/):
onlyfiles = [f for f in listdir(dir) if isfile(join(dir, f)) and f.endswith('.py')]
Then somethig like this to generate class names from each filename:
class_name = ''.join([w.title() for w in filename[:-3].split('_')])
And then we can do:
globals()[class_name]
Of course we need to from strategies include * or something similar before.

Definitely ugly but if should work 90% of the time if competitors name their files correctly :)
I just really like the idea of not having to change the engine to include a new strategy. Maybe finding a way to include all classes from a file given filename would be less hacky. Just some ideas if anyone wants to tackle this.