[PR] Fix: issue #57, add support for "Undecided" match results in outcome processing

[scmisa]

No description provided.

[Kaszanas]

Let's see what copilot says! 🤔

[Copilot]

Pull Request Overview

This PR adds support for handling "Undecided" match results in the SC2 dataset processing pipeline by updating outcome processing functions to recognize and skip replays with undecided results.

• Added "Undecided": -1 mapping to result dictionaries across transform functions
• Implemented early return logic to skip processing when "Undecided" results are detected
• Updated three transform functions to handle the new result type consistently

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 5 comments.

File	Description
src/sc2_datasets/transforms/utils.py	Added "Undecided" result mapping and early return logic in outcome processing
src/sc2_datasets/transforms/pytorch/economy_vs_outcome.py	Added "Undecided" handling to economy statistics transformation
src/sc2_datasets/transforms/mmr_vs_result.py	Added "Undecided" handling to MMR vs result transformation

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Kaszanas]

I think there needs to be a little more thought put into this one. I believe my comments may be helpful ;)

Another thing that we have discussed is maybe having a validator that checks for any undecided games if someone would like to exclude these from modeling.

Check out if you are able to fix the CI. The automatic formatting should be a part of the development environment. Let me know if that works.

[Copilot]

Pull Request Overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[scmisa]

This pull request improves the robustness of several data transformation functions by handling edge cases where a StarCraft II replay result is "Undecided", "Draw", or "Tie". Now, these cases are explicitly checked and skipped, preventing ambiguous or invalid labels from being processed in downstream tasks.

Enhanced handling of ambiguous game results:

• Updated result_dict in mmr_vs_result.py, economy_vs_outcome.py, and utils.py to include mappings for "Undecided", "Draw", and "Tie" results, assigning them a value of -1. [1] [2] [3]
• Added logic to all three modules to check for these ambiguous results and return None (or skip processing) when encountered, ensuring such replays are excluded from further analysis. [1] [2] [3]