Increase inventory capacity with base-100 encoding for values up to 25599

[daveey]

Increase Inventory Capacity with Base-100 Encoding

TL;DR

Increases the maximum inventory capacity from 255 to 65,535 and implements a base-100 encoding scheme for observation tokens to support larger inventory values.

What changed?

• Changed InventoryQuantity from uint8_t to uint16_t to support values up to 65,535
• Updated InventoryDelta from int16_t to int32_t to accommodate the full range of changes
• Implemented a base-100 encoding scheme for inventory observations:
  • inv:{resource} contains the ones digit (value % 100, 0-99)
  • inv:{resource}:ton contains the hundreds digit (value / 100, 0-255)
  • This allows representing values up to 25,599 (255 * 100 + 99)
• Added inventory token encoding/decoding logic in the observation system
• Updated the Python simulator to properly decode the base-100 encoded inventory values
• Removed inventory value capping at 255 in the configuration conversion

How to test?

Run the new tests in test_set_inventory.py which verify:

• Small inventory values (< 100) only emit the ones token
• Medium values (100-255) emit both ones and tons tokens
• Large values (> 255) correctly encode and decode using the base-100 scheme
• The maximum representable value (25,599) works correctly

Why make this change?

The previous 8-bit inventory limit of 255 was too restrictive for many gameplay scenarios. This change allows for much larger inventory quantities while maintaining backward compatibility with existing systems. The base-100 encoding scheme efficiently represents these larger values in the observation space without requiring additional tokens for most common inventory amounts.

[daveey]

• Increase inventory capacity with base-100 encoding for values up to 25599 #4355 👈 (View in Graphite)
• Refactor inventory system into dedicated InventoryConfig class #4356 : 2 other dependent PRs (#4328 , #4489 )
• Add ignore_limits parameter to inventory updates for initial setup #4360
• Rename test_performance to run_performance in mettagrid to avoid running in pytest #4370
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[sasmith]

I think this is no longer relevant?