Fix historical retro earnings split

[konard]

Fixes #5

Summary

• split historical task earnings from the tasks.reward column instead of task_reward transaction totals
• keep retro earnings derived from task reward deltas, current_reward - reward, with legacy transaction fallbacks for databases without task rows
• prevent the cashback net line from showing -3 AP when there is no historical cashback/discount value
• document the historical earnings split in the README

Reproduction

Before this change, imported historical tasks could store the original task payout in reward and the retro-adjusted payout in current_reward, while older task_reward transactions still represented the adjusted total. Historical stats then counted the adjusted transaction amount as tasks and also counted current_reward - reward as retro, double-counting retro.

The regression test inserts historical task rows with reward values of 10 and 2.5, current_reward values of 15 and 3.75, and adjusted task transactions. The expected split is 12.5 AP from tasks and 6.25 AP from retro.

Tests

• PYTHONPATH=src python -m unittest discover -s tests
• python -m compileall -q src tests
• git diff --check

Notes

No UI changes or screenshots are involved.

[konard]

Working session summary

Implemented and pushed the fix to issue-5-317206e5e89e.

Changed:

• Historical task earnings now come from tasks.reward.
• Retro earnings come from current_reward - reward, with legacy transaction fallback.
• Cashback net no longer shows -3 AP when no cashback value exists.
• Added a regression test in tests/test_economy.py.
• Removed the placeholder .gitkeep.

Verification:

• PYTHONPATH=src python -m unittest discover -s tests passed: 25 tests.
• python -m compileall -q src tests passed.
• git diff --check passed.
• GitHub CI passed for Python 3.10, 3.11, and 3.12.

PR is updated and ready for review: #6

---

This summary was automatically extracted from the AI working session output.

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Model: GPT-5.5
• Provider: OpenAI
• Public pricing estimate: $5.288542

📊 Context and tokens usage:

• 168.9K / 1.1M (16%) input tokens, 19.9K / 128K (16%) output tokens

Total: (168.9K + 2.7M cached) input tokens, 19.9K output tokens, $5.288542 cost

🤖 Models used:

• Tool: OpenAI Codex
• Requested: gpt-5.5
• Model: GPT-5.5 (gpt-5.5)

📎 Log file uploaded as Gist (23453KB)

• View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.

[konard]

✅ Ready to merge

This pull request is now ready to be merged:

• All CI checks have passed
• No merge conflicts
• No pending changes

---

Monitored by hive-mind with --auto-restart-until-mergeable flag