Phase 1: Battle Report → Chart

This page is Developer Documentation. It summarizes Phase 1 deliverables and constraints for maintainers.

Phase 1 proves the end-to-end pipeline works:

1. paste a Battle Report (raw text),
2. store it safely with dedupe,
3. analyze it via the Analysis Engine,
4. render a simple time-series chart.

UI

• Dashboard: /
• Paste/import form (POST)
• Date range filter (GET start_date, end_date)
• Chart: coins/real hour over time

See docs/phase2.md for contextual filters, preset labels, overlays, and delta comparisons.

Data Stored

• gamedata.models.BattleReport
• raw Battle Report text is preserved unchanged
• deduplicated via a SHA-256 checksum
• gamedata.models.BattleReportProgress
• stores only Phase 1 metadata: battle date, tier, wave, real time
• accepts labels separated by colons, tabs, or multi-space alignment
• tolerates missing/reordered sections and unknown labels (non-fatal)
• Battle Dates may use ISO, numeric, or month-name formats (e.g. Dec 07, 2025 21:59)
• parses real-game tab-delimited reports (e.g. Battle Date\tDec 14, 2025 01:39)

Analysis Engine

• Entry point: analysis.engine.analyze_runs
• Input: iterable of RunProgress-like objects (duck-typed)
• Output: DTOs in analysis.dto (no database writes)
• Coins are parsed best-effort from raw text via analysis.quantity (e.g. Coins: 4.24M).
• Missing required fields are skipped so empty inputs return an empty result instead of errors.

Limitations (By Design)

• Only extracts Battle Date, Tier, Wave, Real Time
• Ignores all other labels (unknown labels are non-fatal)
• Computes only one metric: coins/real hour