feat: create v2-prep branch with comprehensive planning

Restructured project for V2 refactor:

**Structure Changes:**
- Moved all V1 code to orig/ folder (preserved with git mv)
- Created docs/planning/ directory
- Added orig/README_V1.md explaining V1 preservation

**Planning Documents:**
- 00_V2_MASTER_PLAN.md: Complete architecture overview
  - Executive summary of critical V1 issues
  - High-level component architecture diagrams
  - 5-phase implementation roadmap
  - Success metrics and risk mitigation

- 07_TASK_BREAKDOWN.md: Atomic task breakdown
  - 99+ hours of detailed tasks
  - Every task < 2 hours (atomic)
  - Clear dependencies and success criteria
  - Organized by implementation phase

**V2 Key Improvements:**
- Per-exchange parsers (factory pattern)
- Multi-layer strict validation
- Multi-index pool cache
- Background validation pipeline
- Comprehensive observability

**Critical Issues Addressed:**
- Zero address tokens (strict validation + cache enrichment)
- Parsing accuracy (protocol-specific parsers)
- No audit trail (background validation channel)
- Inefficient lookups (multi-index cache)
- Stats disconnection (event-driven metrics)

Next Steps:
1. Review planning documents
2. Begin Phase 1: Foundation (P1-001 through P1-010)
3. Implement parsers in Phase 2
4. Build cache system in Phase 3
5. Add validation pipeline in Phase 4
6. Migrate and test in Phase 5

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

orig/pkg/scanner/public.go

@@ -0,0 +1,47 @@
+package scanner
+
+import (
+	"github.com/fraktal/mev-beta/internal/config"
+	"github.com/fraktal/mev-beta/internal/logger"
+	"github.com/fraktal/mev-beta/pkg/cache"
+	"github.com/fraktal/mev-beta/pkg/contracts"
+	"github.com/fraktal/mev-beta/pkg/database"
+)
+
+// NewMarketScanner provides a backwards-compatible constructor that accepts
+// optional dependencies in the order (contract executor, database, reserve cache, ...).
+// It falls back to sane defaults when values are omitted so legacy tests and
+// tooling can continue to compile against the public API while more advanced
+// callers should use NewScanner directly for full control.
+func NewMarketScanner(
+	cfg *config.BotConfig,
+	log *logger.Logger,
+	extras ...interface{},
+) *Scanner {
+	var contractExecutor *contracts.ContractExecutor
+	var db *database.Database
+	var reserveCache *cache.ReserveCache
+
+	if len(extras) > 0 {
+		if v, ok := extras[0].(*contracts.ContractExecutor); ok {
+			contractExecutor = v
+		}
+	}
+
+	if len(extras) > 1 {
+		if v, ok := extras[1].(*database.Database); ok {
+			db = v
+		}
+	}
+
+	if len(extras) > 2 {
+		if v, ok := extras[2].(*cache.ReserveCache); ok {
+			reserveCache = v
+		}
+	}
+
+	// Additional parameters beyond the reserve cache are currently ignored but
+	// accepted to keep existing call sites compiling during the migration.
+
+	return NewScanner(cfg, log, contractExecutor, db, reserveCache)
+}