Add money: statement-driven personal finance tracker

A data directory holds one folder per account. Statements dropped into
those folders are parsed into a rebuildable SQLite index, categorised by
ordered glob rules in rules.toml, and browsed or hand-tagged in a Bubble
Tea TUI. Movements between the user's own accounts are marked as
transfers by the same rules and excluded from spending totals.

Manual tags and transfer marks are stored separately from the rule-derived
ones and always win, so editing rules.toml and re-running retag never
destroys hand edits.

Parsers are pluggable. Three are ported from the Python extractors they
replace -- nlb and traderepublic read PDFs via pdftotext -layout, revolut
reads the CSV export -- alongside a configurable-column CSV parser and a
cmd parser that shells out to an external script.

Both ports fix two latent bugs in the originals: the sign character class
rejected the typographic minus U+2212 that some PDF fonts emit, and NLB's
hardcoded continuation indent broke when pdftotext compressed runs of
spaces, so the threshold is now measured from the description column.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

internal/parser/traderepublic_test.go

@@ -0,0 +1,167 @@
+package parser
+
+import (
+	"strings"
+	"testing"
+)
+
+// place builds a fixed-width line by putting each string at a given column,
+// which keeps these fixtures honest about the layout the parser measures.
+func place(cells map[int]string) string {
+	width := 0
+	for col, text := range cells {
+		if end := col + len(text); end > width {
+			width = end
+		}
+	}
+	line := []byte(strings.Repeat(" ", width))
+	for col, text := range cells {
+		copy(line[col:], text)
+	}
+	return string(line)
+}
+
+// Column starts, spaced as a real statement is: wide enough that every token
+// of "05 Jan 2026" stays closest to DATE rather than drifting into TYPE.
+const (
+	colDate     = 2
+	colType     = 20
+	colDesc     = 40
+	colMoneyIn  = 75
+	colMoneyOut = 90
+	colBalance  = 105
+)
+
+func trHeaderLine() string {
+	return place(map[int]string{
+		colDate: "DATE", colType: "TYPE", colDesc: "DESCRIPTION",
+		colMoneyIn: "MONEY IN", colMoneyOut: "MONEY OUT", colBalance: "BALANCE",
+	})
+}
+
+// A statement mixing both layouts: one transaction on a single line, one with
+// its fields stacked across three, and one whose description wraps.
+func trStatement() string {
+	lines := []string{
+		"                    Trade Republic Bank GmbH",
+		"",
+		trHeaderLine(),
+		"",
+		// Everything on one line.
+		place(map[int]string{
+			colDate: "05 Jan 2026", colType: "Deposit", colDesc: "Payment from ACME",
+			colMoneyIn: "€2,500.00", colBalance: "€2,500.00",
+		}),
+		"",
+		// Stacked across three lines.
+		place(map[int]string{
+			colDate: "06 Jan 2026", colMoneyOut: "€45.20", colBalance: "€2,454.80",
+		}),
+		place(map[int]string{colType: "Card"}),
+		place(map[int]string{colDesc: "LIDL SOFIA 4412"}),
+		"",
+		// Wrapped description, hyphen-broken on the first line, plus an IBAN.
+		place(map[int]string{
+			colDate: "10 Jan 2026", colType: "Transfer", colDesc: "Standing order to sav-",
+			colMoneyOut: "€500.00", colBalance: "€1,954.80",
+		}),
+		place(map[int]string{colDesc: "ings SI56123456789012345"}),
+		"",
+		"  Page 1 of 2",
+	}
+	return strings.Join(lines, "\n")
+}
+
+func TestParseTradeRepublicText(t *testing.T) {
+	txns, err := parseTradeRepublicText(trStatement(), 2)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(txns) != 3 {
+		t.Fatalf("got %d transactions, want 3: %+v", len(txns), txns)
+	}
+
+	deposit := txns[0]
+	if deposit.Date != "2026-01-05" {
+		t.Errorf("date = %q, want 2026-01-05", deposit.Date)
+	}
+	if deposit.AmountMinor != 250000 {
+		t.Errorf("money in = %d, want 250000", deposit.AmountMinor)
+	}
+	if deposit.Type != "Deposit" {
+		t.Errorf("type = %q, want Deposit", deposit.Type)
+	}
+	if deposit.Description != "Payment from ACME" {
+		t.Errorf("description = %q", deposit.Description)
+	}
+
+	// The stacked layout must produce the same shape as the single-line one.
+	card := txns[1]
+	if card.Date != "2026-01-06" {
+		t.Errorf("date = %q, want 2026-01-06", card.Date)
+	}
+	if card.AmountMinor != -4520 {
+		t.Errorf("money out = %d, want -4520", card.AmountMinor)
+	}
+	if card.Type != "Card" {
+		t.Errorf("type = %q, want Card", card.Type)
+	}
+	if card.Description != "LIDL SOFIA 4412" {
+		t.Errorf("description = %q", card.Description)
+	}
+	if card.BalanceMinor == nil || *card.BalanceMinor != 245480 {
+		t.Errorf("balance = %v, want 245480", card.BalanceMinor)
+	}
+
+	// A hyphen-broken word rejoins without a space. The hyphen itself is kept,
+	// as the original script does, because descriptions contain real hyphens
+	// that must not be swallowed.
+	transfer := txns[2]
+	if transfer.Description != "Standing order to sav-ings SI56123456789012345" {
+		t.Errorf("description = %q, want the wrapped fragment joined without a space", transfer.Description)
+	}
+	if transfer.Counterparty != "SI56123456789012345" {
+		t.Errorf("counterparty = %q", transfer.Counterparty)
+	}
+	if transfer.AmountMinor != -50000 {
+		t.Errorf("amount = %d, want -50000", transfer.AmountMinor)
+	}
+}
+
+// Pages with no transaction table (cover pages, disclaimers) are skipped.
+func TestParseTradeRepublicSkipsPagesWithoutTable(t *testing.T) {
+	text := "  Some cover page with no table at all\n" + "\f" + trStatement()
+	txns, err := parseTradeRepublicText(text, 2)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(txns) != 3 {
+		t.Errorf("got %d transactions, want 3", len(txns))
+	}
+}
+
+// A figure inside the description must not be mistaken for a money column.
+func TestParseTradeRepublicIgnoresAmountsInDescription(t *testing.T) {
+	text := strings.Join([]string{
+		trHeaderLine(),
+		"",
+		place(map[int]string{
+			colDate: "05 Jan 2026", colType: "Card", colDesc: "Refund of €12.00 order",
+			colMoneyIn: "€12.00", colBalance: "€100.00",
+		}),
+	}, "\n")
+
+	txns, err := parseTradeRepublicText(text, 2)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(txns) != 1 {
+		t.Fatalf("got %d transactions, want 1", len(txns))
+	}
+	if txns[0].AmountMinor != 1200 {
+		t.Errorf("amount = %d, want 1200", txns[0].AmountMinor)
+	}
+	if !strings.Contains(txns[0].Description, "€12.00") {
+		t.Errorf("description = %q, want the inline figure kept", txns[0].Description)
+	}
+}