feat(#18): import de extratos OFX e CSV — parser, dedup, preview e confirmação

Parser OFX tolerante ao formato SGML de bancos brasileiros (sem closing tags, BOM, timezone suffix). Parser CSV com separador de campo/decimal configurável. Dedup por FITID (OFX) ou date+amount+desc normalizado. Fluxo: POST /preview → revisão na UI → POST /confirm salva e loga. Migration 002 adiciona external_id com unique index. Co-Authored-By: Claude Sonnet 4.6 <[email protected]>
2026-05-26 20:17:46 -03:00
parent d6782d51a5
commit 83dff2de78
15 changed files with 1065 additions and 3 deletions
@@ -0,0 +1,99 @@
+package parser
+
+import (
+	"encoding/csv"
+	"fmt"
+	"io"
+	"math"
+	"strconv"
+	"strings"
+	"time"
+
+	"financeiro-carvalho/internal/model"
+)
+
+func ParseCSV(r io.Reader, m model.CSVMapping) ([]model.ImportRow, []string) {
+	if m.DateFormat == "" {
+		m.DateFormat = "02/01/2006"
+	}
+	if m.DecimalSeparator == "" {
+		m.DecimalSeparator = ","
+	}
+	fieldSep := ','
+	if m.FieldSeparator == ";" {
+		fieldSep = ';'
+	}
+
+	reader := csv.NewReader(r)
+	reader.Comma = fieldSep
+	reader.LazyQuotes = true
+	reader.TrimLeadingSpace = true
+
+	var rows []model.ImportRow
+	var errs []string
+	lineNum := 0
+
+	for {
+		record, err := reader.Read()
+		if err == io.EOF {
+			break
+		}
+		lineNum++
+		if err != nil {
+			errs = append(errs, fmt.Sprintf("linha %d: %v", lineNum, err))
+			continue
+		}
+		if lineNum == 1 && m.HasHeader {
+			continue
+		}
+
+		maxCol := max3(m.DateColumn, m.AmountColumn, m.DescriptionColumn)
+		if len(record) <= maxCol {
+			errs = append(errs, fmt.Sprintf("linha %d: colunas insuficientes (%d)", lineNum, len(record)))
+			continue
+		}
+
+		dateStr := strings.TrimSpace(record[m.DateColumn])
+		amtStr := strings.TrimSpace(record[m.AmountColumn])
+		desc := cleanDescription(record[m.DescriptionColumn])
+
+		t, err := time.Parse(m.DateFormat, dateStr)
+		if err != nil {
+			errs = append(errs, fmt.Sprintf("linha %d: data inválida %q", lineNum, dateStr))
+			continue
+		}
+
+		if m.DecimalSeparator == "," {
+			amtStr = strings.ReplaceAll(amtStr, ".", "")
+			amtStr = strings.ReplaceAll(amtStr, ",", ".")
+		}
+		amt, err := strconv.ParseFloat(amtStr, 64)
+		if err != nil {
+			errs = append(errs, fmt.Sprintf("linha %d: valor inválido %q", lineNum, amtStr))
+			continue
+		}
+
+		txType := "expense"
+		if amt > 0 {
+			txType = "income"
+		}
+
+		rows = append(rows, model.ImportRow{
+			Date:        t.Format("2006-01-02"),
+			Amount:      math.Abs(amt),
+			Description: desc,
+			Type:        txType,
+		})
+	}
+	return rows, errs
+}
+
+func max3(a, b, c int) int {
+	if b > a {
+		a = b
+	}
+	if c > a {
+		a = c
+	}
+	return a
+}
@@ -0,0 +1,114 @@
+package parser
+
+import (
+	"bufio"
+	"io"
+	"math"
+	"strconv"
+	"strings"
+	"time"
+
+	"financeiro-carvalho/internal/model"
+)
+
+// ParseOFX parses OFX SGML files as exported by Brazilian banks.
+// It is tolerant of the non-strict SGML format (no closing tags, BOM, etc.).
+func ParseOFX(r io.Reader) ([]model.ImportRow, error) {
+	scanner := bufio.NewScanner(r)
+	scanner.Buffer(make([]byte, 1024*1024), 1024*1024)
+
+	var rows []model.ImportRow
+	var cur *model.ImportRow
+
+	for scanner.Scan() {
+		line := strings.TrimSpace(scanner.Text())
+		line = strings.TrimPrefix(line, "\xef\xbb\xbf") // strip UTF-8 BOM if present
+
+		if line == "" {
+			continue
+		}
+
+		tag, value := splitOFXLine(line)
+
+		switch tag {
+		case "STMTTRN":
+			if cur != nil {
+				rows = append(rows, *cur)
+			}
+			cur = &model.ImportRow{}
+		case "/STMTTRN":
+			if cur != nil {
+				rows = append(rows, *cur)
+				cur = nil
+			}
+		case "DTPOSTED":
+			if cur != nil {
+				cur.Date = parseOFXDate(value)
+			}
+		case "TRNAMT":
+			if cur != nil {
+				amt, _ := strconv.ParseFloat(strings.ReplaceAll(value, ",", "."), 64)
+				cur.Amount = math.Abs(amt)
+				if amt < 0 {
+					cur.Type = "expense"
+				} else {
+					cur.Type = "income"
+				}
+			}
+		case "FITID":
+			if cur != nil {
+				cur.ExternalID = value
+			}
+		case "MEMO":
+			if cur != nil && cur.Description == "" {
+				cur.Description = cleanDescription(value)
+			}
+		case "NAME":
+			if cur != nil {
+				cur.Description = cleanDescription(value)
+			}
+		}
+	}
+	if cur != nil {
+		rows = append(rows, *cur)
+	}
+	return rows, scanner.Err()
+}
+
+// splitOFXLine handles both "<TAG>value" and "TAG:value" formats.
+func splitOFXLine(line string) (tag, value string) {
+	if strings.HasPrefix(line, "<") {
+		end := strings.Index(line, ">")
+		if end < 0 {
+			return "", ""
+		}
+		tag = strings.ToUpper(line[1:end])
+		value = strings.TrimSpace(line[end+1:])
+		return
+	}
+	// header lines like "OFXHEADER:100" — ignore
+	return "", ""
+}
+
+// parseOFXDate handles formats: 20060102, 20060102120000, 20060102120000[-3:BRT]
+func parseOFXDate(s string) string {
+	s = strings.SplitN(s, "[", 2)[0] // strip timezone suffix
+	s = strings.TrimSpace(s)
+	formats := []string{"20060102150405", "20060102"}
+	for _, f := range formats {
+		if len(s) >= len(f) {
+			t, err := time.Parse(f, s[:len(f)])
+			if err == nil {
+				return t.Format("2006-01-02")
+			}
+		}
+	}
+	return s
+}
+
+func cleanDescription(s string) string {
+	s = strings.TrimSpace(s)
+	// remove duplicate whitespace
+	fields := strings.Fields(s)
+	return strings.Join(fields, " ")
+}
@@ -0,0 +1,136 @@
+package parser_test
+
+import (
+	"strings"
+	"testing"
+
+	"financeiro-carvalho/internal/model"
+	"financeiro-carvalho/internal/parser"
+)
+
+const sampleOFX = `OFXHEADER:100
+DATA:OFXSGML
+VERSION:150
+
+<OFX>
+<BANKMSGSRSV1>
+<STMTTRNRS>
+<STMTRS>
+<BANKTRANLIST>
+<STMTTRN>
+<TRNTYPE>DEBIT
+<DTPOSTED>20240315120000[-3:BRT]
+<TRNAMT>-150.90
+<FITID>2024031500001
+<MEMO>PAGTO BOLETO ENERGIA
+</STMTTRN>
+<STMTTRN>
+<TRNTYPE>CREDIT
+<DTPOSTED>20240320
+<TRNAMT>5000.00
+<FITID>2024032000001
+<MEMO>SALARIO EMPRESA
+</STMTTRN>
+</BANKTRANLIST>
+`
+
+func TestParseOFX_BasicFields(t *testing.T) {
+	rows, err := parser.ParseOFX(strings.NewReader(sampleOFX))
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if len(rows) != 2 {
+		t.Fatalf("expected 2 rows, got %d", len(rows))
+	}
+
+	debit := rows[0]
+	if debit.Date != "2024-03-15" {
+		t.Errorf("date: want 2024-03-15, got %q", debit.Date)
+	}
+	if debit.Amount != 150.90 {
+		t.Errorf("amount: want 150.90, got %v", debit.Amount)
+	}
+	if debit.Type != "expense" {
+		t.Errorf("type: want expense, got %q", debit.Type)
+	}
+	if debit.ExternalID != "2024031500001" {
+		t.Errorf("external_id: got %q", debit.ExternalID)
+	}
+	if debit.Description != "PAGTO BOLETO ENERGIA" {
+		t.Errorf("desc: got %q", debit.Description)
+	}
+
+	credit := rows[1]
+	if credit.Type != "income" {
+		t.Errorf("type: want income, got %q", credit.Type)
+	}
+	if credit.Amount != 5000.00 {
+		t.Errorf("amount: want 5000.00, got %v", credit.Amount)
+	}
+}
+
+func TestParseOFX_EmptyFile(t *testing.T) {
+	rows, err := parser.ParseOFX(strings.NewReader(""))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(rows) != 0 {
+		t.Fatalf("expected 0 rows, got %d", len(rows))
+	}
+}
+
+func TestParseCSV_Semicolon(t *testing.T) {
+	// Brazilian bank CSV: semicolon separator, comma as decimal
+	data := "15/03/2024;-150,90;SUPERMERCADO\n20/03/2024;5000,00;SALARIO\n"
+	m := model.CSVMapping{
+		DateColumn:        0,
+		AmountColumn:      1,
+		DescriptionColumn: 2,
+		DateFormat:        "02/01/2006",
+		HasHeader:         false,
+		DecimalSeparator:  ",",
+		FieldSeparator:    ";",
+	}
+	rows, errs := parser.ParseCSV(strings.NewReader(data), m)
+	if len(errs) != 0 {
+		t.Fatalf("unexpected errors: %v", errs)
+	}
+	if len(rows) != 2 {
+		t.Fatalf("expected 2 rows, got %d", len(rows))
+	}
+	if rows[0].Date != "2024-03-15" {
+		t.Errorf("date: got %q", rows[0].Date)
+	}
+	if rows[0].Amount != 150.90 {
+		t.Errorf("amount: want 150.90, got %v", rows[0].Amount)
+	}
+	if rows[0].Type != "expense" {
+		t.Errorf("type: got %q", rows[0].Type)
+	}
+	if rows[1].Type != "income" {
+		t.Errorf("income type: got %q", rows[1].Type)
+	}
+}
+
+func TestParseCSV_WithHeader(t *testing.T) {
+	data := "Data;Valor;Descricao\n15/03/2024;-50,00;FARMACIA\n"
+	m := model.CSVMapping{
+		DateColumn:        0,
+		AmountColumn:      1,
+		DescriptionColumn: 2,
+		DateFormat:        "02/01/2006",
+		HasHeader:         true,
+		DecimalSeparator:  ",",
+		FieldSeparator:    ";",
+	}
+	rows, errs := parser.ParseCSV(strings.NewReader(data), m)
+	if len(errs) != 0 {
+		t.Fatalf("errors: %v", errs)
+	}
+	if len(rows) != 1 {
+		t.Fatalf("expected 1 row (header skipped), got %d", len(rows))
+	}
+	if rows[0].Description != "FARMACIA" {
+		t.Errorf("desc: got %q", rows[0].Description)
+	}
+}