Skip to content

09-txt-conversion.md

Latest commit

 

History

History
113 lines (85 loc) · 3.49 KB

09-txt-conversion.md

File metadata and controls

113 lines (85 loc) · 3.49 KB

TXT-to-XML Conversion

Overview

The SPED TXT format is a legacy pipe-delimited representation of NF-e data used by older fiscal systems and the SEBRAE simplified layout. This module converts TXT to NF-e XML, enabling batch import from legacy software.

Files: convert.ts, txt-structures.ts, valid-txt.ts

TXT Format

Each line represents a data group, prefixed by a tag identifier and delimited by pipes:

A|4.00|NFe|
B|35|200209944445000162|55|101|000123456|1|89012345|3|
C|Empresa Teste|09944445000162|
...
H|1|
I|Produto 1|12345678|01234567890|5102|1|UN|10.50|10.50|
Q|01|10.50|1.6500|0.17|
S|01|10.50|7.6000|0.80|

Tag meanings

Tag XML Group Description
A NFe header Version, namespace
B ide Identification (UF, CNPJ, model, series, number)
C emit Issuer data
E dest Recipient data
H det (header) Item number
I prod Product data
N ICMS ICMS tax data
Q PIS PIS tax data
S COFINS COFINS tax data
W total Totals
X transp Transport
Y pag Payment
Z infAdic Additional info

Supported Layouts (txt-structures.ts)

Layout Description Fields
LOCAL Standard local layout Full field set
LOCAL_V12 Local v1.2 Slight field variations
LOCAL_V13 Local v1.3 Updated fields
SEBRAE Simplified SEBRAE layout Reduced field set

Each layout defines a Record<string, string> mapping tag → pipe-delimited field names:

const LOCAL: Record<string, string> = {
  "A": "versao|Id",
  "B": "cUF|cNF|natOp|mod|serie|nNF|dhEmi|...",
  "C": "xNome|CNPJ|...",
  // ...
};

Validation (valid-txt.ts)

function isValidTxt(content: string): { valid: boolean; errors: string[]; layout: string }

Validates:

  • First line must start with A| (version tag)
  • Must contain at least B| (identification)
  • Detects layout version from field count
  • Checks required tags are present

Converter (convert.ts)

Entry point

function convertTxtToXml(txtContent: string): string

Conversion flow

  1. Validate — Check TXT format and detect layout
  2. Parse lines — Split by |, map to field names using the layout structure
  3. Build XML — For each parsed group, generate the corresponding XML tags
  4. Assemble — Combine all groups into a complete <NFe> XML document

Parser class (NFeParser)

Internal class that maintains state across lines:

class NFeParser {
  private structure: Record<string, string>;  // layout definition
  private infNFeId: string;                    // access key
  private curProd: ParsedFields;               // current product being parsed
  // ... state for each XML group
}

The parser processes lines sequentially. When it encounters H| (item header), it starts a new product. When it encounters I|, N|, Q|, S|, it adds data to the current product.

Output

A valid unsigned NF-e XML document that can then be signed and submitted to SEFAZ, identical to what xml-builder.ts produces from structured data.

Use Cases

  1. Legacy migration — Import invoices from older SPED-compatible systems
  2. Batch processing — Convert multiple TXT files to XML for bulk authorization
  3. SEBRAE integration — Support the simplified layout used by small businesses via SEBRAE tools
  4. Testing — The PHP sped-nfe test suite uses TXT fixtures; this converter enables test porting