Statement PDFs in Spring Boot

Why statements become PDFs

A statement summarises a period that is now closed. Regenerating it later from live data would produce a different document, which defeats the point. That is the difference between a document and a view of a database, and it is why finance and account management keep asking for this.

What a statement has to carry

Before any of the code below matters, the template has to produce a document that is actually a statement. These are the fields that make it one, and the ones a reader or an auditor will look for first.

• the opening balance and the closing balance
• every transaction in the period with date, description and signed amount
• a running balance column, so a reader can find where a discrepancy starts
• the account identifier, usually masked
• the statement period, which must not overlap the previous one

Turning your Thymeleaf template into a document

You already have the markup. TemplateEngine.process("invoice", context) gives you it as a string, which is the only input the render needs. Thymeleaf's @{...} link syntax resolves against the servlet context path. In a rendered document that context does not exist, so build absolute URLs or inline the CSS.

The Spring Boot implementation

java.net.http.HttpClient ships with the runtime, so this adds no dependency. The render happens in a controller returning StreamingResponseBody, and you return ResponseEntity<StreamingResponseBody>.

// StatementPdfController.java
package com.example.statements;

import java.net.URI;
import java.net.http.HttpClient;
import java.net.http.HttpRequest;
import java.net.http.HttpResponse;
import java.time.Duration;

@RestController
public class StatementPdfController {

  private final HttpClient http = HttpClient.newHttpClient();
  private final TemplateEngine templates;

  @Value("${pdfpipe.key}")
  private String apiKey;

  @GetMapping(value = "/statements/{id}.pdf", produces = MediaType.APPLICATION_PDF_VALUE)
  public ResponseEntity<StreamingResponseBody> pdf(@PathVariable String id) throws Exception {
    Context context = new Context();
    context.setVariable("statement", statementService.find(id));
    String html = templates.process("statement", context);

    String body = new ObjectMapper().writeValueAsString(Map.of(
        "html", html,
        "options", Map.of("format", "A4", "printBackground", true)));

    HttpRequest request = HttpRequest.newBuilder(URI.create("https://api.pdfpipe.xyz/v1/pdf"))
        .header("Authorization", "Bearer " + apiKey)
        .header("Content-Type", "application/json")
        .timeout(Duration.ofSeconds(60))
        .POST(HttpRequest.BodyPublishers.ofString(body))
        .build();

    // ofInputStream, not ofByteArray. A long statement on the heap is how a
    // container sized for JSON payloads meets the garbage collector.
    HttpResponse<InputStream> upstream =
        http.send(request, HttpResponse.BodyHandlers.ofInputStream());

    return ResponseEntity.ok()
        .header("Content-Disposition", "attachment; filename=\"statement-" + id + ".pdf\"")
        .body(out -> upstream.body().transferTo(out));
  }
}

The CSS that makes a statement page correctly

The layout problem specific to this document is that a long transaction table spanning many pages, needing repeated headers and a running balance that stays legible after a break. These rules handle it.

/* Line items run past one page. Repeat the header, keep the
   totals block whole, and never strand a single row. */
thead { display: table-header-group; }
tfoot { display: table-footer-group; }
tr    { break-inside: avoid; }

.totals {
  break-inside: avoid;
  break-before: auto;
}

@page {
  size: A4;
  margin: 18mm 16mm 22mm;
}

What goes wrong in Spring Boot

RestTemplate returning byte[] materialises the whole document on the heap, and a servlet container sized for JSON will meet the garbage collector after a handful of concurrent report renders. Use the reactive client or the JDK client with ofInputStream.

What people try first

Most Spring Boot projects reach for Flying Saucer, which renders XHTML strictly and rejects markup a browser would accept, or iText, whose AGPL licence makes commercial use a paid decision. That works until it is running on more than one machine, at which point the browser becomes the thing you operate rather than the thing you use.

Where the statement lives afterwards

Rendering is the short part. Statements are the document most often requested again years later, usually for a loan application or an audit. Regenerating one from current data would produce different figures, so the rendered file is the record and has to be stored as issued.

Getting the document right

• Check the closing balance, which has to reconcile with the transactions listed above it.
• Generation is triggered by the end of a billing or accounting period, so size the timeout for that path rather than for a health check.
• This is a batch shape: render the run through the batch endpoint rather than firing thousands of individual requests.
• Backgrounds are painted by default here, so a design that uses colour needs nothing set. Only an explicit print_background of false turns them off.

Frequently asked

Related

Other Spring Boot documents, and the same statement in other stacks.