stream method

@override Stream<String> stream(File file)

Extracts text content from a file and returns it as a stream of strings.

Implementing classes must define how to parse their specific file format. @param file The file to extract text from

Implementation

@override
Stream<String> stream(File file) async* {
  File sidecar = File("${file.path}.ocrmypdf.txt");
  File outputPDF = File("${file.path}.ocrmypdf.pdf");

  ProcessResult result = await Process.run("ocrmypdf", [
    "--sidecar",
    sidecar.path,
    //"--jobs", "1",
    //"--force-ocr",
    // "--invalidate-digital-signatures",
    "--output-type", "pdf",
    //"--optimize", "0",
    file.path,
    outputPDF.path,
  ]);

  if (result.stdout != null && result.stdout.isNotEmpty) {
    print("[OCRMyPDF]: ${result.stdout}");
  }
  if (result.stderr != null && result.stderr.isNotEmpty) {
    print("[OCRMyPDF] STDERR: ${result.stderr}");
  }

  if (!await sidecar.exists()) {
    info(
      "OCR Unnecessary for ${file.path}. Text already exists. Reading out sidecar...",
    );

    File of = outputPDF;
    if (!await outputPDF.exists()) {
      of = file;
    }

    PdfDocument document = PdfDocument(inputBytes: await of.readAsBytes());
    int pageCount = document.pages.count;
    IOSink sink = sidecar.openWrite();
    for (int i = 0; i < pageCount; i++) {
      sink.writeln(
        PdfTextExtractor(
          document,
        ).extractText(startPageIndex: i, layoutText: true),
      );
    }

    document.dispose();
    await sink.flush();
    await sink.close();
  } else {
    info(
      "OCR complete for ${file.path}. Text saved to ${sidecar.path} and PDF saved to ${outputPDF.path}. Feeding text into pandoc",
    );
  }

  yield* const TextFileStringer().stream(sidecar);

  try {
    await outputPDF.delete();
  } catch (e) {}
  try {
    await sidecar.delete();
  } catch (e) {}
}