Checking text in a PDF file with LoadRunner

A friend was performance testing a new SAP system that had a self-service portal where employees could view and download a PDF version of their payslip. He asked me how he could add a check to his LoadRunner script to make sure that the information in the PDF was correct (employee name, pay grade, etc). Manually checking all the PDFs at the end of the run was not a scalable solution (and just checking a sample of them would have missed any intermittent errors).

Fortunately, it was a simple matter to write some code in VuGen that would solve his problem.

My solution uses the popen() function to call a command-line program from LoadRunner which converts the PDF to text. The text can then be verified using simple string comparison functions. The command line program I used is pdftotext, which is part of the open-source Xpdf project. I have mirrored the file, so you can download pdftotext (from Xpdf v3.03) directly from my website.

Here is the example code:

char *strstr( const char *string1, const char *string2); // Explicit declaration required for C functions that do not return integers.
#define BUFFER_SIZE 16384 // 16 KB. If this gets too big, use malloc() instead to avoid "too many local variables" compilation error.

Action()
{
    int pdf_size; // size of the downloaded PDF (determined at runtime)
    int fp_save; // file pointer for saving the PDF
    int fp_popen; // file/stream pointer for the output from popen()
    int count; // number of characters that have been read from the popen() stream
    char buffer[BUFFER_SIZE]; // allocate memory for the output of popen()
	
    lr_start_transaction("download_pdf");

    // Save entire HTTP response body
    // Note that the PDF download must be in a step by itself in order to save the contents. It cannot be an EXTRARES.
    web_reg_save_param_ex(
        "ParamName=PdfContents",
        "LB/BIN=",
        "RB/BIN=",
        SEARCH_FILTERS,
        "Scope=Body",
        LAST);

    web_reg_find("Text=PDF", LAST); // Basic check
    web_url("Download PDF", 
        "URL=https://www.myloadtest.com/wp-content/uploads/migrated-resources/hp-discover-2012-tb2337-moncrieff.pdf", 
        "Resource=0", // Not a resource 
        "RecContentType=application/pdf", 
        LAST);	

    pdf_size = web_get_int_property(HTTP_INFO_DOWNLOAD_SIZE);
    if (pdf_size < 100000){
        lr_error_message("Downloaded PDF is smaller than expected."); // By itself, this is not a very good check.
        lr_exit(LR_EXIT_ACTION_AND_CONTINUE, LR_FAIL);
    }
	
    lr_end_transaction("download_pdf", LR_AUTO);
    lr_think_time(10);
    lr_start_transaction("check_pdf");
	
    // Save file.
    lr_save_string(lr_eval_string("C:\\LoadRunner\\Data\\{TimeStamp}{VuserId}.pdf"), "FileName"); // Ensure that the file name is unique to avoid overwriting the contents
    fp_save = fopen(lr_eval_string("{FileName}"), "wb"); // open file in "write, binary" mode.
    fwrite(lr_eval_string("{PdfContents}"), pdf_size, 1, fp_save);
    fclose(fp_save);
	
    // Run "C:\LoadRunner\Lib\pdftotext {FileName} -" and save the output to a parameter.
    fp_popen = popen(lr_eval_string("C:\\LoadRunner\\Lib\\pdftotext {FileName} -"), "r");
    count = fread(buffer, sizeof(char), BUFFER_SIZE, fp_popen);
    if (feof(fp_popen) == 0) {
        lr_error_message("Did not reach the end of the input stream when reading. Try increasing BUFFER_SIZE.");
        return -1;
    }
    buffer[count] = NULL; // Make sure the string in the buffer is null-terminated.
    lr_save_string(buffer, "PdfText");
    lr_output_message("The Windows version is: %s", buffer);
    pclose(fp_popen);
 
    // Check that the PDF parameter contains the expected text ("LoadRunner").
    if (strstr(lr_eval_string("{PdfText}"), "LoadRunner") == NULL) {
        lr_error_message("PDF file does not contain the expected string.");
        lr_exit(LR_EXIT_ACTION_AND_CONTINUE, LR_FAIL);
    }
	
    // Cleanup. Delete the file.
    system(lr_eval_string("del /q {FileName}"));
	
    lr_end_transaction("check_pdf", LR_AUTO);
	
    return 0;
}

Two caveats:

1. The code above has no error checking/handling. It is very quick and dirty.
2. pdftotext is a text extraction utility, not an OCR program, so it cannot extract text that is actually an image (like a scanned document).