Php R Bundle Laravel Package

Getting Started

1. Install R on the host (not just a Composer package). Verify the web/PHP user can execute it:

   which R
   R --version
   sudo -u www-data R --version
   

2. Install the bundle (this is a Symfony 2/3-era package):

   composer require kachkaev/php-r-bundle
   

   It depends on kachkaev/php-r, which provides the actual RClient.
3. Register the bundle in app/AppKernel.php:

   new Kachkaev\PHPRBundle\KachkaevPHPRBundle(),
   

4. Configure the R binary path in app/config/config.yml:

   kachkaev_php_r:
       r_binary: /usr/bin/R
   

   Check the bundle's DependencyInjection/Configuration.php and Resources/config/services.yml for the exact config keys and service ID — the bundle exposes the client as a service (commonly aliased to kachkaev_php_r or the RClient class).
5. Look first at Resources/config/services.yml to confirm the service ID and whether it is public. Then inject it:

   use Kachkaev\PHPR\RClient;
   
   class ReportController
   {
       public function summaryAction()
       {
           /** @var RClient $r */
           $r = $this->get('kachkaev_php_r'); // or type-hint RClient if autowiring is enabled
   
           $result = $r->run('mean(c(1, 2, 3, 4))');
   
           return new Response($result->getOutput());
       }
   }
   

6. First real use case: move an existing analytical calculation (averages, regressions, forecasting, chart generation) out of PHP and into an R script, then call it from a Symfony service/controller and render or persist the result.

Implementation Patterns

• One client per request, many R statements. The client wraps an R process/session. Assign variables and run multiple statements on the same client instead of spawning a new process per expression:

  $r->run('values <- c(10, 20, 30, 40)');
  $r->run('avg <- mean(values)');
  echo $r->run('avg')->getOutput();
  

• Use JSON/CSV as the PHP↔R contract. Text output from R includes [1] prefixes and formatting noise. Prefer jsonlite in R and json_decode in PHP, or write/read a temp CSV:

  $script = <<<'R'
  library(jsonlite)
  toJSON(list(mean = mean(values), sd = sd(values)), auto_unbox = TRUE)
  R;
  $data = json_decode($r->run($script)->getOutput(), true);
  

• Pass data via files or escaped variables, never raw interpolation. For user-supplied values, write to a temp file and read.csv()/readRDS() inside R, or carefully quote/escape values before assigning R variables.
• Wrap the client in your own service. Keep RClient out of controllers; create e.g. AppBundle\Analytics\RStatsService that owns script generation, error handling, and result parsing. This makes testing and later replacement much easier.
• Use R for file-producing tasks. Charts are a common fit: call png('/tmp/chart.png'), build the plot, dev.off(), then serve/attach the file from PHP.
• Handle R failures explicitly. Check output, stderr, and exceptions around each run() call. Log the generated script and R output so failures are reproducible.
• Test by stubbing the R service. Unit-test your wrapper with a fake implementation; keep a small integration test suite that requires R installed in CI.

Gotchas and Tips

• Archived and old. Last release was 2017, zero dependents, and it targets Symfony 2/3 and PHP 5/7-era APIs. Expect incompatibilities with Symfony 4+ and PHP 8. Pin versions, or fork/copy the tiny service definition into your own project.
• Requires proc_open/process execution. Many hosts disable proc_open, restrict open_basedir, or run PHP-FPM as a user without access to /usr/bin/R. Test the exact runtime user, not your shell user.
• Performance is process-bound. Every request that touches R pays process startup cost. Cache results, push heavy work to a queue/worker, or use a long-lived R server (Rserve) instead of CLI invocation for high traffic.
• Security: treat R script generation like SQL injection. R can call system(), read files, and execute shell commands. Never concatenate raw request data into a script. Whitelist operations and pass parameters through files or strictly escaped variables.
• Output parsing is fragile. getOutput() returns R's printed representation, not typed PHP data. Normalize whitespace and [1] prefixes, and prefer JSON for anything non-trivial.
• Path differences. Linux/macOS use /usr/bin/R; Windows uses Rscript.exe. Make the binary path environment-specific config, not hard-coded.
• Debugging workflow: log the generated R script, run it manually with Rscript script.R, and compare stderr/exit codes. Most failures are missing R packages, wrong working directory, or permissions.
• Migration tip: because the bundle is thin, replacing it with a symfony/process wrapper (run Rscript --vanilla script.R) is usually straightforward and removes a dead dependency.