first commit

2026-02-11 14:15:08 +01:00
parent a4742c2c38
commit aa7d362bc3
58 changed files with 9999 additions and 0 deletions
--- a/src/Command/KnowledgeIngestCommand.php
+++ b/src/Command/KnowledgeIngestCommand.php
@@ -0,0 +1,116 @@
+<?php
+// src/Command/KnowledgeIngestCommand.php
+
+declare(strict_types=1);
+
+namespace App\Command;
+
+use App\Knowledge\Ingest\KnowledgeIngestService;
+use Symfony\Component\Console\Attribute\AsCommand;
+use Symfony\Component\Console\Command\Command;
+use Symfony\Component\Console\Input\InputArgument;
+use Symfony\Component\Console\Input\InputInterface;
+use Symfony\Component\Console\Input\InputOption;
+use Symfony\Component\Console\Output\OutputInterface;
+use Symfony\Component\Finder\Finder;
+
+#[AsCommand(
+    name: 'mto:agent:knowledge:ingest',
+    description: 'Ingest one or multiple markdown/text documents into file-based knowledge chunks'
+)]
+final class KnowledgeIngestCommand extends Command
+{
+    public function __construct(
+        private readonly KnowledgeIngestService $ingest,
+        private readonly string $uploadsDir,
+    ) {
+        parent::__construct();
+    }
+
+    protected function configure(): void
+    {
+        $this
+            ->addArgument(
+                'file',
+                InputArgument::OPTIONAL,
+                'Path to a single .txt/.md file'
+            )
+            ->addOption(
+                'all',
+                null,
+                InputOption::VALUE_NONE,
+                'Ingest all .md files from the uploads directory'
+            )
+            ->addOption(
+                'optimize',
+                'o',
+                InputOption::VALUE_NONE,
+                'Optimize chunks for retrieval quality'
+            );
+    }
+
+    protected function execute(InputInterface $input, OutputInterface $output): int
+    {
+        $files = [];
+        $optimize = (bool) $input->getOption('optimize');
+
+        if ($input->getOption('all')) {
+            if (!is_dir($this->uploadsDir)) {
+                $output->writeln('<error>❌ uploads directory not found</error>');
+                return Command::FAILURE;
+            }
+
+            $finder = new Finder();
+            $finder
+                ->files()
+                ->in($this->uploadsDir)
+                ->name('*.md');
+
+            if (!$finder->hasResults()) {
+                $output->writeln('<comment>ℹ️ No .md files found in uploads/</comment>');
+                return Command::SUCCESS;
+            }
+
+            foreach ($finder as $file) {
+                $files[] = $file->getRealPath();
+            }
+
+            $output->writeln(sprintf(
+                '📂 Ingesting %d markdown files from uploads (%s)',
+                count($files),
+                $optimize ? 'optimized' : 'standard'
+            ));
+        } else {
+            $file = $input->getArgument('file');
+
+            if (!$file) {
+                $output->writeln('<error>❌ Either provide a file or use --all</error>');
+                return Command::FAILURE;
+            }
+
+            $files[] = (string) $file;
+        }
+
+        $totalWritten = 0;
+
+        foreach ($files as $filePath) {
+            $output->writeln('➡️  Ingesting: ' . $filePath);
+
+            $written = $this->ingest->ingestFile(
+                $filePath,
+                optimize: $optimize
+            );
+
+            $totalWritten += count($written);
+
+            foreach ($written as $chunk) {
+                $output->writeln('   - ' . $chunk);
+            }
+        }
+
+        $output->writeln('');
+        $output->writeln('✅ Total written chunks: ' . $totalWritten);
+
+        return Command::SUCCESS;
+    }
+}