MtoRagSystem/config/services.yaml

# ------------------------------------------------------------
# Parameters
# ------------------------------------------------------------

parameters:
  # ------------------------------------------------------------
  # Root
  # ------------------------------------------------------------

  mto.root: '%kernel.project_dir%'
  mto.kernel.dir: '%mto.root%'

  # ------------------------------------------------------------
  # Knowledge Root (ZENTRAL)
  # ------------------------------------------------------------

  mto.knowledge.root: '%mto.root%/var/knowledge'

  mto.knowledge.ndjson: '%mto.knowledge.root%/index.ndjson'
  mto.knowledge.index_meta: '%mto.knowledge.root%/index_meta.json'
  mto.knowledge.vector_index: '%mto.knowledge.root%/vector.index'
  mto.knowledge.vector_index_meta: '%mto.knowledge.root%/vector.index.meta.json'
  mto.runtime.meta: '%mto.knowledge.root%/index_runtime.json'
  mto.knowledge.upload: '%mto.knowledge.root%/uploads'

  # Backward compatibility alias
  mto.vector.data.upload.path: '%mto.knowledge.upload%'

  # ------------------------------------------------------------
  # Index Configuration (Fallback Guardrails)
  # ------------------------------------------------------------

  mto.index.chunk_size: 800
  mto.index.chunk_overlap: 100
  mto.index.embedding_model: 'all-MiniLM-L6-v2'
  mto.index.embedding_dimension: 768
  mto.index.scoring_version: 1

  # ------------------------------------------------------------
  # Python / Vector Runtime
  # ------------------------------------------------------------

  mto.vector.python_bin: '/var/www/html/.venv/bin/python3'
  mto.vector.ingest_script: '%mto.root%/src/Vector/vector_ingest.py'
  mto.vector.search_script: '%mto.root%/src/Vector/vector_search.py'
  mto.vector.timeout: 600


# ------------------------------------------------------------
# Services
# ------------------------------------------------------------

services:

  _defaults:
    autowire: true
    autoconfigure: true
    bind:
      Psr\Log\LoggerInterface $agentLogger: '@monolog.logger.agent'
      string $projectDir: '%mto.root%'

  App\:
    resource: '../src/'
    exclude:
      - '../src/DependencyInjection/'
      - '../src/Entity/'
      - '../src/Kernel.php'

  App\Controller\:
    resource: '../src/Controller/'
    tags: [ 'controller.service_arguments' ]

  # ------------------------------------------------------------
  # AI Agent – Infrastructure
  # ------------------------------------------------------------

  App\Infrastructure\OllamaClient:
    arguments:
      $apiUrl: '%env(AI_LLM_API_URL)%'
      $model: '%env(AI_LLM_MODEL)%'
      $timeoutSeconds: 600
      $configProvider: '@App\Service\ModelGenerationConfigProvider'

  # ------------------------------------------------------------
  # AI Agent – Context & Runner
  # ------------------------------------------------------------

  App\Context\ContextService:
    arguments:
      $historyDir: '%env(AI_HISTORY_DIR)%'
      $projectDir: '%mto.root%'

  App\Agent\AgentRunner:
    arguments:
      $debug: '%env(bool:AI_DEBUG)%'
      $logPrompt: '%env(bool:AI_LOG_PROMPT)%'
      $logContext: '%env(bool:AI_LOG_CONTEXT)%'

  # ------------------------------------------------------------
  # NDJSON Retrieval Stack
  # ------------------------------------------------------------

  App\Knowledge\Retrieval\NdjsonChunkLookup: ~

  App\Knowledge\Retrieval\NdjsonKeywordSearch: ~

  App\Knowledge\Retrieval\NdjsonHybridRetriever:
    arguments:
      $maxChunks: 3
      $vectorTopK: 5

  App\Knowledge\Retrieval\CachedRetriever:
    arguments:
      $inner: '@App\Knowledge\Retrieval\NdjsonHybridRetriever'
      $cache: '@cache.app'
      $ttlSeconds: 600

  App\Knowledge\Retrieval\RetrieverInterface:
    alias: App\Knowledge\Retrieval\CachedRetriever

  # ------------------------------------------------------------
  # Index Configuration Provider (DB + Fallback)
  # ------------------------------------------------------------

  App\Index\IndexConfigurationProvider:
    arguments:
      $repository: '@App\Repository\IngestProfileRepository'
      $fallbackChunkSize: '%mto.index.chunk_size%'
      $fallbackChunkOverlap: '%mto.index.chunk_overlap%'
      $fallbackEmbeddingModel: '%mto.index.embedding_model%'
      $fallbackEmbeddingDimension: '%mto.index.embedding_dimension%'
      $fallbackScoringVersion: '%mto.index.scoring_version%'

  # ------------------------------------------------------------
  # Index Meta Manager (uses Provider)
  # ------------------------------------------------------------

  App\Index\IndexMetaManager:
    arguments:
      $metaPath: '%mto.knowledge.index_meta%'
      $runTimePath: '%mto.runtime.meta%'
      $provider: '@App\Index\IndexConfigurationProvider'

  # ------------------------------------------------------------
  # Vector Layer
  # ------------------------------------------------------------

  App\Vector\VectorSearchClient:
    arguments:
      $pythonBin: '%mto.vector.python_bin%'
      $scriptPath: '%mto.vector.search_script%'
      $vectorIndexPath: '%mto.knowledge.vector_index%'
      $vectorMetaPath: '%mto.knowledge.vector_index_meta%'
      $indexMetaPath: '%mto.knowledge.index_meta%'
      $agentLogger: '@monolog.logger.agent'

  App\Vector\VectorIndexBuilder:
    arguments:
      $pythonBin: '%mto.vector.python_bin%'
      $scriptPath: '%mto.vector.ingest_script%'
      $indexNdjsonPath: '%mto.knowledge.ndjson%'
      $indexMetaPath: '%mto.knowledge.index_meta%'
      $vectorIndexPath: '%mto.knowledge.vector_index%'
      $timeoutSeconds: '%mto.vector.timeout%'
      $configurationProvider: '@App\Index\IndexConfigurationProvider'

  # ------------------------------------------------------------
  # Admin Utilities
  # ------------------------------------------------------------

  App\Service\Admin\IndexNdjsonInspector:
    arguments:
      $ndJsonPath: '%mto.knowledge.ndjson%'
      $indexMetaPath: '%mto.knowledge.index_meta%'