haystack/test/samples/pipeline/test.haystack-pipeline.yml

version: ignore

components:
  - name: Reader
    type: FARMReader
    params:
      no_ans_boost: -10
      model_name_or_path: deepset/roberta-base-squad2
      num_processes: 0
  - name: ESRetriever
    type: BM25Retriever
    params:
      document_store: DocumentStore
  - name: DocumentStore
    type: ElasticsearchDocumentStore
    params:
      index: haystack_test
      label_index: haystack_test_label
  - name: PDFConverter
    type: PDFToTextConverter
    params:
      remove_numeric_tables: false
  - name: TextConverter
    type: TextConverter
  - name: Preprocessor
    type: PreProcessor
    params:
      clean_whitespace: true
  - name: IndexTimeDocumentClassifier
    type: TransformersDocumentClassifier
    params:
      batch_size: 16
      use_gpu: false
  - name: QueryTimeDocumentClassifier
    type: TransformersDocumentClassifier
    params:
      use_gpu: false


pipelines:
  - name: query_pipeline
    nodes:
      - name: ESRetriever
        inputs: [Query]
      - name: Reader
        inputs: [ESRetriever]

  - name: query_pipeline_with_document_classifier
    nodes:
      - name: ESRetriever
        inputs: [Query]
      - name: QueryTimeDocumentClassifier
        inputs: [ESRetriever]
      - name: Reader
        inputs: [QueryTimeDocumentClassifier]

  - name: indexing_pipeline
    nodes:
      - name: PDFConverter
        inputs: [File]
      - name: Preprocessor
        inputs: [PDFConverter]
      - name: ESRetriever
        inputs: [Preprocessor]
      - name: DocumentStore
        inputs: [ESRetriever]

  - name: indexing_text_pipeline
    nodes:
      - name: TextConverter
        inputs: [File]
      - name: Preprocessor
        inputs: [TextConverter]
      - name: ESRetriever
        inputs: [Preprocessor]
      - name: DocumentStore
        inputs: [ESRetriever]

  - name: indexing_pipeline_with_classifier
    nodes:
      - name: PDFConverter
        inputs: [File]
      - name: Preprocessor
        inputs: [PDFConverter]
      - name: IndexTimeDocumentClassifier
        inputs: [Preprocessor]
      - name: ESRetriever
        inputs: [IndexTimeDocumentClassifier]
      - name: DocumentStore
        inputs: [ESRetriever]