# 🤖 PIPELINE DEFINITION
nlp:
  "@core": pipeline

  lang: eds

  components:
    normalizer:
      '@factory': eds.normalizer

    sentencizer:
      '@factory': eds.sentences

    covid:
      '@factory': eds.covid

    qualifier:
      '@factory': eds.span_classifier
      attributes: { "_.negation": [ "sosy" ] }
      span_getter: ["ents", "gold_spans"]
      context_getter: { '@misc': eds.span_context_getter, "context_words": 30, "context_sents": 1 }

      embedding:
        '@factory': eds.span_pooler

        embedding:
          '@factory': eds.text_cnn
          kernel_sizes: [ 3 ]

          embedding:
            '@factory': eds.transformer
            model: hf-internal-testing/tiny-bert
            window: 128
            stride: 96

# 📈 SCORERS
scorer:
  speed: true
  qual:
    '@metrics': "eds.span_attributes"
    span_getter: ${nlp.components.qualifier.span_getter}
    qualifiers: ${nlp.components.qualifier.attributes}

# 🎛️ OPTIMIZER
# (disabled to test the default optimizer)
# optimizer:
#   "@optimizers": adam
#   groups:
#     "*.transformer.*":
#       lr: 1e-3
#       schedules:
#         "@schedules": linear
#         "warmup_rate": 0.1
#         "start_value": 0
#     "*":
#       lr: 1e-3
#       schedules:
#         "@schedules": linear
#         "warmup_rate": 0.1
#         "start_value": 1e-3

# 📚 DATA
train_data:
  data:
    "@readers": json
    path: ./dataset.jsonl
    converter:
      - '@factory': 'myproject.custom_dict2doc'
        span_setter : 'gold_spans'
        span_attributes : ['negation']
        bool_attributes : ['negation']  # default json to doc converter
      - '@factory': eds.sentences
        nlp: ${nlp}
      # - '@factory': eds.split
      #   nlp: null
      #   max_length: 10
      #   randomize: 0.3
      #   # sentence regex:
      #   regex: '\\s*(?:\\n\\s*)+()[A-Z]|[.!?]\\s+()[A-Z]'
        # regex: '\\n{2,}'
  shuffle: dataset
  batch_size: 4 docs
  pipe_names: [ "qualifier" ]
  sub_batch_size: 10 words

val_data:
  "@readers": json
  path: ./dataset.jsonl
  converter:
    - '@factory': myproject.custom_dict2doc
      span_setter : 'gold_spans'
      span_attributes : ['negation']
      bool_attributes : ['negation']  # default standoff to doc converter

# 🚀 TRAIN SCRIPT OPTIONS
train:
  nlp: ${ nlp }
  train_data: ${ train_data }
  val_data: ${ val_data }
  max_steps: 40
  validation_interval: 10
  max_grad_norm: 1.0
  scorer: ${ scorer }
  num_workers: 1