operators:
- id: hf_source_positive_seed
  config:
    type: hf_source
    dataset: open-r1/codeforces
    split: train

- id: rename_task_column_positive_seed
  config:
    type: function
    function: data_strategies.commons.force_rename_columns
    function_config:
      column_maps:
        description: instruction_seed
  input_ids:
  -  hf_source_positive_seed

- id: remove_nones_positive
  config:
    type: function
    function: data_strategies.commons.remove_nones
    function_config:
      column: instruction_seed
  input_ids:
  - rename_task_column_positive_seed

- id: decontaminate_positive
  config:
    type: cpu_function
    sharded: false
    num_cpus: 32
    function:  data_strategies._A1_Source.utils.decontaminate_fuzzy_w_ngram
    function_config:
      column: instruction_seed
      eval_datasets:
        - HuggingFaceH4/MATH-500
        - Maxwell-Jia/AIME_2024
        - AI-MO/aimo-validation-amc
        - livecodebench/code_generation_lite
        - mlfoundations-dev/AIME2025_combined
        - cais/hle
        - open-r1/codeforces
        - Idavidrein/gpqa
        - daman1209arora/jeebench
        - mlfoundations-dev/mmlu_pro_eval_full
        - Qwen/CodeElo
        - open-r1/ioi
      eval_columns:
        - problem
        - Problem
        - problem
        - question_content
        - question
        - question
        - description
        - Question
        - question
        - prompt
        - description
        - statement
      eval_splits:
        - test
        - train
        - train
        - test
        - train
        - test
        - test
        - train
        - test
        - test
        - test
        - test
      eval_subsets:
        Idavidrein/gpqa: gpqa_diamond
      similarity_threshold: 75.0
      ngram_size: 13
  input_ids:
  - remove_nones_positive

- id: hf_source_negative_seed
  config:
    type: load_preexisting
    framework_name: a1_code_sql_create_context

- id: decontaminate_negative
  config:
    type: cpu_function
    sharded: false
    num_cpus: 32
    function:  data_strategies._A1_Source.utils.decontaminate_fuzzy_w_ngram
    function_config:
      column: instruction_seed
      eval_datasets:
        - HuggingFaceH4/MATH-500
        - Maxwell-Jia/AIME_2024
        - AI-MO/aimo-validation-amc
        - livecodebench/code_generation_lite
        - mlfoundations-dev/AIME2025_combined
        - cais/hle
        - open-r1/codeforces
        - Idavidrein/gpqa
        - daman1209arora/jeebench
        - mlfoundations-dev/mmlu_pro_eval_full
        - Qwen/CodeElo
        - open-r1/ioi
      eval_columns:
        - problem
        - Problem
        - problem
        - question_content
        - question
        - question
        - description
        - Question
        - question
        - prompt
        - description
        - statement
      eval_splits:
        - test
        - train
        - train
        - test
        - train
        - test
        - test
        - train
        - test
        - test
        - test
        - test
      eval_subsets:
        Idavidrein/gpqa: gpqa_diamond
      similarity_threshold: 75.0
      ngram_size: 13
  input_ids:
  -  hf_source_negative_seed

- id: train_fasttext_classifier
  config:
    type: train_fasttext
    text_column: instruction_seed
    positive_input_ids:
      - decontaminate_positive
    negative_input_ids:
      - decontaminate_negative
    dim: 256
    epoch: 3
    lr: 0.1
    word_ngrams: 2
    min_count: 3
    hf_repo_id: mlfoundations-dev/instruction_filtering_fast_text_classifier_b2_code_pos_codeforces_neg_sql
  input_ids:
    - decontaminate_positive
    - decontaminate_negative