onescience_run_manifest.yaml

onescience_manifest_version: "0.1"
resource_type: dataset
resource:
  id: OneScience/eagle
  name: eagle
  domain: cfd
  domain_tags: [cfd, eagle, turbulent_flow, unstructured_mesh]
  task: eagle_flow_forecasting_dataset
  task_tags: [train_data, eval_data, visualization_input]
  modalities: [unstructured_mesh, temporal_field]
  input_formats: [npz, npy]
  output_formats: []
  summary: Eagle 数据集标准包，包含 Cre/Spl/Tri 三类二维非定常流体仿真样本及 GraphViT 所需聚类文件。
  language: zh
  license: unknown

platform_resource:
  primary:
    platform: modelscope
    repo_id: OneScience/eagle
    repo_type: dataset
    url: https://modelscope.cn/datasets/OneScience/eagle
    revision: main
    readme_path: README.md
    manifest_path: onescience_run_manifest.yaml
  mirrors: []
  access:
    visibility: public
    license: null

website_integration:
  enabled: true
  click_target:
    platform: modelscope
    resource_url: https://modelscope.cn/datasets/OneScience/eagle
  llm_handoff:
    readme_required: true
    manifest_required: true
    download_readme_first: true
    resolve_related_models: true
    default_run_goal: minimal_validation
    default_workdir: .
    cwd_note: 如果使用 modelscope download --cache_dir 下载，请先 cd 到实际下载后的数据集仓库根目录，模型侧再把 ONESCIENCE_EAGLE_DATA_DIR 指向 data/Eagle_dataset。

runtime:
  enabled: true
  onescience_domain: cfd
  supported_execution: [local_dataset_validation, model_training_input, model_evaluation_input]
  environment:
    exported_env:
      ONESCIENCE_EAGLE_DATA_DIR: <dataset_repo_root>/data/Eagle_dataset
  dependencies:
    python_packages: [numpy, ruamel.yaml]

onescience:
  repo: https://gitee.com/onescience-ai/onescience
  official_links:
    gitee:
      doc: https://gitee.com/onescience-ai/onescience-doc
      onescience: https://gitee.com/onescience-ai/onescience
      skills: https://gitee.com/onescience-ai/oneskills
    github:
      doc: https://github.com/onescience-ai/OneScience-doc
      onescience: https://github.com/onescience-ai/OneScience
      skills: https://github.com/onescience-ai/oneskills
  install:
    required_by_default: false
    command: bash install.sh cfd
  compatibility:
    examples_path: onescience/examples/cfd/EagleMeshTransformer
    status: examples_compatible
  datapipe: onescience.datapipes.cfd.EagleDatapipe

runtime_package:
  kind: standard_runtime_package
  package_root: .
  standard_layout:
    workdir: .
    data_dir: data/Eagle_dataset
    metadata_dir: metadata
    output_dir: validation_outputs
  apply_policy:
    mode: direct_use
    target: session_workdir
    overwrite: false
    protect_installed_onescience: true
  entry_files:
    - data/Eagle_dataset
    - metadata/schema.yaml
    - metadata/train.txt
    - metadata/valid.txt
    - metadata/test.txt
    - files_sha256.jsonl
    - scripts/validate_eagle_dataset.py
  entrypoints:
    preflight: scripts/validate_eagle_dataset.py
    inference: null
    train: null
    finetune: null
    evaluate: null
    visualize: null
    deploy: null

files:
  dataset_files:
    - id: eagle_data_root
      path: data/Eagle_dataset
      role: train_eval_data
      description_zh: EAGLE 原始仿真和聚类数据根目录，包含 Cre/Spl/Tri。
      required: true
      required_for: [preflight, train, evaluate, visualize]
      source:
        platform: modelscope
        repo_id: OneScience/eagle
        repo_type: dataset
        path: data/Eagle_dataset
        revision: main
        download_method: command_ref
        command_ref: commands.download.download_dataset
      local_path: data/Eagle_dataset
    - id: eagle_schema
      path: metadata/schema.yaml
      role: schema
      description_zh: 描述 sim.npz、triangles.npy 和 constrained_kmeans 文件的字段、shape 和 dtype。
      required: true
      required_for: [preflight]
      source: {platform: modelscope, repo_id: OneScience/eagle, repo_type: dataset, path: metadata/schema.yaml, revision: main, download_method: command_ref, command_ref: commands.download.download_dataset}
      local_path: metadata/schema.yaml
    - id: eagle_sha256_inventory
      path: files_sha256.jsonl
      role: file_size_sha256_inventory
      description_zh: 每个数据文件的相对路径、大小和 SHA256，用于上传前和下载后完整性校验。
      required: true
      required_for: [preflight]
      source: {platform: modelscope, repo_id: OneScience/eagle, repo_type: dataset, path: files_sha256.jsonl, revision: main, download_method: command_ref, command_ref: commands.download.download_dataset}
      local_path: files_sha256.jsonl
  config_files: []
  sample_files:
    - id: eagle_split_metadata
      path: metadata/*.txt
      role: split_files
      description_zh: EagleMeshTransformer 使用的 train/valid/test 相对路径列表。
      required: true
      required_for: [train, evaluate]
      source: {platform: modelscope, repo_id: OneScience/eagle, repo_type: dataset, path: metadata/, revision: main, download_method: command_ref, command_ref: commands.download.download_dataset}
      local_path: metadata/

dataset:
  format: [npz, npy]
  sample_unit: Cre/Spl/Tri 下的一个 case/trajectory 目录
  schema:
    path: metadata/schema.yaml
  splits:
    train:
      path: metadata/train.txt
      count: 947
    validation:
      path: metadata/valid.txt
      count: 118
    test:
      path: metadata/test.txt
      count: 118
  size:
    sample_directories: 1200
    files: 7200
    geometry_directories:
      Cre: 400
      Spl: 400
      Tri: 400

relations:
  compatible_models:
    - id: OneScience/EagleMeshTransformer
      name: EagleMeshTransformer
      role: train_data
      required_for: [preflight, train, inference, evaluate, visualize]
      resource_ref:
        platform: modelscope
        repo_id: OneScience/EagleMeshTransformer
        repo_type: model
        url: https://modelscope.cn/models/OneScience/EagleMeshTransformer
        revision: main
        readme_path: README.md
        manifest_path: onescience_run_manifest.yaml
      expected_local_path: ${ONESCIENCE_EAGLE_DATA_DIR}
      expected_model_env: ONESCIENCE_EAGLE_DATA_DIR
      expected_model_config: conf/graphvit_eagle.yaml
  required_by_models:
    - resource_ref:
        platform: modelscope
        repo_id: OneScience/EagleMeshTransformer
        repo_type: model
        url: https://modelscope.cn/models/OneScience/EagleMeshTransformer
        revision: main
        readme_path: README.md
        manifest_path: onescience_run_manifest.yaml

run_matrix:
  scenarios:
    - name: validate_structure
      capability: dataset_preflight
      default: true
      required_datasets:
        - id: OneScience/eagle
          role: train_data
          local_path: data/Eagle_dataset
      required_model_files: []
      required_dataset_files:
        - data/Eagle_dataset/Cre|Spl|Tri/<case>/<traj>/sim.npz
        - data/Eagle_dataset/Cre|Spl|Tri/<case>/<traj>/triangles.npy
        - data/Eagle_dataset/Cre|Spl|Tri/<case>/<traj>/constrained_kmeans_10.npy
        - data/Eagle_dataset/Cre|Spl|Tri/<case>/<traj>/constrained_kmeans_20.npy
        - data/Eagle_dataset/Cre|Spl|Tri/<case>/<traj>/constrained_kmeans_30.npy
        - data/Eagle_dataset/Cre|Spl|Tri/<case>/<traj>/constrained_kmeans_40.npy
      preconditions: [当前工作目录为数据集仓库根目录。]
      command_refs: [commands.preflight.validate_dataset]
      outputs:
        - '控制台输出 "[OK] dataset validation completed"'
    - name: validate_integrity
      capability: dataset_integrity
      required_datasets:
        - id: OneScience/eagle
          role: train_data
          local_path: data/Eagle_dataset
      required_model_files: []
      required_dataset_files:
        - files_sha256.jsonl
        - "data/Eagle_dataset/**"
      preconditions:
        - 当前工作目录为数据集仓库根目录。
      command_refs: [commands.preflight.validate_dataset_full_hash]
      outputs:
        - '控制台输出 "[OK] checksum manifest verified in size+sha256 mode"'
    - name: provide_to_model
      capability: train
      required_datasets:
        - id: OneScience/eagle
          role: train_data
          local_path: data/Eagle_dataset
      required_model_files:
        - "OneScience/EagleMeshTransformer:conf/graphvit_eagle.yaml"
      required_dataset_files:
        - "data/Eagle_dataset/**/sim.npz"
        - "data/Eagle_dataset/**/triangles.npy"
        - "data/Eagle_dataset/**/constrained_kmeans_40.npy"
      preconditions:
        - 模型仓库已下载。
      command_refs: [commands.prepare.export_dataset_env]
      outputs:
        - 模型侧 scripts/preflight_check.py 可以通过数据目录检查。

capabilities:
  dataset_validation: true
  train_input: true
  inference_input: true
  evaluation_input: true
  visualization_input: true

commands:
  download:
    - name: download_dataset
      cwd: session_workdir
      command: modelscope download --dataset OneScience/eagle
    - name: download_model
      cwd: session_workdir
      command: modelscope download --model OneScience/EagleMeshTransformer
  prepare:
    - name: export_dataset_env
      cwd: .
      command: export ONESCIENCE_EAGLE_DATA_DIR=/path/to/OneScience_eagle/data/Eagle_dataset
  preflight:
    - name: validate_dataset
      cwd: .
      command: python scripts/validate_eagle_dataset.py
    - name: validate_dataset_full_hash
      cwd: .
      command: python scripts/validate_eagle_dataset.py --full-hash
  train: []
  inference: []
  evaluate: []
  visualize: []
  finetune: []
  deploy: []

expected_outputs:
  validate_dataset: ["[OK] dataset validation completed"]
  validate_dataset_full_hash: ["[OK] checksum manifest verified in size+sha256 mode: 7200 files"]
  model_consumption: [模型包 scripts/preflight_check.py 通过数据目录检查。]

diagnostics:
  - id: missing_data_root
    symptom: dataset root not found
    action: 确认当前目录为 OneScience/eagle 数据集仓库根目录，且 data/Eagle_dataset 存在。
  - id: missing_sample_file
    symptom: missing required file
    action: 对照 files_sha256.jsonl 重新上传或重新下载缺失样本目录。
  - id: checksum_mismatch
    symptom: sha256 mismatch
    action: 数据文件损坏或未完整上传，需重新同步对应文件。
  - id: schema_mismatch
    symptom: unexpected pointcloud shape 或 unexpected dtype
    action: 检查是否误用了非 Eagle v1 数据，或文件在解压/上传过程中被替换。

domain_extension:
  cfd:
    geometry_types: [Cre, Spl, Tri]
    mesh_type: unstructured_triangular_mesh
    variables: [pointcloud, mask, VX, VY, PS, PG]
    temporal_steps_per_sample: 990
    cluster_files: [constrained_kmeans_10.npy, constrained_kmeans_20.npy, constrained_kmeans_30.npy, constrained_kmeans_40.npy]