Upload folder using huggingface_hub

.gitattributes

@@ -1,60 +1,97 @@
 *.7z filter=lfs diff=lfs merge=lfs -text
 *.arrow filter=lfs diff=lfs merge=lfs -text
-*.avro filter=lfs diff=lfs merge=lfs -text
 *.bin filter=lfs diff=lfs merge=lfs -text
+*.bin.* filter=lfs diff=lfs merge=lfs -text
 *.bz2 filter=lfs diff=lfs merge=lfs -text
-*.ckpt filter=lfs diff=lfs merge=lfs -text
 *.ftz filter=lfs diff=lfs merge=lfs -text
 *.gz filter=lfs diff=lfs merge=lfs -text
 *.h5 filter=lfs diff=lfs merge=lfs -text
 *.joblib filter=lfs diff=lfs merge=lfs -text
 *.lfs.* filter=lfs diff=lfs merge=lfs -text
-*.lz4 filter=lfs diff=lfs merge=lfs -text
-*.mds filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
 *.model filter=lfs diff=lfs merge=lfs -text
 *.msgpack filter=lfs diff=lfs merge=lfs -text
-*.npy filter=lfs diff=lfs merge=lfs -text
-*.npz filter=lfs diff=lfs merge=lfs -text
 *.onnx filter=lfs diff=lfs merge=lfs -text
 *.ot filter=lfs diff=lfs merge=lfs -text
 *.parquet filter=lfs diff=lfs merge=lfs -text
 *.pb filter=lfs diff=lfs merge=lfs -text
-*.pickle filter=lfs diff=lfs merge=lfs -text
-*.pkl filter=lfs diff=lfs merge=lfs -text
 *.pt filter=lfs diff=lfs merge=lfs -text
 *.pth filter=lfs diff=lfs merge=lfs -text
 *.rar filter=lfs diff=lfs merge=lfs -text
-*.safetensors filter=lfs diff=lfs merge=lfs -text
 saved_model/**/* filter=lfs diff=lfs merge=lfs -text
-*.tar.* filter=lfs diff=lfs merge=lfs -text
 *.tar filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.mat filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.hdf5 filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
 *.tflite filter=lfs diff=lfs merge=lfs -text
 *.tgz filter=lfs diff=lfs merge=lfs -text
-*.wasm filter=lfs diff=lfs merge=lfs -text
 *.xz filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
-*.zst filter=lfs diff=lfs merge=lfs -text
-*tfevents* filter=lfs diff=lfs merge=lfs -text
-# Audio files - uncompressed
-*.pcm filter=lfs diff=lfs merge=lfs -text
-*.sam filter=lfs diff=lfs merge=lfs -text
-*.raw filter=lfs diff=lfs merge=lfs -text
-# Audio files - compressed
-*.aac filter=lfs diff=lfs merge=lfs -text
-*.flac filter=lfs diff=lfs merge=lfs -text
-*.mp3 filter=lfs diff=lfs merge=lfs -text
-*.ogg filter=lfs diff=lfs merge=lfs -text
-*.wav filter=lfs diff=lfs merge=lfs -text
-# Image files - uncompressed
-*.bmp filter=lfs diff=lfs merge=lfs -text
-*.gif filter=lfs diff=lfs merge=lfs -text
-*.png filter=lfs diff=lfs merge=lfs -text
-*.tiff filter=lfs diff=lfs merge=lfs -text
-# Image files - compressed
+*.zstandard filter=lfs diff=lfs merge=lfs -text
+*.tfevents* filter=lfs diff=lfs merge=lfs -text
+*.db* filter=lfs diff=lfs merge=lfs -text
+*.ark* filter=lfs diff=lfs merge=lfs -text
+**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text
+**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text
+**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
 *.jpg filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
 *.jpeg filter=lfs diff=lfs merge=lfs -text
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
 *.webp filter=lfs diff=lfs merge=lfs -text
-# Video files - compressed
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+*.wma filter=lfs diff=lfs merge=lfs -text
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.m4a filter=lfs diff=lfs merge=lfs -text
+*.m3u8 filter=lfs diff=lfs merge=lfs -text
+*.amr filter=lfs diff=lfs merge=lfs -text
+*.audio filter=lfs diff=lfs merge=lfs -text
+*.avi filter=lfs diff=lfs merge=lfs -text
+*.flv filter=lfs diff=lfs merge=lfs -text
 *.mp4 filter=lfs diff=lfs merge=lfs -text
-*.webm filter=lfs diff=lfs merge=lfs -text
+*.mpg filter=lfs diff=lfs merge=lfs -text
+*.asf filter=lfs diff=lfs merge=lfs -text
+*.mov filter=lfs diff=lfs merge=lfs -text
+*.mpeg filter=lfs diff=lfs merge=lfs -text
+*.3gp filter=lfs diff=lfs merge=lfs -text
+*.wmv filter=lfs diff=lfs merge=lfs -text
+*.rmvb filter=lfs diff=lfs merge=lfs -text
+*.rm filter=lfs diff=lfs merge=lfs -text
+*.ts filter=lfs diff=lfs merge=lfs -text
+*.mkv filter=lfs diff=lfs merge=lfs -text
+*.flash filter=lfs diff=lfs merge=lfs -text
+*.vob filter=lfs diff=lfs merge=lfs -text
+*.pdf filter=lfs diff=lfs merge=lfs -text
+*.ost filter=lfs diff=lfs merge=lfs -text
+*.pst filter=lfs diff=lfs merge=lfs -text
+*.doc filter=lfs diff=lfs merge=lfs -text
+*.docx filter=lfs diff=lfs merge=lfs -text
+*.txt filter=lfs diff=lfs merge=lfs -text
+*.ppt filter=lfs diff=lfs merge=lfs -text
+*.pptx filter=lfs diff=lfs merge=lfs -text
+*.xls filter=lfs diff=lfs merge=lfs -text
+*.xlsx filter=lfs diff=lfs merge=lfs -text
+*.vsd filter=lfs diff=lfs merge=lfs -text
+*.vsdx filter=lfs diff=lfs merge=lfs -text
+ 
+ 
+dataset_infos.json ignore
+*.csv filter=lfs diff=lfs merge=lfs -text
+*.tsv filter=lfs diff=lfs merge=lfs -text
+
+data/cylinder_flow/test.tfrecord filter=lfs diff=lfs merge=lfs -text
+
+data/cylinder_flow/train.tfrecord filter=lfs diff=lfs merge=lfs -text
+
+data/cylinder_flow/valid.tfrecord filter=lfs diff=lfs merge=lfs -text
+
+files_sha256.jsonl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,86 @@
+---
+license: other
+#User-Defined Tags
+tags:
+  - CylinderFlow
+  - computational fluid dynamics
+  - incompressible flow
+language:
+  - en
+  - zh
+---
+<p align="center">
+  <strong>
+    <span style="font-size: 30px;">Cylinder Flow</span>
+  </strong>
+</p>
+
+## Dataset Description
+
+The Cylinder Flow dataset is sourced from DeepMind's MeshGraphNets benchmark and describes flow around a cylinder (vortex shedding) on a two-dimensional unstructured triangular mesh. The data was generated using COMSOL simulations. Each trajectory contains 600 time steps with a time-step size of `dt=0.01` and records the fluid velocity and pressure fields.
+
+Paper: [Learning Mesh-Based Simulation with Graph Networks](https://arxiv.org/abs/2010.03409)
+
+## Supported Tasks
+
+| Scenario | Description |
+|---|---|
+| Flow field time-series prediction | Predict subsequent velocity and pressure fields from historical mesh states. |
+| Vortex shedding simulation | Learn the evolution of transient flow around a cylinder. |
+| Graph neural network simulation | Provide training and evaluation data for mesh-based models such as MeshGraphNet. |
+| Mesh-based physics modeling | Learn physical relationships between nodes and edges on an unstructured triangular mesh. |
+
+## Dataset Format and Structure
+
+The data uses the TFRecord format and is divided into training, validation, and test sets:
+
+```text
+data/cylinder_flow/
+  train.tfrecord
+  valid.tfrecord
+  test.tfrecord
+  meta.json
+  stats/
+```
+
+Each record corresponds to an unstructured mesh trajectory. The main fields are as follows:
+
+| Field | shape | dtype | Description |
+|---|---|---|---|
+| `cells` | `[1, -1, 3]` | `int32` | Node indices of triangular cells. |
+| `mesh_pos` | `[1, -1, 2]` | `float32` | Two-dimensional coordinates of mesh nodes. |
+| `node_type` | `[1, -1, 1]` | `int32` | Node types, such as interior, boundary, inlet, and outlet. |
+| `velocity` | `[600, -1, 2]` | `float32` | Two-dimensional velocity fields over 600 time steps. |
+| `pressure` | `[600, -1, 1]` | `float32` | Pressure fields over 600 time steps. |
+
+`stats/` stores normalization statistics for edge features, node velocities, velocity differences, and pressure.
+
+## How to Use the Dataset
+
+This dataset is compatible with the `OneScience-Sugon/MeshGraphNet` model. Download the dataset and model:
+
+```bash
+hf download --dataset OneScience-Sugon/cylinder_flow --local-dir ./cylinder_flow
+hf download --model OneScience-Sugon/MeshGraphNet --local-dir ./MeshGraphNet
+```
+
+The validation script included with the dataset can be used to check the directory structure, metadata, and the first TFRecord sample:
+
+```bash
+cd cylinder_flow
+python scripts/validate_cylinder_flow_dataset.py
+```
+
+Append the `--verify-sha256` option to perform full SHA256 verification.
+
+## Official OneScience Information
+
+| Platform | OneScience Main Repository | Skills Repository |
+|---|---|---|
+| Gitee | https://gitee.com/onescience-ai/onescience | https://gitee.com/onescience-ai/oneskills |
+| GitHub | https://github.com/onescience-ai/OneScience | https://github.com/onescience-ai/oneskills |
+
+## Citation and License
+
+- Original Cylinder Flow paper: [Learning Mesh-Based Simulation with Graph Networks](https://arxiv.org/abs/2010.03409)
+- This repository retains information about the data source and organizes the data for automated runs on OneScience ModelScope. Before public distribution or redistribution, confirm the licensing requirements of the upstream project.

README_zh.md

@@ -0,0 +1,86 @@
+---
+license: other
+#用户自定义标签
+tags:
+  - CylinderFlow
+  - computational fluid dynamics
+  - incompressible flow
+language:
+  - en
+  - zh
+---
+<p align="center">
+  <strong>
+    <span style="font-size: 30px;">Cylinder Flow</span>
+  </strong>
+</p>
+
+## 数据集描述
+
+Cylinder Flow 数据集来源于 DeepMind 的 MeshGraphNets 基准，用于描述二维非结构化三角网格上的圆柱绕流（vortex shedding）。数据由 COMSOL 仿真生成，每条轨迹包含 600 个时间步，时间步长为 `dt=0.01`，记录流体速度场和压力场。
+
+论文：[Learning Mesh-Based Simulation with Graph Networks](https://arxiv.org/abs/2010.03409)
+
+## 数据集支持的任务
+
+| 场景 | 说明 |
+|---|---|
+| 流场时序预测 | 根据历史网格状态预测后续速度和压力场。 |
+| 涡脱落模拟 | 学习圆柱绕流中的瞬态流动演化。 |
+| 图神经网络仿真 | 为 MeshGraphNet 等网格模型提供训练和评测数据。 |
+| 网格物理建模 | 在非结构化三角网格上学习节点和边的物理关系。 |
+
+## 数据集的格式和结构
+
+数据采用 TFRecord 格式，按训练、验证和测试划分：
+
+```text
+data/cylinder_flow/
+  train.tfrecord
+  valid.tfrecord
+  test.tfrecord
+  meta.json
+  stats/
+```
+
+每条记录对应一条非结构化网格轨迹，主要字段如下：
+
+| 字段 | shape | dtype | 说明 |
+|---|---|---|---|
+| `cells` | `[1, -1, 3]` | `int32` | 三角形单元的节点索引。 |
+| `mesh_pos` | `[1, -1, 2]` | `float32` | 网格节点的二维坐标。 |
+| `node_type` | `[1, -1, 1]` | `int32` | 节点类型，如内部、边界、入口和出口。 |
+| `velocity` | `[600, -1, 2]` | `float32` | 600 个时间步的二维速度场。 |
+| `pressure` | `[600, -1, 1]` | `float32` | 600 个时间步的压力场。 |
+
+`stats/` 中保存边特征、节点速度、速度差分和压力的归一化统计量。
+
+## 数据集使用方式
+
+本数据集适配 `OneScience-Sugon/MeshGraphNet` 模型。下载数据集和模型：
+
+```bash
+hf download --dataset OneScience-Sugon/cylinder_flow --local-dir ./cylinder_flow
+hf download --model OneScience-Sugon/MeshGraphNet --local-dir ./MeshGraphNet
+```
+
+可使用数据集内的校验脚本检查目录结构、元信息和 TFRecord 首条样本：
+
+```bash
+cd cylinder_flow
+python scripts/validate_cylinder_flow_dataset.py
+```
+
+完整 SHA256 校验可附加 `--verify-sha256` 参数。
+
+## OneScience 官方信息
+
+| 平台 | OneScience 主仓库 | Skills 仓库 |
+|---|---|---|
+| Gitee | https://gitee.com/onescience-ai/onescience | https://gitee.com/onescience-ai/oneskills |
+| GitHub | https://github.com/onescience-ai/OneScience | https://github.com/onescience-ai/oneskills |
+
+## 引用与许可证
+
+- Cylinder Flow 原始论文：[Learning Mesh-Based Simulation with Graph Networks](https://arxiv.org/abs/2010.03409)
+- 本仓库保留数据来源说明，并面向 OneScience ModelScope 自动运行场景进行整理；公开分发或二次发布前，请根据上游项目确认许可证要求。

data/cylinder_flow/meta.json

@@ -0,0 +1,60 @@
+{
+  "simulator": "comsol",
+  "dt": 0.01,
+  "collision_radius": null,
+  "features": {
+    "cells": {
+      "type": "static",
+      "shape": [
+        1,
+        -1,
+        3
+      ],
+      "dtype": "int32"
+    },
+    "mesh_pos": {
+      "type": "static",
+      "shape": [
+        1,
+        -1,
+        2
+      ],
+      "dtype": "float32"
+    },
+    "node_type": {
+      "type": "static",
+      "shape": [
+        1,
+        -1,
+        1
+      ],
+      "dtype": "int32"
+    },
+    "velocity": {
+      "type": "dynamic",
+      "shape": [
+        600,
+        -1,
+        2
+      ],
+      "dtype": "float32"
+    },
+    "pressure": {
+      "type": "dynamic",
+      "shape": [
+        600,
+        -1,
+        1
+      ],
+      "dtype": "float32"
+    }
+  },
+  "field_names": [
+    "cells",
+    "mesh_pos",
+    "node_type",
+    "velocity",
+    "pressure"
+  ],
+  "trajectory_length": 600
+}

data/cylinder_flow/stats/edge_stats.json

@@ -0,0 +1,12 @@
+{
+    "edge_mean": [
+        -1.8337848600724627e-12,
+        1.9015786666652656e-13,
+        0.019958416000008583
+    ],
+    "edge_std": [
+        0.01587940752506256,
+        0.014747538603842258,
+        0.008444361388683319
+    ]
+}

data/cylinder_flow/stats/node_stats.json

@@ -0,0 +1,24 @@
+{
+    "velocity_mean": [
+        0.5875466465950012,
+        0.0005995402461849153
+    ],
+    "velocity_diff_mean": [
+        -0.0006314588827081025,
+        2.141517825293704e-06
+    ],
+    "pressure_mean": [
+        0.02848142944276333
+    ],
+    "velocity_std": [
+        0.5751428008079529,
+        0.14606069028377533
+    ],
+    "pressure_std": [
+        0.3340202569961548
+    ],
+    "velocity_diff_std": [
+        0.021763920783996582,
+        0.020587600767612457
+    ]
+}

data/cylinder_flow/test.tfrecord

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8522932a23da6ccdee996c56158e4b908f7091f0ade11e8acea700be321af8c3
+size 1355376404

data/cylinder_flow/train.tfrecord

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea63bf985ef131d542b763dded130f242040b9389a59b76f5e2d61a3f40c6ada
+size 13645805387

data/cylinder_flow/valid.tfrecord

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79171fa282fcd4fd5ff93d13e04ade43e83c1b11a0b96eee0951e4e0ed37f3ca
+size 1363987289

data_integrity_summary.json

@@ -0,0 +1,104 @@
+{
+  "source_root": "cfd_dataset/vortex_shedding_mgn/cylinder_flow",
+  "standardized_root": "modelscope_standardized/dataset/cfd_cylinder_flow",
+  "file_count": 6,
+  "all_files_match": true,
+  "files": [
+    {
+      "name": "meta.json",
+      "source": {
+        "path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/meta.json",
+        "name": "meta.json",
+        "size": 883,
+        "sha256": "2a3e39429a55a0cf47355717cc07f4b292629daeb48a89abd518ea0402033e96"
+      },
+      "standardized": {
+        "path": "data/cylinder_flow/meta.json",
+        "name": "meta.json",
+        "size": 883,
+        "sha256": "2a3e39429a55a0cf47355717cc07f4b292629daeb48a89abd518ea0402033e96"
+      },
+      "matches": true
+    },
+    {
+      "name": "train.tfrecord",
+      "source": {
+        "path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/train.tfrecord",
+        "name": "train.tfrecord",
+        "size": 13645805387,
+        "sha256": "ea63bf985ef131d542b763dded130f242040b9389a59b76f5e2d61a3f40c6ada"
+      },
+      "standardized": {
+        "path": "data/cylinder_flow/train.tfrecord",
+        "name": "train.tfrecord",
+        "size": 13645805387,
+        "sha256": "ea63bf985ef131d542b763dded130f242040b9389a59b76f5e2d61a3f40c6ada"
+      },
+      "matches": true
+    },
+    {
+      "name": "valid.tfrecord",
+      "source": {
+        "path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/valid.tfrecord",
+        "name": "valid.tfrecord",
+        "size": 1363987289,
+        "sha256": "79171fa282fcd4fd5ff93d13e04ade43e83c1b11a0b96eee0951e4e0ed37f3ca"
+      },
+      "standardized": {
+        "path": "data/cylinder_flow/valid.tfrecord",
+        "name": "valid.tfrecord",
+        "size": 1363987289,
+        "sha256": "79171fa282fcd4fd5ff93d13e04ade43e83c1b11a0b96eee0951e4e0ed37f3ca"
+      },
+      "matches": true
+    },
+    {
+      "name": "test.tfrecord",
+      "source": {
+        "path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/test.tfrecord",
+        "name": "test.tfrecord",
+        "size": 1355376404,
+        "sha256": "8522932a23da6ccdee996c56158e4b908f7091f0ade11e8acea700be321af8c3"
+      },
+      "standardized": {
+        "path": "data/cylinder_flow/test.tfrecord",
+        "name": "test.tfrecord",
+        "size": 1355376404,
+        "sha256": "8522932a23da6ccdee996c56158e4b908f7091f0ade11e8acea700be321af8c3"
+      },
+      "matches": true
+    },
+    {
+      "name": "stats/edge_stats.json",
+      "source": {
+        "path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/stats/edge_stats.json",
+        "name": "stats/edge_stats.json",
+        "size": 235,
+        "sha256": "ab0860bf8bece19460278095592c6bb90fa06e7ae77274530be48b3b9d3d799a"
+      },
+      "standardized": {
+        "path": "data/cylinder_flow/stats/edge_stats.json",
+        "name": "stats/edge_stats.json",
+        "size": 235,
+        "sha256": "ab0860bf8bece19460278095592c6bb90fa06e7ae77274530be48b3b9d3d799a"
+      },
+      "matches": true
+    },
+    {
+      "name": "stats/node_stats.json",
+      "source": {
+        "path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/stats/node_stats.json",
+        "name": "stats/node_stats.json",
+        "size": 479,
+        "sha256": "1b595ea651193e3e22191dbb5afdfabd8bc96dc8dcfea4d73d92c99512eaae86"
+      },
+      "standardized": {
+        "path": "data/cylinder_flow/stats/node_stats.json",
+        "name": "stats/node_stats.json",
+        "size": 479,
+        "sha256": "1b595ea651193e3e22191dbb5afdfabd8bc96dc8dcfea4d73d92c99512eaae86"
+      },
+      "matches": true
+    }
+  ]
+}

files_sha256.jsonl

@@ -0,0 +1,6 @@
+{"path": "data/cylinder_flow/meta.json", "sha256": "2a3e39429a55a0cf47355717cc07f4b292629daeb48a89abd518ea0402033e96", "size": 883, "source_path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/meta.json"}
+{"path": "data/cylinder_flow/train.tfrecord", "sha256": "ea63bf985ef131d542b763dded130f242040b9389a59b76f5e2d61a3f40c6ada", "size": 13645805387, "source_path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/train.tfrecord"}
+{"path": "data/cylinder_flow/valid.tfrecord", "sha256": "79171fa282fcd4fd5ff93d13e04ade43e83c1b11a0b96eee0951e4e0ed37f3ca", "size": 1363987289, "source_path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/valid.tfrecord"}
+{"path": "data/cylinder_flow/test.tfrecord", "sha256": "8522932a23da6ccdee996c56158e4b908f7091f0ade11e8acea700be321af8c3", "size": 1355376404, "source_path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/test.tfrecord"}
+{"path": "data/cylinder_flow/stats/edge_stats.json", "sha256": "ab0860bf8bece19460278095592c6bb90fa06e7ae77274530be48b3b9d3d799a", "size": 235, "source_path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/stats/edge_stats.json"}
+{"path": "data/cylinder_flow/stats/node_stats.json", "sha256": "1b595ea651193e3e22191dbb5afdfabd8bc96dc8dcfea4d73d92c99512eaae86", "size": 479, "source_path": "cfd_dataset/vortex_shedding_mgn/cylinder_flow/stats/node_stats.json"}

metadata/cylinder_flow_schema.json

@@ -0,0 +1,45 @@
+{
+  "dataset": "DeepMind CylinderFlow / Vortex Shedding",
+  "modelscope_dataset_id": "OneScience/cylinder_flow",
+  "format": "tfrecord",
+  "data_root": "data/cylinder_flow",
+  "splits": {
+    "train": "data/cylinder_flow/train.tfrecord",
+    "valid": "data/cylinder_flow/valid.tfrecord",
+    "test": "data/cylinder_flow/test.tfrecord"
+  },
+  "metadata": "data/cylinder_flow/meta.json",
+  "stats": {
+    "edge": "data/cylinder_flow/stats/edge_stats.json",
+    "node": "data/cylinder_flow/stats/node_stats.json"
+  },
+  "trajectory_length": 600,
+  "dt": 0.01,
+  "simulator": "comsol",
+  "features": {
+    "cells": {"type": "static", "shape": [1, -1, 3], "dtype": "int32"},
+    "mesh_pos": {"type": "static", "shape": [1, -1, 2], "dtype": "float32"},
+    "node_type": {"type": "static", "shape": [1, -1, 1], "dtype": "int32"},
+    "velocity": {"type": "dynamic", "shape": [600, -1, 2], "dtype": "float32"},
+    "pressure": {"type": "dynamic", "shape": [600, -1, 1], "dtype": "float32"}
+  },
+  "normalization_stats": {
+    "edge_stats": ["edge_mean", "edge_std"],
+    "node_stats": [
+      "velocity_mean",
+      "velocity_std",
+      "velocity_diff_mean",
+      "velocity_diff_std",
+      "pressure_mean",
+      "pressure_std"
+    ]
+  },
+  "default_model_config_usage": {
+    "train_samples": 400,
+    "train_steps": 300,
+    "val_samples": 10,
+    "val_steps": 300,
+    "test_samples": 10,
+    "test_steps": 300
+  }
+}

onescience_relations.yaml

@@ -0,0 +1,15 @@
+relations:
+  required_datasets: []
+  optional_datasets: []
+  compatible_models:
+    - id: OneScience/Vortex_shedding_mgn
+      resource_ref:
+        platform: modelscope
+        repo_id: OneScience/Vortex_shedding_mgn
+        repo_type: model
+        url: https://modelscope.cn/models/OneScience/Vortex_shedding_mgn
+        revision: main
+        readme_path: README.md
+        manifest_path: onescience_run_manifest.yaml
+      expected_local_path: data/cylinder_flow
+      expected_env: ONESCIENCE_CYLINDER_FLOW_DATA_DIR

onescience_run_manifest.yaml

@@ -0,0 +1,434 @@
+onescience_manifest_version: "0.1"
+resource_type: dataset
+
+resource:
+  id: OneScience/cylinder_flow
+  name: cylinder_flow
+  domain: cfd
+  domain_tags:
+    - cfd
+    - vortex_shedding
+    - cylinder_flow
+    - unstructured_mesh
+  task: transient_cylinder_flow_dataset
+  task_tags:
+    - train_data
+    - eval_data
+    - inference_input
+    - dataset_validation
+  modalities:
+    - unstructured_mesh
+    - graph
+    - transient_flow_field
+  input_formats:
+    - tfrecord
+    - json
+  output_formats:
+    - tfrecord
+    - json
+  summary: DeepMind CylinderFlow 圆柱绕流数据集，包含 train/valid/test TFRecord、meta.json 和 MeshGraphNet 归一化统计量。
+
+platform_resource:
+  primary:
+    platform: modelscope
+    repo_id: OneScience/cylinder_flow
+    repo_type: dataset
+    url: https://modelscope.cn/datasets/OneScience/cylinder_flow
+    revision: main
+    readme_path: README.md
+    manifest_path: onescience_run_manifest.yaml
+  mirrors: []
+  access:
+    visibility: public
+    license: unknown
+
+website_integration:
+  enabled: true
+  click_target:
+    platform: modelscope
+    resource_url: https://modelscope.cn/datasets/OneScience/cylinder_flow
+  llm_handoff:
+    readme_required: true
+    manifest_required: true
+    download_readme_first: true
+    resolve_related_models: true
+    default_run_goal: dataset_validation
+    cwd_note: 如果使用 modelscope download --cache_dir 下载数据集，请先 cd 到实际下载后的数据集仓库根目录；模型侧默认读取 data/cylinder_flow。
+
+runtime:
+  enabled: true
+  onescience_domain: cfd
+  min_onescience_version: null
+  supported_execution:
+    - local_dataset_validation
+    - model_training_input
+    - model_inference_input
+  environment:
+    exported_env:
+      ONESCIENCE_CYLINDER_FLOW_DATA_DIR: <dataset_repo_root>/data/cylinder_flow
+  dependencies:
+    python: ">=3.10"
+    python_packages:
+      - tensorflow
+      - numpy
+      - pyyaml
+
+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
+  source_paths:
+    - onescience/src/onescience/datapipes/cfd/deepmind_cylinderflow.py
+  compatibility:
+    examples_path: onescience/examples/cfd/Vortex_shedding_mgn
+    status: examples_compatible
+  datapipe: onescience.datapipes.cfd.DeepMind_CylinderFlowDatapipe
+
+runtime_package:
+  kind: standard_runtime_package
+  package_root: .
+  standard_layout:
+    workdir: .
+    data_dir: data/cylinder_flow
+    metadata_dir: metadata
+    output_dir: validation_outputs
+  apply_policy:
+    mode: direct_use
+    target: session_workdir
+    overwrite: false
+    protect_installed_onescience: true
+  entry_files:
+    - data/cylinder_flow/meta.json
+    - data/cylinder_flow/train.tfrecord
+    - data/cylinder_flow/valid.tfrecord
+    - data/cylinder_flow/test.tfrecord
+    - data/cylinder_flow/stats/edge_stats.json
+    - data/cylinder_flow/stats/node_stats.json
+    - metadata/cylinder_flow_schema.json
+    - scripts/validate_cylinder_flow_dataset.py
+    - files_sha256.jsonl
+  entrypoints:
+    preflight: scripts/validate_cylinder_flow_dataset.py
+    validate: scripts/validate_cylinder_flow_dataset.py
+    inference: null
+    train: null
+    finetune: null
+    evaluate: null
+    visualize: null
+    deploy: null
+
+files:
+  model_files: []
+  dataset_files:
+    - path: data/cylinder_flow/train.tfrecord
+      role: train_split
+      description_zh: CylinderFlow 训练轨迹 TFRecord，供 MeshGraphNet 训练使用。
+      required: true
+      required_for: [dataset_validation, train]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: data/cylinder_flow/train.tfrecord
+    - path: data/cylinder_flow/valid.tfrecord
+      role: validation_split
+      description_zh: CylinderFlow 验证轨迹 TFRecord，供训练过程验证使用。
+      required: true
+      required_for: [dataset_validation, train, evaluate]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: data/cylinder_flow/valid.tfrecord
+    - path: data/cylinder_flow/test.tfrecord
+      role: test_split
+      description_zh: CylinderFlow 测试轨迹 TFRecord，供推理、评测和可视化使用。
+      required: true
+      required_for: [dataset_validation, inference, evaluate, visualize]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: data/cylinder_flow/test.tfrecord
+    - path: data/cylinder_flow/meta.json
+      role: tfrecord_schema_metadata
+      description_zh: TFRecord 字段、shape、dtype、轨迹长度和仿真步长元数据。
+      required: true
+      required_for: [dataset_validation, train, inference, evaluate]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: data/cylinder_flow/meta.json
+    - path: data/cylinder_flow/stats/edge_stats.json
+      role: edge_normalization_stats
+      description_zh: MeshGraphNet 边特征归一化均值和标准差。
+      required: true
+      required_for: [dataset_validation, train, inference, evaluate]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: data/cylinder_flow/stats/edge_stats.json
+    - path: data/cylinder_flow/stats/node_stats.json
+      role: node_normalization_stats
+      description_zh: MeshGraphNet 节点速度、速度差和压力归一化均值与标准差。
+      required: true
+      required_for: [dataset_validation, train, inference, evaluate]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: data/cylinder_flow/stats/node_stats.json
+    - path: files_sha256.jsonl
+      role: file_size_sha256_inventory
+      description_zh: 整理后数据文件的大小和 SHA256 清单，用于完整性校验。
+      required: true
+      required_for: [dataset_validation]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: files_sha256.jsonl
+    - path: data_integrity_summary.json
+      role: source_to_standardized_integrity_summary
+      description_zh: 原始数据与整理后数据的文件名、大小、SHA256 对照摘要。
+      required: true
+      required_for: [dataset_validation]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: data_integrity_summary.json
+  config_files:
+    - path: metadata/cylinder_flow_schema.json
+      role: dataset_schema
+      description_zh: 描述 CylinderFlow TFRecord 字段、shape、dtype、split、metadata 和 stats。
+      required: true
+      required_for: [dataset_validation, train, inference]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: metadata/cylinder_flow_schema.json
+  sample_files:
+    - path: scripts/validate_cylinder_flow_dataset.py
+      role: dataset_validation_entry
+      description_zh: 数据集读取验证脚本，检查文件结构、shape、dtype、stats、可读性和 SHA256。
+      required: true
+      required_for: [dataset_validation]
+      source:
+        platform: modelscope
+        repo_id: OneScience/cylinder_flow
+        repo_type: dataset
+        download_method: command_ref
+        command_ref: commands.download.download_dataset
+      local_path: scripts/validate_cylinder_flow_dataset.py
+
+relations:
+  required_datasets: []
+  optional_datasets: []
+  compatible_models:
+    - id: OneScience/Vortex_shedding_mgn
+      role: train_eval_inference_model
+      required_for: [train, inference, evaluate, visualize]
+      resource_ref:
+        platform: modelscope
+        repo_id: OneScience/Vortex_shedding_mgn
+        repo_type: model
+        url: https://modelscope.cn/models/OneScience/Vortex_shedding_mgn
+        revision: main
+        readme_path: README.md
+        manifest_path: onescience_run_manifest.yaml
+      expected_local_path: data/cylinder_flow
+      expected_env: ONESCIENCE_CYLINDER_FLOW_DATA_DIR
+
+run_matrix:
+  scenarios:
+    - name: validate_structure
+      capability: dataset_validation
+      default: true
+      required_datasets:
+        - id: OneScience/cylinder_flow
+          role: train_eval_inference_data
+          local_path: data/cylinder_flow
+      required_model_files: []
+      required_dataset_files:
+        - data/cylinder_flow/meta.json
+        - data/cylinder_flow/train.tfrecord
+        - data/cylinder_flow/valid.tfrecord
+        - data/cylinder_flow/test.tfrecord
+        - metadata/cylinder_flow_schema.json
+      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] CylinderFlow dataset validation passed
+    - name: validate_integrity_full_hash
+      capability: dataset_validation
+      default: false
+      required_datasets:
+        - id: OneScience/cylinder_flow
+          role: train_eval_inference_data
+          local_path: data/cylinder_flow
+      required_model_files: []
+      required_dataset_files:
+        - files_sha256.jsonl
+        - data/cylinder_flow/train.tfrecord
+        - data/cylinder_flow/valid.tfrecord
+        - data/cylinder_flow/test.tfrecord
+      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] CylinderFlow dataset validation passed
+    - name: provide_to_vortex_shedding_mgn_model
+      capability: train
+      default: true
+      required_datasets:
+        - id: OneScience/cylinder_flow
+          role: train_data
+          local_path: data/cylinder_flow
+      required_model_files:
+        - OneScience/Vortex_shedding_mgn:conf/mgn_cylinderflow.yaml
+      required_dataset_files:
+        - data/cylinder_flow/train.tfrecord
+        - data/cylinder_flow/valid.tfrecord
+        - data/cylinder_flow/test.tfrecord
+      preconditions:
+        - 模型仓库 OneScience/Vortex_shedding_mgn 已下载。
+        - 将本仓库 data/cylinder_flow 放到模型包根目录 data/cylinder_flow，或设置 ONESCIENCE_CYLINDER_FLOW_DATA_DIR 指向本目录用于预检。
+      command_refs:
+        - commands.prepare.export_dataset_env
+      outputs:
+        - 模型侧 scripts/preflight_vortex_shedding_mgn.py 可以读取数据集。
+
+capabilities:
+  dataset_validation: true
+  train_input: true
+  inference_input: true
+  evaluation_input: true
+  visualization_input: true
+  preflight: true
+  inference: false
+  train: false
+  finetune: false
+  evaluate: false
+  visualize: false
+  deploy: false
+
+commands:
+  download:
+    - name: download_dataset
+      command: modelscope download --dataset OneScience/cylinder_flow
+      cwd: session_workdir
+    - name: download_model
+      command: modelscope download --model OneScience/Vortex_shedding_mgn
+      cwd: session_workdir
+  prepare:
+    - name: export_dataset_env
+      command: export ONESCIENCE_CYLINDER_FLOW_DATA_DIR=/path/to/OneScience_cylinder_flow/data/cylinder_flow
+      cwd: .
+  preflight:
+    - name: validate_dataset
+      command: python scripts/validate_cylinder_flow_dataset.py --dataset-root . --skip-tfrecord-read
+      cwd: .
+    - name: validate_dataset_full_hash
+      command: python scripts/validate_cylinder_flow_dataset.py --dataset-root . --verify-sha256 --skip-tfrecord-read
+      cwd: .
+    - name: validate_dataset_read_first_record
+      command: python scripts/validate_cylinder_flow_dataset.py --dataset-root .
+      cwd: .
+  inference: []
+  train: []
+  finetune: []
+  evaluate: []
+  visualize: []
+  deploy: []
+
+expected_outputs:
+  - name: validate_dataset
+    paths: []
+    success_criteria:
+      - 控制台输出 [OK] CylinderFlow dataset validation passed
+  - name: validate_dataset_full_hash
+    paths:
+      - files_sha256.jsonl
+    success_criteria:
+      - SHA256 与 files_sha256.jsonl 完全一致。
+  - name: model_consumption
+    paths: []
+    success_criteria:
+      - 模型包 scripts/preflight_vortex_shedding_mgn.py 通过数据路径和 schema 检查。
+
+diagnostics:
+  - name: missing_data_root
+    symptom: missing required file 或 missing data directory
+    action: 确认当前目录为 OneScience/cylinder_flow 数据集仓库根目录，且 data/cylinder_flow 存在。
+  - name: missing_tfrecord
+    symptom: "missing required file: data/cylinder_flow/*.tfrecord"
+    action: 对照 files_sha256.jsonl 重新上传或重新下载缺失 TFRecord。
+  - name: schema_mismatch
+    symptom: meta.json feature shape mismatch 或 dtype mismatch
+    action: 确认没有用其他 CylinderFlow 数据覆盖当前文件。
+  - name: checksum_mismatch
+    symptom: sha256 mismatch
+    action: 删除损坏文件后重新下载 OneScience/cylinder_flow。
+  - name: model_cannot_find_data
+    symptom: 模型侧报告 data/cylinder_flow 不存在
+    action: 将本仓库 data/cylinder_flow 复制到模型包根目录 data/cylinder_flow，或修改模型包配置副本中的 data_dir 与 stats_dir。
+
+domain_extension:
+  cfd:
+    dataset_family: DeepMind CylinderFlow
+    physics: transient incompressible cylinder wake / vortex shedding
+    simulator: comsol
+    file_count: 6
+    trajectory_length: 600
+    dt: 0.01
+    schema:
+      cells:
+        shape: [1, -1, 3]
+        dtype: int32
+      mesh_pos:
+        shape: [1, -1, 2]
+        dtype: float32
+      node_type:
+        shape: [1, -1, 1]
+        dtype: int32
+      velocity:
+        shape: [600, -1, 2]
+        dtype: float32
+      pressure:
+        shape: [600, -1, 1]
+        dtype: float32

scripts/validate_cylinder_flow_dataset.py

@@ -0,0 +1,182 @@
+#!/usr/bin/env python3
+"""Validate the standardized DeepMind CylinderFlow dataset package."""
+
+from __future__ import annotations
+
+import argparse
+import hashlib
+import json
+from pathlib import Path
+
+
+EXPECTED_SPLITS = ["train", "valid", "test"]
+EXPECTED_REQUIRED = {
+    "data/cylinder_flow/meta.json",
+    "data/cylinder_flow/train.tfrecord",
+    "data/cylinder_flow/valid.tfrecord",
+    "data/cylinder_flow/test.tfrecord",
+    "data/cylinder_flow/stats/edge_stats.json",
+    "data/cylinder_flow/stats/node_stats.json",
+}
+
+
+def fail(message: str) -> None:
+    raise SystemExit(f"[FAIL] {message}")
+
+
+def sha256_file(path: Path) -> str:
+    digest = hashlib.sha256()
+    with path.open("rb") as f:
+        for chunk in iter(lambda: f.read(1024 * 1024), b""):
+            digest.update(chunk)
+    return digest.hexdigest()
+
+
+def parse_args() -> argparse.Namespace:
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--dataset-root", default=".", help="dataset package root")
+    parser.add_argument(
+        "--verify-sha256",
+        action="store_true",
+        help="recompute and verify SHA256 values from files_sha256.jsonl",
+    )
+    parser.add_argument(
+        "--skip-tfrecord-read",
+        action="store_true",
+        help="skip optional TensorFlow first-record readability check",
+    )
+    return parser.parse_args()
+
+
+def check_files(dataset_root: Path) -> dict:
+    sizes = {}
+    for rel_path in sorted(EXPECTED_REQUIRED):
+        path = dataset_root / rel_path
+        if not path.is_file():
+            fail(f"missing required file: {rel_path}")
+        if path.stat().st_size <= 0:
+            fail(f"empty required file: {rel_path}")
+        sizes[rel_path] = path.stat().st_size
+    return sizes
+
+
+def check_metadata(data_dir: Path) -> dict:
+    meta = json.loads((data_dir / "meta.json").read_text(encoding="utf-8"))
+    if meta.get("simulator") != "comsol":
+        fail("meta.json simulator must be comsol")
+    if meta.get("trajectory_length") != 600:
+        fail("meta.json trajectory_length must be 600")
+    expected_fields = ["cells", "mesh_pos", "node_type", "velocity", "pressure"]
+    if meta.get("field_names") != expected_fields:
+        fail("meta.json field_names mismatch")
+    expected_shapes = {
+        "cells": [1, -1, 3],
+        "mesh_pos": [1, -1, 2],
+        "node_type": [1, -1, 1],
+        "velocity": [600, -1, 2],
+        "pressure": [600, -1, 1],
+    }
+    expected_dtypes = {
+        "cells": "int32",
+        "mesh_pos": "float32",
+        "node_type": "int32",
+        "velocity": "float32",
+        "pressure": "float32",
+    }
+    for key in expected_fields:
+        feature = meta.get("features", {}).get(key)
+        if not feature:
+            fail(f"meta.json missing feature {key}")
+        if feature.get("shape") != expected_shapes[key]:
+            fail(f"meta.json feature {key} shape mismatch")
+        if feature.get("dtype") != expected_dtypes[key]:
+            fail(f"meta.json feature {key} dtype mismatch")
+    return meta
+
+
+def check_stats(stats_dir: Path) -> None:
+    edge = json.loads((stats_dir / "edge_stats.json").read_text(encoding="utf-8"))
+    node = json.loads((stats_dir / "node_stats.json").read_text(encoding="utf-8"))
+    for key in ["edge_mean", "edge_std"]:
+        if not isinstance(edge.get(key), list) or len(edge[key]) != 3:
+            fail(f"edge_stats.json {key} must be length 3")
+    for key in ["velocity_mean", "velocity_std", "velocity_diff_mean", "velocity_diff_std"]:
+        if not isinstance(node.get(key), list) or len(node[key]) != 2:
+            fail(f"node_stats.json {key} must be length 2")
+    for key in ["pressure_mean", "pressure_std"]:
+        if not isinstance(node.get(key), list) or len(node[key]) != 1:
+            fail(f"node_stats.json {key} must be length 1")
+
+
+def verify_inventory(dataset_root: Path, sizes: dict, verify_sha256: bool) -> None:
+    inventory_path = dataset_root / "files_sha256.jsonl"
+    if not inventory_path.is_file():
+        fail("missing files_sha256.jsonl")
+    seen = {}
+    for line in inventory_path.read_text(encoding="utf-8").splitlines():
+        if not line.strip():
+            continue
+        item = json.loads(line)
+        rel_path = item["path"]
+        path = dataset_root / rel_path
+        if not path.is_file():
+            fail(f"inventory path missing on disk: {rel_path}")
+        if path.stat().st_size != item["size"]:
+            fail(f"size mismatch for {rel_path}")
+        if rel_path in sizes and sizes[rel_path] != item["size"]:
+            fail(f"required file size mismatch for {rel_path}")
+        if verify_sha256 and sha256_file(path) != item["sha256"]:
+            fail(f"sha256 mismatch for {rel_path}")
+        seen[rel_path] = item
+    missing = sorted(EXPECTED_REQUIRED - set(seen))
+    if missing:
+        fail(f"inventory missing required files: {missing}")
+
+
+def check_tfrecord_first_record(data_dir: Path, meta: dict) -> None:
+    try:
+        import numpy as np
+        import tensorflow.compat.v1 as tf
+    except Exception as exc:  # pragma: no cover - optional runtime dependency
+        print(f"[WARN] TensorFlow first-record check skipped: {exc}")
+        return
+
+    feature_dict = {k: tf.io.VarLenFeature(tf.string) for k in meta["field_names"]}
+    for split in EXPECTED_SPLITS:
+        record_iter = iter(tf.data.TFRecordDataset(str(data_dir / f"{split}.tfrecord")).take(1))
+        try:
+            raw = next(record_iter)
+        except StopIteration:
+            fail(f"{split}.tfrecord contains no records")
+        features = tf.io.parse_single_example(raw, feature_dict)
+        velocity = np.frombuffer(features["velocity"].values[0].numpy(), dtype=np.float32)
+        pressure = np.frombuffer(features["pressure"].values[0].numpy(), dtype=np.float32)
+        mesh_pos = np.frombuffer(features["mesh_pos"].values[0].numpy(), dtype=np.float32)
+        cells = np.frombuffer(features["cells"].values[0].numpy(), dtype=np.int32)
+        if velocity.size % (meta["trajectory_length"] * 2) != 0:
+            fail(f"{split}.tfrecord velocity shape incompatible with metadata")
+        nodes = velocity.size // (meta["trajectory_length"] * 2)
+        if pressure.size != meta["trajectory_length"] * nodes:
+            fail(f"{split}.tfrecord pressure node count mismatch")
+        if mesh_pos.size != nodes * 2:
+            fail(f"{split}.tfrecord mesh_pos node count mismatch")
+        if cells.size % 3 != 0:
+            fail(f"{split}.tfrecord cells are not triangular")
+        print(f"[OK] {split}.tfrecord first record: nodes={nodes}, cells={cells.size // 3}")
+
+
+def main() -> None:
+    args = parse_args()
+    dataset_root = Path(args.dataset_root).resolve()
+    data_dir = dataset_root / "data" / "cylinder_flow"
+    sizes = check_files(dataset_root)
+    meta = check_metadata(data_dir)
+    check_stats(data_dir / "stats")
+    verify_inventory(dataset_root, sizes, args.verify_sha256)
+    if not args.skip_tfrecord_read:
+        check_tfrecord_first_record(data_dir, meta)
+    print("[OK] CylinderFlow dataset validation passed")
+
+
+if __name__ == "__main__":
+    main()