Redesign dataset workbench management

backend/app/main.py

@@ -26,14 +26,14 @@ from .coverage import get_coverage_report
 from .jobs import cancel_job, create_job
 from .modules.results.service import scan_results, scan_training_curves
 from .modules.system.service import disk_usage, get_conda_envs, get_gpus, get_runtime_readiness
-from .modules.dataset.service import create_dataset, generate_yolo_dataset_yaml, list_uploaded_datasets, save_upload, validate_dataset
+from .modules.dataset.service import copy_dataset, create_dataset, delete_dataset, describe_dataset, generate_yolo_dataset_yaml, list_uploaded_datasets, save_upload, validate_dataset
 from .modules.weights.service import load_manifest, sync_weights, verify_weights
 from .agents.evaluation_agent import evaluate_project
 from .agents.user_agent import latest_user_agent_report, run_user_agent
 from .agents.validation_agent import validate_project
 from .paths import ensure_inside
 from .progress import progress_from_log_path
-from .schemas import DatasetCreate, DatasetYoloYamlRequest, JobCreate, ProfileCreate, WeightSyncRequest
+from .schemas import DatasetCopyRequest, DatasetCreate, DatasetYoloYamlRequest, JobCreate, ProfileCreate, WeightSyncRequest
 
 app = FastAPI(title="Seg Data Server Net", version="0.1.0")
 
@@ -149,9 +149,36 @@ def api_datasets() -> list[dict]:
     return list_uploaded_datasets()
 
 
+@app.get("/api/datasets/{dataset_name}")
+def api_dataset(dataset_name: str) -> dict:
+    try:
+        return describe_dataset(dataset_name)
+    except Exception as exc:
+        raise HTTPException(status_code=400, detail=str(exc)) from exc
+
+
 @app.post("/api/datasets")
 def api_create_dataset(dataset: DatasetCreate) -> dict:
-    return create_dataset(dataset.name, dataset.description)
+    try:
+        return create_dataset(dataset.name, dataset.description)
+    except Exception as exc:
+        raise HTTPException(status_code=400, detail=str(exc)) from exc
+
+
+@app.post("/api/datasets/{dataset_name}/copy")
+def api_copy_dataset(dataset_name: str, request: DatasetCopyRequest) -> dict:
+    try:
+        return copy_dataset(dataset_name, request.name, request.description)
+    except Exception as exc:
+        raise HTTPException(status_code=400, detail=str(exc)) from exc
+
+
+@app.delete("/api/datasets/{dataset_name}")
+def api_delete_dataset(dataset_name: str) -> dict:
+    try:
+        return delete_dataset(dataset_name)
+    except Exception as exc:
+        raise HTTPException(status_code=400, detail=str(exc)) from exc
 
 
 @app.post("/api/datasets/{dataset_name}/upload/{kind}")

backend/app/modules/dataset/service.py

@@ -16,6 +16,7 @@ from ...config import settings
 DATASET_KINDS = ("images", "labels", "masks")
 IMAGE_EXTS = {".png", ".jpg", ".jpeg", ".bmp", ".tif", ".tiff"}
 LABEL_EXTS = {".txt", ".json", ".yaml", ".yml"}
+VIDEO_EXTS = {".mp4", ".avi", ".mov", ".mkv", ".wmv", ".flv", ".webm", ".m4v"}
 ARCHIVE_SUFFIXES = (".zip", ".tar", ".tar.gz", ".tgz")
 ARCHIVE_ALIASES = {
     "images": {"image", "images", "img", "imgs", "ori", "original", "originals"},
@@ -47,7 +48,7 @@ def safe_filename(value: str | None) -> str:
 
 def upload_exts_for_kind(kind: str) -> set[str]:
     if kind == "images":
-        return IMAGE_EXTS
+        return IMAGE_EXTS | VIDEO_EXTS
     if kind == "masks":
         return IMAGE_EXTS | LABEL_EXTS
     if kind == "labels":
@@ -138,6 +139,8 @@ def metadata_path(name: str) -> Path:
 def create_dataset(name: str, description: str = "") -> dict:
     safe_name = slugify(name)
     root = dataset_dir(safe_name)
+    if root.exists():
+        raise ValueError(f"dataset already exists: {safe_name}")
     for kind in DATASET_KINDS:
         (root / kind).mkdir(parents=True, exist_ok=True)
     meta = {
@@ -155,6 +158,44 @@ def create_dataset(name: str, description: str = "") -> dict:
     return describe_dataset(safe_name)
 
 
+def delete_dataset(name: str) -> dict:
+    safe_name = slugify(name)
+    root = dataset_dir(safe_name)
+    if not root.exists() or not root.is_dir():
+        raise ValueError(f"dataset not found: {safe_name}")
+    shutil.rmtree(root)
+    return {"deleted": True, "name": safe_name}
+
+
+def copy_dataset(source_name: str, target_name: str, description: str | None = None) -> dict:
+    source_safe = slugify(source_name)
+    target_safe = slugify(target_name)
+    source = dataset_dir(source_safe)
+    target = dataset_dir(target_safe)
+    if not source.exists() or not source.is_dir():
+        raise ValueError(f"dataset not found: {source_safe}")
+    if target.exists():
+        raise ValueError(f"dataset already exists: {target_safe}")
+    shutil.copytree(source, target)
+    meta = _load_meta(source_safe)
+    meta.update(
+        {
+            "name": target_safe,
+            "description": description if description is not None else f"Copy of {source_safe}",
+            "created_at": datetime.now(timezone.utc).isoformat(),
+            "copied_from": source_safe,
+            "root": str(target.relative_to(settings.project_root)),
+            "layout": {
+                "images": str((target / "images").relative_to(settings.project_root)),
+                "labels": str((target / "labels").relative_to(settings.project_root)),
+                "masks": str((target / "masks").relative_to(settings.project_root)),
+            },
+        }
+    )
+    metadata_path(target_safe).write_text(json.dumps(meta, ensure_ascii=False, indent=2), encoding="utf-8")
+    return describe_dataset(target_safe)
+
+
 def _load_meta(name: str) -> dict:
     path = metadata_path(name)
     if path.exists():
@@ -192,6 +233,22 @@ def _image_shape(path: Path) -> dict | None:
         return None
 
 
+def _file_info(path: Path, kind: str) -> dict:
+    suffix = path.suffix.lower()
+    stem = path.stem
+    return {
+        "name": path.name,
+        "stem": stem,
+        "pair_stem": normalize_mask_stem(stem) if kind == "masks" else stem,
+        "path": str(path.resolve()),
+        "relative_path": str(path.resolve().relative_to(settings.project_root)),
+        "size": path.stat().st_size,
+        "ext": suffix,
+        "previewable": suffix in IMAGE_EXTS,
+        "media_type": "image" if suffix in IMAGE_EXTS else "video" if suffix in VIDEO_EXTS else "file",
+    }
+
+
 def _validate_yolo_txt(path: Path) -> dict:
     errors = []
     classes: set[int] = set()
@@ -232,20 +289,14 @@ def describe_dataset(name: str) -> dict:
     absolute_layout = {kind: str((root / kind).resolve()) for kind in DATASET_KINDS}
     counts = {}
     samples = {}
+    files_by_kind = {}
     for kind in sorted(DATASET_KINDS):
         files = list(_iter_files(root / kind))
         counts[kind] = len(files)
-        samples[kind] = [
-            {
-                "name": path.name,
-                "path": str(path.resolve()),
-                "relative_path": str(path.resolve().relative_to(settings.project_root)),
-                "size": path.stat().st_size,
-                "previewable": path.suffix.lower() in IMAGE_EXTS,
-            }
-            for path in files[:80]
-        ]
-    return {**meta, "absolute_layout": absolute_layout, "counts": counts, "samples": samples}
+        file_items = [_file_info(path, kind) for path in files[:2000]]
+        files_by_kind[kind] = file_items
+        samples[kind] = file_items[:80]
+    return {**meta, "absolute_layout": absolute_layout, "counts": counts, "samples": samples, "files": files_by_kind}
 
 
 def validate_dataset(name: str) -> dict:

backend/app/schemas.py

@@ -49,5 +49,10 @@ class DatasetCreate(BaseModel):
     description: str = ""
 
 
+class DatasetCopyRequest(BaseModel):
+    name: str
+    description: str | None = None
+
+
 class DatasetYoloYamlRequest(BaseModel):
     class_names: list[str] | None = None

backend/tests/test_dataset_service.py

@@ -4,9 +4,10 @@ import zipfile
 
 import cv2
 import numpy as np
+import pytest
 from fastapi import UploadFile
 
-from app.modules.dataset.service import create_dataset, describe_dataset, generate_yolo_dataset_yaml, save_upload, validate_dataset
+from app.modules.dataset.service import copy_dataset, create_dataset, delete_dataset, describe_dataset, generate_yolo_dataset_yaml, save_upload, validate_dataset
 
 
 def test_create_dataset_layout(tmp_path, monkeypatch):
@@ -19,6 +20,32 @@ def test_create_dataset_layout(tmp_path, monkeypatch):
     assert created["counts"] == {"images": 0, "labels": 0, "masks": 0}
     described = describe_dataset("case_01")
     assert described["layout"]["images"].endswith("images")
+    assert described["files"]["images"] == []
+    with pytest.raises(ValueError, match="already exists"):
+        create_dataset("case 01", "duplicate")
+
+
+def test_copy_and_delete_dataset(tmp_path, monkeypatch):
+    from types import SimpleNamespace
+    from app.modules.dataset import service
+
+    monkeypatch.setattr(service, "settings", SimpleNamespace(project_root=tmp_path))
+    create_dataset("case base", "demo")
+    root = service.dataset_dir("case_base")
+    (root / "images" / "sample.png").write_bytes(b"image")
+
+    copied = copy_dataset("case_base", "case copy", "copy")
+
+    assert copied["name"] == "case_copy"
+    assert copied["description"] == "copy"
+    assert copied["counts"]["images"] == 1
+    with pytest.raises(ValueError, match="already exists"):
+        copy_dataset("case_base", "case_copy")
+
+    deleted = delete_dataset("case_copy")
+
+    assert deleted == {"deleted": True, "name": "case_copy"}
+    assert not service.dataset_dir("case_copy").exists()
 
 
 def test_validate_dataset_and_generate_yolo_yaml(tmp_path, monkeypatch):