Commit d8fb40a6 authored by Jan Reimes's avatar Jan Reimes
Browse files

🔧 feat(workspaces): add workspace directory and sources customization options

parent 607631aa

src/tdoc_crawler/workspaces/crud.py

+30 −3
Original line number Diff line number Diff line
@@ -2,6 +2,8 @@ 


from __future__ import annotations



import shutil



from tdoc_crawler.config import resolve_cache_manager

from tdoc_crawler.config.workspace_registry import WorkspaceRegistry

from tdoc_crawler.logging import get_logger
@@ -33,13 +35,21 @@ def get_workspace(workspace: str | None) -> WorkspaceRegistry | None: 
    return registry.get_workspace(normalized_workspace)





def create_workspace(workspace: str | None, auto_build: bool = False, description: str = "") -> WorkspaceRegistry:

def create_workspace(

    workspace: str | None,

    auto_build: bool = False,

    description: str = "",

    workspace_dir: str | None = None,

    sources_dirname: str | None = None,

) -> WorkspaceRegistry:

    """Create a workspace entry after canonical normalization.



    Args:

        workspace: Workspace name to create.

        auto_build: When True, newly added members are automatically processed.

        description: Optional workspace description.

        workspace_dir: Custom workspace data directory path.

        sources_dirname: Custom sources subdirectory name (default: "sources").



    Returns:

        WorkspaceRegistry with created workspace.
@@ -47,7 +57,13 @@ def create_workspace(workspace: str | None, auto_build: bool = False, descriptio 
    normalized_workspace = normalize_workspace_name(workspace)

    registry = _get_registry()

    try:

        registry.create_workspace(normalized_workspace, description=description, auto_build=auto_build)

        registry.create_workspace(

            normalized_workspace,

            description=description,

            auto_build=auto_build,

            workspace_dir=workspace_dir,

            sources_dirname=sources_dirname,

        )

        registry.save()

    except ValueError:

        pass  # Workspace already exists
@@ -74,11 +90,12 @@ def list_workspaces() -> list[WorkspaceRegistry]: 
    return list(registry.workspaces.values())





def delete_workspace(workspace: str | None) -> bool:

def delete_workspace(workspace: str | None, delete_artifacts: bool = False) -> bool:

    """Delete a workspace.



    Args:

        workspace: Workspace name to delete.

        delete_artifacts: When True, remove workspace directory and files.



    Returns:

        True if deleted, False if not found or if attempting to delete default.
@@ -93,6 +110,16 @@ def delete_workspace(workspace: str | None) -> bool: 
        _logger.warning("Workspace '%s' not found — nothing to delete", normalized_workspace)

        return False



    if delete_artifacts:

        try:

            manager = resolve_cache_manager()

            ws_dir = manager.workspace_llm_wiki_dir(normalized_workspace).parent

            if ws_dir.exists():

                shutil.rmtree(ws_dir)

                _logger.info("Deleted workspace artifacts at '%s'", ws_dir)

        except Exception:

            _logger.debug("Skipping artifact deletion: CacheManager not available")



    deleted = registry.delete_workspace(normalized_workspace)

    if not deleted:

        return False