coulomb-social/coulomb_social/apps/spaces/import_tree.py

"""Import an ADR-0004 space tree into the content plane + Space index.

Source layout (export or hand-authored):

    <tree>/
      pages/*.md
      assets/*   (optional)

Does not fetch Bubble; only copies local files. Bodies may be private —
callers should keep source trees out of git.
"""

from __future__ import annotations

import re
import shutil
from dataclasses import dataclass, field
from pathlib import Path

from django.db import transaction
from django.utils.text import slugify

from coulomb_social.apps.members.models import Member

from . import pageops
from .models import Space, SpaceMembership

_SLUG_SAFE = re.compile(r"[^a-z0-9]+")


@dataclass
class ImportReport:
    space_slug: str
    title: str
    pages_copied: int = 0
    assets_copied: int = 0
    space_created: bool = False
    space_updated: bool = False
    content_path: str = ""
    warnings: list[str] = field(default_factory=list)
    page_slugs: list[str] = field(default_factory=list)


def _read_index_meta(pages: Path) -> tuple[str, str, str]:
    index = pages / "index.md"
    if not index.is_file():
        return "", "", ""
    text = index.read_text(encoding="utf-8")
    meta, _ = pageops.parse_frontmatter(text)
    title = str(meta.get("title") or "").strip()
    abstractor = str(meta.get("abstractor") or meta.get("description") or "").strip()
    visual = str(meta.get("visual") or "").strip()
    return title, abstractor, visual


def normalize_import_slug(value: str) -> str:
    s = slugify(value or "")[:80]
    if not s:
        raise pageops.PageOpsError("Space slug cannot be empty")
    return s


def import_space_tree(
    tree: Path,
    *,
    tenant_id: str,
    slug: str | None = None,
    member: Member | None = None,
    title: str | None = None,
    replace: bool = False,
) -> ImportReport:
    """Copy tree into CONTENT_ROOT and upsert Space row.

    If ``replace`` is False and the destination pages dir already has files,
    raises PageOpsError unless the tree is empty.
    """
    tree = tree.expanduser().resolve()
    if not tree.is_dir():
        raise pageops.PageOpsError(f"Tree not found: {tree}")
    pages_src = tree / "pages"
    if not pages_src.is_dir():
        raise pageops.PageOpsError(f"Missing pages/ under {tree}")

    idx_title, idx_abs, idx_vis = _read_index_meta(pages_src)
    space_slug = normalize_import_slug(slug or tree.name)
    display_title = (title or idx_title or space_slug).strip()
    abstractor = idx_abs
    visual = idx_vis

    dest = pageops.space_dir(space_slug)
    dest_pages = pageops.pages_dir(space_slug)
    dest_assets = pageops.assets_dir(space_slug)

    if dest_pages.is_dir() and any(dest_pages.glob("*.md")) and not replace:
        raise pageops.PageOpsError(
            f"Destination already has pages at {dest_pages}; pass replace=True to overwrite"
        )

    report = ImportReport(space_slug=space_slug, title=display_title)

    if replace and dest.is_dir():
        shutil.rmtree(dest)

    pageops.ensure_space_tree(space_slug)

    # Copy pages
    for path in sorted(pages_src.glob("*.md")):
        target = dest_pages / path.name
        shutil.copy2(path, target)
        report.pages_copied += 1
        report.page_slugs.append(path.stem)

    # Copy assets if present
    assets_src = tree / "assets"
    if assets_src.is_dir():
        for path in sorted(assets_src.rglob("*")):
            if not path.is_file():
                continue
            rel = path.relative_to(assets_src)
            target = dest_assets / rel
            target.parent.mkdir(parents=True, exist_ok=True)
            shutil.copy2(path, target)
            report.assets_copied += 1

    if not (dest_pages / "index.md").is_file():
        pageops.init_space_content(
            space_slug,
            title=display_title,
            abstractor=abstractor,
            visual=visual,
        )
        report.warnings.append("index.md missing in source; synthesized")
        report.pages_copied += 1
        report.page_slugs.append("index")

    report.content_path = str(dest)

    with transaction.atomic():
        space, created = Space.objects.update_or_create(
            tenant_id=tenant_id,
            slug=space_slug,
            defaults={
                "title": display_title,
                "description": abstractor,
                "visual": visual,
                "is_active": True,
                "content_root": "pages",
                "default_branch": "main",
            },
        )
        report.space_created = created
        report.space_updated = not created
        if member is not None:
            SpaceMembership.objects.get_or_create(
                space=space,
                member=member,
                defaults={"role": SpaceMembership.Role.OWNER},
            )
            if space.created_by_id is None:
                space.created_by = member
                space.save(update_fields=["created_by", "updated_at"])

    # Sanity: count readable pages
    try:
        listed = pageops.list_pages(space_slug)
        if len(listed) != report.pages_copied:
            report.warnings.append(
                f"list_pages returned {len(listed)} vs copied {report.pages_copied}"
            )
    except pageops.PageOpsError as exc:
        report.warnings.append(f"list_pages failed: {exc.message}")

    return report