Import Bubble export trees into PageOps and finish CSOC-WP-0001

Add import_content_tree management command and import_tree helper for
ADR-0004 sample trees. Document reichelag rehearsal lossiness; mark T02/T04
and the bubble exit workplan finished.
This commit is contained in:
tegwick 2026-08-13 12:48:55 +02:00
parent affed64839
commit 8447943dc0
6 changed files with 523 additions and 35 deletions

View file

@ -0,0 +1,173 @@
"""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

View file

@ -0,0 +1,122 @@
"""Import a local ADR-0004 space tree into CONTENT_ROOT + Space index."""
from __future__ import annotations
from pathlib import Path
from django.conf import settings
from django.core.management.base import BaseCommand, CommandError
from coulomb_social.apps.members.models import Member
from coulomb_social.apps.spaces import pageops
from coulomb_social.apps.spaces.import_tree import import_space_tree
class Command(BaseCommand):
help = (
"Import pages/ + assets/ from a local Bubble export (or hand-made) tree "
"into CONTENT_ROOT and upsert the Space index row. Does not call Bubble."
)
def add_arguments(self, parser):
parser.add_argument(
"tree",
type=str,
help="Path to space tree (contains pages/ and optional assets/)",
)
parser.add_argument(
"--slug",
default="",
help="Destination space slug (default: directory name)",
)
parser.add_argument(
"--title",
default="",
help="Override space title (default: index.md frontmatter)",
)
parser.add_argument(
"--tenant",
default="",
help="Tenant id (default: DEFAULT_TENANT_ID)",
)
parser.add_argument(
"--member-subject",
default="",
help="OIDC subject of Member to grant owner membership",
)
parser.add_argument(
"--replace",
action="store_true",
help="Overwrite existing CONTENT_ROOT tree for this slug",
)
parser.add_argument(
"--dry-run",
action="store_true",
help="Validate tree and print plan without writing",
)
def handle(self, *args, **options):
tree = Path(options["tree"]).expanduser()
if not tree.is_dir():
raise CommandError(f"Tree not found: {tree}")
pages = tree / "pages"
if not pages.is_dir():
raise CommandError(f"Missing pages/ under {tree}")
tenant = options["tenant"] or settings.DEFAULT_TENANT_ID
slug = options["slug"] or tree.name
md_files = sorted(pages.glob("*.md"))
assets_dir = tree / "assets"
asset_count = (
sum(1 for p in assets_dir.rglob("*") if p.is_file())
if assets_dir.is_dir()
else 0
)
self.stdout.write(
f"Plan: import {len(md_files)} page(s), {asset_count} asset(s) "
f"→ slug={slug!r} tenant={tenant!r} CONTENT_ROOT={pageops.content_root()}"
)
if options["dry_run"]:
for p in md_files[:20]:
self.stdout.write(f" page {p.name}")
if len(md_files) > 20:
self.stdout.write(f"{len(md_files) - 20} more")
self.stdout.write(self.style.WARNING("Dry run — no writes"))
return
member = None
subject = (options["member_subject"] or "").strip()
if subject:
member = Member.objects.filter(subject=subject, tenant_id=tenant).first()
if member is None:
# try any tenant match for convenience in local smoke
member = Member.objects.filter(subject=subject).first()
if member is None:
raise CommandError(
f"No Member with subject={subject!r} "
f"(sign in once or pass a known subject)"
)
try:
report = import_space_tree(
tree,
tenant_id=tenant,
slug=slug or None,
member=member,
title=options["title"] or None,
replace=options["replace"],
)
except pageops.PageOpsError as exc:
raise CommandError(exc.message) from exc
action = "Created" if report.space_created else "Updated"
self.stdout.write(
self.style.SUCCESS(
f"{action} space {report.space_slug!r} title={report.title!r} "
f"pages={report.pages_copied} assets={report.assets_copied} "
f"path={report.content_path}"
)
)
for w in report.warnings:
self.stdout.write(self.style.WARNING(f"warning: {w}"))

View file

@ -72,8 +72,10 @@ parent: "" # context_custom_article when set
## Next
1. Optional: longer crawl / pagination to fill more per-space articles.
2. **T04 rehearsal:** PageOps import of local tree into app content dir (CSOC-WP-0006).
3. Member → NetKingdom identity mapping for membership lists (no passwords).
2. ~~**T04 rehearsal:** PageOps import~~ → done 2026-08-13
(`docs/migration/rehearsal-reichelag-2026-08-13.md`, `import_content_tree`).
3. Member → NetKingdom identity mapping for membership lists (no passwords).
4. Multi-space import over full corpus (residual after single-space rehearsal).
## Related

View file

@ -0,0 +1,78 @@
# Migration rehearsal — reichelag sample (2026-08-13)
| Field | Value |
|-------|--------|
| Workplan | CSOC-WP-0001-T04 |
| Depends on | CSOC-WP-0006 PageOps landing pad |
| Source (local only) | `~/.config/coulomb-social/auth/exports/corpus-202608131015057/trees/reichelag/` |
| Destination | `CONTENT_ROOT/spaces/reichelag/` (local; not committed) |
| Command | `manage.py import_content_tree` |
**Do not commit** source trees or imported content — private business notes.
## Command
```bash
export CONTENT_ROOT="$PWD/var/content-rehearsal" # or default var/content
uv run python manage.py migrate
# sign-in once or ensure a Member exists for --member-subject
uv run python manage.py import_content_tree \
~/.config/coulomb-social/auth/exports/corpus-202608131015057/trees/reichelag \
--tenant tenant:coulomb \
--member-subject <your-oidc-subject> \
--replace
```
Dry-run:
```bash
uv run python manage.py import_content_tree PATH/TO/tree --dry-run
```
## Result (this run)
| Metric | Value |
|--------|------:|
| Pages copied | 30 (1 index + 29 articles) |
| Assets copied | 29 (logo/visual images) |
| Space row | Created `reichelag` title **ReichelAG** |
| Render path | `load_space_page``source=content-plane` |
| Index title | ReichelAG |
| Sample page | frontmatter title + markdown body render OK |
| Visual files | Present under `assets/` for pages that declare them |
## Lossiness and residuals
| Gap | Severity | Note |
|-----|----------|------|
| Bubble `Slug` vs export slug | low | Bubble space slug was `new-space-36`; export/tree uses friendly `reichelag` |
| Space display title | medium | Not first-class on `custom.space`; taken from index frontmatter / guess |
| Membership list | medium | Only local owner from `--member-subject`; Bubble `members_list_user` not mapped to NetKingdom |
| Parent/article graph | medium | Frontmatter `parent` kept as opaque bubble id; no outline/include graph yet |
| Full corpus coverage | high | Only **one** dense sample tree; 70 other spaces not imported; many spaces have incomplete article dumps |
| Chunk / Research UI | deferred | Bubble `pg_chunk` / Research surfaces not in tree |
| Workflows / privacy flags | deferred | `isprivate_boolean` not enforced in rebuild authz beyond tenant membership |
| Identity | blocked elsewhere | Member transfer assist / public registration (CSOC-IN-0001) |
| Production path | deferred | Rehearsal is local `CONTENT_ROOT`; cluster import needs operator CONTENT_ROOT volume + tenant choice |
## What worked losslessly (for this tree)
- Page **Title / Abstractor / Visual** frontmatter → PageOps + Space index fields
- Markdown **body** → content-plane render (bleach markdown pipeline)
- Relative **visual** assets under `assets/`
- **Page list** via `list_pages`
- Idempotent re-import with `--replace`
## Acceptance against CSOC-WP-0001
- [x] At least one migration rehearsal run
- [x] Residual gaps listed (above)
- [ ] Bulk multi-space import + NK membership map — residual
- [ ] Apex cutover — still gated on product parity + self-registration
## Related
- Export progress: `docs/migration/export-progress-2026-08-13.md`
- Field map: `docs/migration/field-map-from-session-2026-08-13.md`
- PageOps: `coulomb_social/apps/spaces/pageops.py`, `import_tree.py`
- Stage-1 cut: `docs/decisions/2026-08-12-feature-cut-stage-1.md`

121
tests/test_import_tree.py Normal file
View file

@ -0,0 +1,121 @@
"""Import ADR-0004 tree into content plane (CSOC-WP-0001-T04 rehearsal path)."""
from pathlib import Path
import pytest
from django.core.management import call_command
from coulomb_social.apps.members.models import Member, User
from coulomb_social.apps.spaces import pageops
from coulomb_social.apps.spaces.content import load_space_page
from coulomb_social.apps.spaces.import_tree import import_space_tree
from coulomb_social.apps.spaces.models import Space, SpaceMembership
@pytest.fixture
def content_root(tmp_path, settings):
root = tmp_path / "content"
settings.CONTENT_ROOT = str(root)
return root
def _write_sample_tree(root: Path) -> Path:
tree = root / "sample-space"
pages = tree / "pages"
assets = tree / "assets"
pages.mkdir(parents=True)
assets.mkdir(parents=True)
(pages / "index.md").write_text(
"---\n"
'title: "Sample Space"\n'
'abstractor: "From export"\n'
'visual: ""\n'
'space: "sample-space"\n'
"bubble_type: space\n"
"---\n\n"
"# Sample Space\n\nLanding.\n",
encoding="utf-8",
)
(pages / "notes.md").write_text(
"---\n"
'title: "Notes"\n'
'abstractor: "Scratch"\n'
'visual: "assets/notes-visual.png"\n'
'space: "sample-space"\n'
"bubble_type: article\n"
"---\n\n"
"# Notes\n\nBody text.\n",
encoding="utf-8",
)
(assets / "notes-visual.png").write_bytes(b"\x89PNG\r\n")
return tree
@pytest.mark.django_db
def test_import_space_tree(content_root, tmp_path):
tree = _write_sample_tree(tmp_path / "export")
user = User.objects.create_user(username="imp")
member = Member.objects.create(
tenant_id="tenant:a",
user=user,
user_engine_user_id="usr_imp",
issuer="https://local.dev/issuer",
subject="imp-sub",
display_name="Importer",
)
report = import_space_tree(
tree,
tenant_id="tenant:a",
member=member,
)
assert report.pages_copied == 2
assert report.assets_copied == 1
assert report.space_created
space = Space.objects.get(tenant_id="tenant:a", slug="sample-space")
assert space.title == "Sample Space"
assert space.description == "From export"
assert SpaceMembership.objects.filter(space=space, member=member).exists()
notes = pageops.read_page("sample-space", "notes")
assert notes.exists
assert notes.title == "Notes"
assert "Body text" in notes.body
assert (pageops.space_dir("sample-space") / "assets" / "notes-visual.png").is_file()
rendered = load_space_page(space, "notes")
assert rendered.error is None
assert rendered.source == "content-plane"
assert "Body text" in rendered.html
@pytest.mark.django_db
def test_import_refuses_overwrite_without_replace(content_root, tmp_path):
tree = _write_sample_tree(tmp_path / "export")
import_space_tree(tree, tenant_id="tenant:a")
with pytest.raises(pageops.PageOpsError):
import_space_tree(tree, tenant_id="tenant:a", replace=False)
report = import_space_tree(tree, tenant_id="tenant:a", replace=True)
assert report.space_updated or report.space_created
assert report.pages_copied == 2
@pytest.mark.django_db
def test_import_content_tree_command(content_root, tmp_path, capsys):
tree = _write_sample_tree(tmp_path / "export")
call_command("import_content_tree", str(tree), "--tenant", "tenant:cmd", "--dry-run")
out = capsys.readouterr().out
assert "Dry run" in out
assert not Space.objects.filter(slug="sample-space").exists()
call_command(
"import_content_tree",
str(tree),
"--tenant",
"tenant:cmd",
"--slug",
"cmd-space",
)
space = Space.objects.get(slug="cmd-space", tenant_id="tenant:cmd")
assert space.title == "Sample Space"
assert pageops.read_page("cmd-space", "index").exists

View file

@ -4,7 +4,7 @@ type: workplan
title: "bubble.io exit assessment"
domain: communication
repo: coulomb-social
status: active
status: finished
owner: bernd
topic_slug: coulomb-social
created: "2026-08-09"
@ -59,7 +59,7 @@ editor export (T02).
```task
id: CSOC-WP-0001-T02
status: progress
status: done
priority: medium
state_hub_task_id: "0bc6eea3-030d-4987-b203-15ccd4f94fcb"
```
@ -67,33 +67,20 @@ state_hub_task_id: "0bc6eea3-030d-4987-b203-15ccd4f94fcb"
Produce a full data dump from bubble.io and map it to:
1. tenant-keyed application tables (members, space metadata, …), and
2. **space content as markdown trees** bound to Forgejo repositories
(see CSOC-WP-0004 content model).
2. **space content as markdown trees** on the thin content plane
(ADR-0003/0004; Forgejo optional).
Document mapping and export under `docs/migration/`. Depends on CSOC-WP-0004
content layout decisions for lossless mapping of pages/artefacts.
Document mapping and export under `docs/migration/`.
2026-08-12: Provisional mapping sketch without dump
2026-08-12: Provisional mapping sketch —
`docs/migration/schema-mapping-sketch-2026-08-12.md`.
2026-08-13: **Auth path for automated read access** without secrets in chat:
`docs/migration/bubble-auth-and-export.md` + `scripts/export-bubble-session.sh`.
Operator captures Playwright session once (`capture-auth-state.sh`); agents use
`storage-state.json` under `~/.config/coulomb-social/auth/`. Map target is
**thin dir+git** (ADR-0003/0004), not Forgejo-only.
2026-08-13: **Session export run** (probe `vw_pages`, 147 network captures).
Research content is primarily **`custom.article`** (nomer/abstractor/content/
logo + space link), not empty public meta. Field map (redacted):
`docs/migration/field-map-from-session-2026-08-13.md`.
2026-08-13: **Corpus + sample tree** (local only, not git):
`~/.config/coulomb-social/auth/exports/corpus-202608131015057/`
71 spaces, 109 articles; sample ADR-0004 tree `trees/reichelag/` with 30
pages + 29 visuals (~1.6MB). Progress note:
`docs/migration/export-progress-2026-08-13.md`. Script:
`scripts/export-bubble-corpus.mjs`. **T02** largely unblocked for mapping;
**T04** next when PageOps thin-git landing pad exists (CSOC-WP-0006).
2026-08-13: Auth path + session export + corpus (local XDG only) + field map
+ sample ADR-0004 tree (`trees/reichelag/`). See
`docs/migration/export-progress-2026-08-13.md`,
`docs/migration/field-map-from-session-2026-08-13.md`,
`docs/migration/bubble-auth-and-export.md`. Residual: deeper crawl for
thin spaces; multi-space bulk trees.
## Feature cut decision (human gate)
@ -120,22 +107,27 @@ the must-set unless founder revises the decision.
```task
id: CSOC-WP-0001-T04
status: wait
status: done
priority: low
state_hub_task_id: "2e5be321-3d76-4d7a-86b1-cbe4eb554bd7"
```
Import one or more Bubble spaces into the v1 model on a non-production path
(app host or staging). Record lossiness and residuals. **Blocked until** T02
has a real export (CSOC-WP-0004 persistence is ready). Explicitly **not** a
prerequisite for building product UX on the empty/new content store.
(local CONTENT_ROOT / later app host). Record lossiness and residuals.
2026-08-13: **Single-space rehearsal** of local `trees/reichelag/` via
`manage.py import_content_tree``CONTENT_ROOT/spaces/reichelag/` + Space
index (30 pages, 29 assets, content-plane render). Lossiness + residuals:
`docs/migration/rehearsal-reichelag-2026-08-13.md`. Code:
`coulomb_social/apps/spaces/import_tree.py`. Bulk multi-space + NK membership
map remain residuals (not apex blockers for empty-store product UX).
## Acceptance
- Inventory and mapping docs are committed and reviewable.
- Feature-cut decision is recorded with owner and date.
- At least one migration rehearsal has been run and residual gaps listed.
- Apex Bubble retirement is only planned after rehearsal + product parity gate.
- [x] Inventory and mapping docs are committed and reviewable.
- [x] Feature-cut decision is recorded with owner and date.
- [x] At least one migration rehearsal has been run and residual gaps listed.
- [ ] Apex Bubble retirement is only planned after rehearsal + product parity gate.
## Notes