Complete publish preview Phase 2A with retention, refresh and cache consistency

backend/app/core/config.py

@@ -35,6 +35,7 @@ class Settings(BaseSettings):
     svg_display_technical_text_patterns: str = "debug,tech,helper,tmp,service"
 
     storage_root_dir: str = "/data"
+    publish_preview_retention_per_variant: int = 2
 
     model_config = SettingsConfigDict(
         env_file=".env",

backend/app/repositories/scheme_artifacts.py

@@ -1,6 +1,6 @@
 from uuid import uuid4
 
-from sqlalchemy import asc, desc, select
+from sqlalchemy import asc, delete, desc, select
 
 from app.db.session import AsyncSessionLocal
 from app.models.scheme_artifact import SchemeArtifactRecord
@@ -99,3 +99,16 @@ async def get_latest_scheme_artifact(
 
         result = await session.execute(stmt)
         return result.scalar_one_or_none()
+
+
+async def delete_scheme_artifacts_by_artifact_ids(artifact_ids: list[str]) -> int:
+    if not artifact_ids:
+        return 0
+
+    async with AsyncSessionLocal() as session:
+        stmt = delete(SchemeArtifactRecord).where(
+            SchemeArtifactRecord.artifact_id.in_(artifact_ids)
+        )
+        result = await session.execute(stmt)
+        await session.commit()
+        return int(result.rowcount or 0)

backend/app/services/publish_preview.py

@@ -18,6 +18,24 @@ from app.services.scheme_validation import build_scheme_validation_report
 from app.services.structure_diff import build_structure_diff
 
 
+def _serialize_artifacts(artifacts_rows: list) -> dict:
+    return {
+        "total": len(artifacts_rows),
+        "items": [
+            {
+                "artifact_id": row.artifact_id,
+                "artifact_type": row.artifact_type,
+                "artifact_variant": row.artifact_variant,
+                "status": row.status,
+                "storage_path": row.storage_path,
+                "meta_json": row.meta_json,
+                "created_at": row.created_at.isoformat(),
+            }
+            for row in artifacts_rows
+        ],
+    }
+
+
 async def build_publish_preview_bundle(
     *,
     scheme_id: str,
@@ -62,21 +80,7 @@ async def build_publish_preview_bundle(
         except Exception:
             unpriced += 1
 
-    artifacts = {
-        "total": len(artifacts_rows),
-        "items": [
-            {
-                "artifact_id": row.artifact_id,
-                "artifact_type": row.artifact_type,
-                "artifact_variant": row.artifact_variant,
-                "status": row.status,
-                "storage_path": row.storage_path,
-                "meta_json": row.meta_json,
-                "created_at": row.created_at.isoformat(),
-            }
-            for row in artifacts_rows
-        ],
-    }
+    artifacts = _serialize_artifacts(artifacts_rows)
 
     pricing_coverage = {
         "snapshot_available": snapshot_available,
@@ -126,10 +130,21 @@ async def get_or_build_publish_preview_bundle(
         scheme_version_id=scheme_version_id,
         baseline_override_scheme_version_id=baseline_override_scheme_version_id,
     )
-    await save_publish_preview_artifact(
+
+    save_result = await save_publish_preview_artifact(
         scheme_id=scheme_id,
         scheme_version_id=scheme_version_id,
         payload=payload,
         baseline_scheme_version_id=payload["structure_diff"]["baseline_scheme_version_id"],
     )
+
+    artifacts_rows = await list_scheme_artifacts(scheme_version_id=scheme_version_id)
+    payload["artifacts"] = _serialize_artifacts(artifacts_rows)
+    payload["summary"]["has_artifacts"] = payload["artifacts"]["total"] > 0
+    payload["summary"]["preview_cache_cleanup"] = save_result["cleanup"]
+
+    artifact = save_result["artifact"]
+    path = Path(artifact.storage_path)
+    path.write_text(json.dumps(payload, ensure_ascii=False, indent=2), encoding="utf-8")
+
     return payload

backend/app/services/publish_preview_cache.py

@@ -5,7 +5,11 @@ from pathlib import Path
 from uuid import uuid4
 
 from app.core.config import settings
-from app.repositories.scheme_artifacts import create_scheme_artifact, list_scheme_artifacts
+from app.repositories.scheme_artifacts import (
+    create_scheme_artifact,
+    delete_scheme_artifacts_by_artifact_ids,
+    list_scheme_artifacts,
+)
 
 
 def _preview_storage_dir() -> Path:
@@ -14,6 +18,64 @@ def _preview_storage_dir() -> Path:
     return path
 
 
+def _cleanup_preview_file(storage_path: str) -> None:
+    path = Path(storage_path)
+    try:
+        if path.exists() and path.is_file():
+            path.unlink()
+    except FileNotFoundError:
+        pass
+
+    parent = path.parent
+    preview_root = Path(settings.storage_preview_dir)
+
+    try:
+        if parent != preview_root and parent.exists():
+            parent.rmdir()
+    except OSError:
+        pass
+
+
+async def cleanup_publish_preview_artifacts(
+    *,
+    scheme_version_id: str,
+    baseline_scheme_version_id: str | None,
+) -> dict:
+    retention = max(1, settings.publish_preview_retention_per_variant)
+    variant = baseline_scheme_version_id or "default"
+
+    rows = await list_scheme_artifacts(
+        scheme_version_id=scheme_version_id,
+        artifact_type="publish_preview",
+        artifact_variant=variant,
+    )
+    if len(rows) <= retention:
+        return {
+            "retention": retention,
+            "deleted_count": 0,
+            "deleted_artifact_ids": [],
+        }
+
+    rows_sorted = sorted(
+        rows,
+        key=lambda row: (row.created_at, row.id),
+        reverse=True,
+    )
+    to_delete = rows_sorted[retention:]
+
+    deleted_artifact_ids = [row.artifact_id for row in to_delete]
+    for row in to_delete:
+        _cleanup_preview_file(row.storage_path)
+
+    deleted_count = await delete_scheme_artifacts_by_artifact_ids(deleted_artifact_ids)
+
+    return {
+        "retention": retention,
+        "deleted_count": deleted_count,
+        "deleted_artifact_ids": deleted_artifact_ids,
+    }
+
+
 async def save_publish_preview_artifact(
     *,
     scheme_id: str,
@@ -37,7 +99,16 @@ async def save_publish_preview_artifact(
             "summary": payload.get("summary"),
         },
     )
-    return artifact
+
+    cleanup = await cleanup_publish_preview_artifacts(
+        scheme_version_id=scheme_version_id,
+        baseline_scheme_version_id=baseline_scheme_version_id,
+    )
+
+    return {
+        "artifact": artifact,
+        "cleanup": cleanup,
+    }
 
 
 async def get_latest_publish_preview_artifact(
@@ -45,15 +116,13 @@ async def get_latest_publish_preview_artifact(
     scheme_version_id: str,
     baseline_scheme_version_id: str | None,
 ):
-    rows = await list_scheme_artifacts(scheme_version_id=scheme_version_id)
-    variant = baseline_scheme_version_id or "default"
-
-    matching = [
-        row for row in rows
-        if row.artifact_type == "publish_preview" and row.artifact_variant == variant
-    ]
-    if not matching:
+    rows = await list_scheme_artifacts(
+        scheme_version_id=scheme_version_id,
+        artifact_type="publish_preview",
+        artifact_variant=baseline_scheme_version_id or "default",
+    )
+    if not rows:
         return None
 
-    matching.sort(key=lambda row: (row.created_at, row.id), reverse=True)
-    return matching[0]
+    rows.sort(key=lambda row: (row.created_at, row.id), reverse=True)
+    return rows[0]

backend/docs/publish-preview-phase-2a-status.md

@@ -0,0 +1,30 @@
+# Publish Preview — Phase 2A Status
+
+## Status
+
+Phase 2A is mostly closed.
+
+## Confirmed
+
+- retention policy works
+- `refresh=true` works
+- cached read works
+- response contract is more consistent
+- cleanup actually removes old records from the database
+
+## Retention result
+
+For the current draft version, exactly 2 preview artifacts remain in the database, matching the configured retention policy.
+
+## Response contract
+
+The response now explicitly returns:
+
+- `preview_cache_cleanup.retention=2`
+- `deleted_count`
+- `deleted_artifact_ids`
+
+## Runtime state
+
+- logs are clean
+- no new exceptions were observed