fix(files): a file is added when it is done, not when it was asked for

A download that failed for days and then went through sat days down the library, under its
request date, and read as missing. A job now records when it became done (`ready_at`), stamped
where the expiry already is (`folders.stamp_done`, `finalize_if_running`) and on an insert that is
already done; a check constraint holds every path to done to it. The library sorts and dates its
"Added" by it; the queue keeps the request order.

The migration dates a file from before by its "ready to watch" notice where the inbox still holds
one, else by its request date, where it already sat.
This commit is contained in:
2026-10-12 22:19:37 +02:00
parent 132fb56428
commit 96509fe356
14 changed files with 204 additions and 21 deletions
+5 -4
View File
@@ -16,7 +16,7 @@ file in the TikTok folder would be a lie the folder's name tells.
Retention is the folder's: a source folder keeps a file `download_retention_days` (the value when the
folder was made; the user can change it), Uploads and own folders keep it until it is deleted. A
done file carries its own `expires_at`, stamped by `done_expiry` when it finishes — unless the user
kept it.
kept it — and its `ready_at`, the moment it finished (`stamp_done`).
"""
from datetime import datetime, timedelta, timezone
@@ -136,8 +136,9 @@ def done_expiry(db: Session, folder_id: int | None, kept: bool) -> datetime | No
return None if days is None else datetime.now(timezone.utc) + timedelta(days=days)
def stamp_expiry(db: Session, job: DownloadJob) -> None:
"""Stamp a job that just became done (in this session) with its expiry."""
def stamp_done(db: Session, job: DownloadJob) -> None:
"""Stamp a job that just became done (in this session): when it did, and its expiry."""
job.ready_at = datetime.now(timezone.utc)
job.expires_at = done_expiry(db, job.folder_id, job.kept)
job.expiry_warned = False
@@ -163,7 +164,7 @@ def upload_target(db: Session, user_id: int, folder_id: int) -> FileFolder:
def restamp(db: Session, job: DownloadJob) -> None:
"""Start a file's clock again from now: it moved to another folder, or its user stopped keeping
it. A file not done yet gets its expiry when it finishes (`stamp_expiry`), so it has none now."""
it. A file not done yet gets its expiry when it finishes (`stamp_done`), so it has none now."""
done = job.status == JobStatus.DONE
job.expires_at = done_expiry(db, job.folder_id, job.kept) if done else None
job.expiry_warned = False
+4 -3
View File
@@ -205,7 +205,7 @@ def enqueue(
if asset.status == "ready":
job.status = JobStatus.DONE
job.progress = 100
folders.stamp_expiry(db, job)
folders.stamp_done(db, job)
asset.last_access_at = datetime.now(timezone.utc)
db.commit()
@@ -285,7 +285,7 @@ def enqueue_edit(
if asset.status == "ready": # identical clip already produced for this user → instant
job.status = JobStatus.DONE
job.progress = 100
folders.stamp_expiry(db, job)
folders.stamp_done(db, job)
asset.last_access_at = datetime.now(timezone.utc)
db.commit()
@@ -374,7 +374,7 @@ def enqueue_clip(
if asset.status == "ready": # the same range was already cut for this user → instant
job.status = JobStatus.DONE
job.progress = 100
folders.stamp_expiry(db, job)
folders.stamp_done(db, job)
asset.last_access_at = datetime.now(timezone.utc)
db.commit()
@@ -399,6 +399,7 @@ def finalize_if_running(db: Session, job_id: int, **fields) -> bool:
return False
db.flush() # asset-then-job lock order: push pending asset writes before we lock the job row
if fields.get("status") == JobStatus.DONE:
fields = {**fields, "ready_at": datetime.now(timezone.utc)}
# Done is when a file's retention starts: its folder's, unless the user kept it.
held = db.execute(
select(DownloadJob.folder_id, DownloadJob.kept).where(DownloadJob.id == job_id)
+1 -1
View File
@@ -140,7 +140,7 @@ def ingest_upload(
if folder_id is not None:
job.folder_id = folder_id
if not convert:
folders.stamp_expiry(db, job)
folders.stamp_done(db, job)
service.incr_ref_count(db, asset.id)
db.flush()
audit.record(db, user_id, AuditAction.FILE_UPLOAD, target_type="file", target_id=job.id)
+21 -2
View File
@@ -1,4 +1,4 @@
from datetime import date, datetime
from datetime import date, datetime, timezone
from sqlalchemy import (
BigInteger,
@@ -949,6 +949,13 @@ class FileFolder(Base, TimestampMixin):
retention_days: Mapped[int | None] = mapped_column(Integer)
def _ready_on_insert(context) -> datetime | None:
"""A job inserted already done (an upload kept as it came) was ready the moment it was made; one
that becomes done later is stamped then (`folders.stamp_done`, `service.finalize_if_running`)."""
done = context.get_current_parameters().get("status") == "done"
return datetime.now(timezone.utc) if done else None
class DownloadJob(Base, TimestampMixin, UpdatedAtMixin):
"""A user's request to download one source — the per-user layer over `MediaAsset`.
@@ -967,7 +974,13 @@ class DownloadJob(Base, TimestampMixin, UpdatedAtMixin):
point at the parent download the clip was cut from; `edit_spec` is the (trim/crop) recipe."""
__tablename__ = "download_jobs"
__table_args__ = (Index("ix_download_jobs_claim", "status", "queue_pos"),)
__table_args__ = (
Index("ix_download_jobs_claim", "status", "queue_pos"),
# Every way a job becomes done stamps when (`folders.stamp_done`, `finalize_if_running`).
CheckConstraint(
"status <> 'done' OR ready_at IS NOT NULL", name="ck_download_jobs_ready_at"
),
)
id: Mapped[int] = mapped_column(primary_key=True)
user_id: Mapped[int] = mapped_column(ForeignKey("users.id", ondelete="CASCADE"), index=True)
@@ -1026,6 +1039,12 @@ class DownloadJob(Base, TimestampMixin, UpdatedAtMixin):
kept: Mapped[bool] = mapped_column(Boolean, default=False, server_default="false")
expires_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), index=True)
expiry_warned: Mapped[bool] = mapped_column(Boolean, default=False, server_default="false")
# When the file became done — what the library's "Added" means: a download that failed for days
# and then went through is new to its user, so it sorts as new. `created_at` stays when it was
# asked for (the queue's order). Set again when a file is made anew (a retry, a re-download).
ready_at: Mapped[datetime | None] = mapped_column(
DateTime(timezone=True), default=_ready_on_insert
)
class DownloadShare(Base, TimestampMixin):
+4 -3
View File
@@ -78,6 +78,7 @@ def serialize(job: DownloadJob, asset: MediaAsset | None) -> dict:
"error_detail": job.error_detail,
"queue_pos": job.queue_pos,
"created_at": iso(job.created_at),
"ready_at": iso(job.ready_at),
"source_kind": job.source_kind,
"source_ref": job.source_ref,
"source_url": service.reference_url(job, asset),
@@ -545,10 +546,10 @@ def list_files(
func.coalesce(DownloadJob.display_uploader, MediaAsset.uploader, "").ilike(like),
)
)
# A file shared with the caller was added to THEIR list when it was shared, not when its owner
# downloaded it.
# A file is added to its owner's list when it is done, not when it was asked for (a download that
# failed for days and then went through is new); one shared with the caller, when it was shared.
key = {
"added": DownloadShare.created_at if shared else DownloadJob.created_at,
"added": DownloadShare.created_at if shared else DownloadJob.ready_at,
"name": collated(name, lang_for_user(user)),
"size": MediaAsset.size_bytes,
}[sort]