Enforce one ArchiveResult per plugin

This commit is contained in:
Nick Sweeting 2026-08-28 08:56:07 -07:00
parent 42ee971dfb
commit 43c2dc9896
No known key found for this signature in database
12 changed files with 484 additions and 282 deletions

View File

@ -389,13 +389,6 @@ def _queue_archiveresult_snapshot_maintenance(snapshot: Snapshot) -> None:
snapshot.safe_update(updates, refresh=False)
def _merge_archiveresult_output_file_maps(results: list[ArchiveResult]) -> dict[str, dict[str, Any]]:
output_files: dict[str, dict[str, Any]] = {}
for result in results:
output_files.update(result.output_file_map())
return output_files
def _write_archiveresult_files(
request: HttpRequest,
snapshot: Snapshot,
@ -535,15 +528,8 @@ def create_archiveresult(
normalized_status = ArchiveResult.normalize_status(status)
parsed_output_json = _parse_archiveresult_output_json(output_json)
hook = hook_name or ARCHIVERESULT_UPLOAD_HOOK_NAME
matching_results = list(
ArchiveResult.objects.filter(
snapshot=snapshot,
plugin=plugin_name,
hook_name=hook,
).order_by("created_at", "id"),
)
existing_result = matching_results[0] if matching_results else None
existing_output_files = _merge_archiveresult_output_file_maps(matching_results)
existing_result = ArchiveResult.objects.filter(snapshot=snapshot, plugin=plugin_name).first()
existing_output_files = dict(existing_result.output_files or {}) if existing_result else {}
output_files = _write_archiveresult_files(
request,
snapshot,
@ -554,22 +540,13 @@ def create_archiveresult(
now = timezone.now()
with transaction.atomic():
matching_results = list(
ArchiveResult.objects.filter(
snapshot=snapshot,
plugin=plugin_name,
hook_name=hook,
).order_by("created_at", "id"),
)
if matching_results:
existing_result = matching_results[0]
Snapshot.objects.select_for_update().get(pk=snapshot.pk)
existing_result = ArchiveResult.objects.filter(snapshot=snapshot, plugin=plugin_name).first()
if existing_result:
output_files = {
**_merge_archiveresult_output_file_maps(matching_results),
**dict(existing_result.output_files or {}),
**output_files,
}
duplicate_ids = [result.id for result in matching_results[1:]]
if duplicate_ids:
ArchiveResult.objects.filter(id__in=duplicate_ids).delete()
result = existing_result
else:
existing_result = None
@ -588,6 +565,7 @@ def create_archiveresult(
output_size, output_mimetypes = _summarize_archiveresult_output_files(output_files)
output_file_paths = list(output_files.keys())
result.status = normalized_status
result.hook_name = hook
result.output_str = output_str or (output_file_paths[0] if output_file_paths else "")
result.output_json = parsed_output_json
result.output_files = output_files

View File

@ -67,7 +67,7 @@ def create_archiveresults(
Reads Snapshot records from stdin and emits ArchiveResult request JSONL.
Pass-through: Non-Snapshot/ArchiveResult records are output unchanged.
If --plugin is specified, only emits requests for that plugin.
Otherwise, emits requests for all enabled snapshot hooks.
Otherwise, emits one request for each enabled Snapshot plugin.
Exit codes:
0: Success
@ -151,11 +151,9 @@ def create_archiveresults(
else:
config = get_config(crawl=snapshot.crawl, snapshot=snapshot)
hooks = discover_hooks("Snapshot", config=config)
for hook_path in hooks:
hook_name = hook_path.stem
plugin_name = hook_path.parent.name
for plugin_name in dict.fromkeys(hook_path.parent.name for hook_path in hooks):
if not is_tty:
write_record(build_archiveresult_request(snapshot.id, plugin_name, hook_name=hook_name, status=status))
write_record(build_archiveresult_request(snapshot.id, plugin_name, status=status))
created_count += 1
rprint(f"[green]Created {created_count} archive result request records[/green]", file=sys.stderr)
@ -256,9 +254,11 @@ def update_archiveresults(
# Apply updates from CLI flags
if status:
result.status = status
result.save()
if status == ArchiveResult.StatusChoices.QUEUED:
result.reset_for_retry()
else:
result.status = status
result.save(update_fields=["status", "modified_at"])
updated_count += 1
if not is_tty:

View File

@ -122,7 +122,6 @@ def run_plugins(
from archivebox.core.models import Snapshot
from archivebox.core.models import ArchiveResult
from archivebox.services.runner import run_crawl
from abx_dl.models import discover_plugins
is_tty = sys.stdout.isatty()
@ -193,30 +192,21 @@ def run_plugins(
if snapshot_id in existing_snapshot_ids
for plugin_name in plugin_names
)
plugins_by_name = discover_plugins(runtime="archivebox")
requested_rows: set[tuple[str, str, str]] = set()
for snapshot_id, plugin_name in requested_pairs:
plugin = plugins_by_name.get(plugin_name)
hooks = plugin.filter_hooks("Snapshot") if plugin is not None else []
if hooks:
requested_rows.update((snapshot_id, plugin_name, hook.name) for hook in hooks)
else:
requested_rows.add((snapshot_id, plugin_name, ""))
queued_rows: set[tuple[str, str, str]] = set()
if preserve_queued and requested_rows:
queued_rows: set[tuple[str, str]] = set()
if preserve_queued and requested_pairs:
queued_rows = {
(str(snapshot_id), plugin_name, hook_name)
for snapshot_id, plugin_name, hook_name in ArchiveResult.objects.filter(
(str(snapshot_id), plugin_name)
for snapshot_id, plugin_name in ArchiveResult.objects.filter(
snapshot_id__in=existing_snapshot_ids,
plugin__in={plugin_name for _snapshot_id, plugin_name, _hook_name in requested_rows},
plugin__in={plugin_name for _snapshot_id, plugin_name in requested_pairs},
status=ArchiveResult.StatusChoices.QUEUED,
).values_list("snapshot_id", "plugin", "hook_name")
).values_list("snapshot_id", "plugin")
}
rows_to_queue = requested_rows - queued_rows
rows_to_queue = requested_pairs - queued_rows
reset_fields = {
"status": ArchiveResult.StatusChoices.QUEUED,
"hook_name": "",
"output_str": "",
"output_json": None,
"output_files": {},
@ -227,28 +217,28 @@ def run_plugins(
"modified_at": timezone.now(),
}
if rows_to_queue and plugins_list:
rows_to_reset_by_hook: dict[tuple[str, str], set[str]] = defaultdict(set)
for snapshot_id, plugin_name, hook_name in rows_to_queue:
rows_to_reset_by_hook[(plugin_name, hook_name)].add(snapshot_id)
for (plugin_name, hook_name), plugin_snapshot_ids in rows_to_reset_by_hook.items():
ArchiveResult.objects.filter(snapshot_id__in=plugin_snapshot_ids, plugin=plugin_name, hook_name=hook_name).update(
rows_to_reset_by_plugin: dict[str, set[str]] = defaultdict(set)
for snapshot_id, plugin_name in rows_to_queue:
rows_to_reset_by_plugin[plugin_name].add(snapshot_id)
for plugin_name, plugin_snapshot_ids in rows_to_reset_by_plugin.items():
ArchiveResult.objects.filter(snapshot_id__in=plugin_snapshot_ids, plugin=plugin_name).update(
**reset_fields,
)
elif rows_to_queue and requested_plugins_by_snapshot:
snapshot_ids_by_hook: dict[tuple[str, str], set[str]] = defaultdict(set)
for snapshot_id, plugin_name, hook_name in rows_to_queue:
snapshot_ids_by_hook[(plugin_name, hook_name)].add(snapshot_id)
for (plugin_name, hook_name), plugin_snapshot_ids in snapshot_ids_by_hook.items():
ArchiveResult.objects.filter(snapshot_id__in=plugin_snapshot_ids, plugin=plugin_name, hook_name=hook_name).update(
snapshot_ids_by_plugin: dict[str, set[str]] = defaultdict(set)
for snapshot_id, plugin_name in rows_to_queue:
snapshot_ids_by_plugin[plugin_name].add(snapshot_id)
for plugin_name, plugin_snapshot_ids in snapshot_ids_by_plugin.items():
ArchiveResult.objects.filter(snapshot_id__in=plugin_snapshot_ids, plugin=plugin_name).update(
**reset_fields,
)
existing_rows = (
{
(str(snapshot_id), plugin_name, hook_name)
for snapshot_id, plugin_name, hook_name in ArchiveResult.objects.filter(
(str(snapshot_id), plugin_name)
for snapshot_id, plugin_name in ArchiveResult.objects.filter(
snapshot_id__in=existing_snapshot_ids,
plugin__in={plugin_name for _snapshot_id, plugin_name, _hook_name in rows_to_queue},
).values_list("snapshot_id", "plugin", "hook_name")
plugin__in={plugin_name for _snapshot_id, plugin_name in rows_to_queue},
).values_list("snapshot_id", "plugin")
}
if rows_to_queue
else set()
@ -260,10 +250,10 @@ def run_plugins(
ArchiveResult(
snapshot_id=snapshot_id,
plugin=plugin_name,
hook_name=hook_name,
hook_name="",
status=ArchiveResult.StatusChoices.QUEUED,
)
for snapshot_id, plugin_name, hook_name in sorted(missing_rows)
for snapshot_id, plugin_name in sorted(missing_rows)
],
batch_size=500,
)
@ -271,13 +261,13 @@ def run_plugins(
processed_count = len(existing_snapshot_ids)
queue_at = timezone.now()
if existing_snapshot_ids:
if requested_rows:
# Search indexing on a sealed Snapshot is the only targeted hook
if requested_pairs:
# Search indexing on a sealed Snapshot is the only targeted plugin
# allowed to bypass the normal lifecycle. Every other requested
# plugin requeues its Snapshot through the unified state machine.
affected_snapshot_ids = {snapshot_id for snapshot_id, _plugin_name, _hook_name in rows_to_queue}
affected_snapshot_ids = {snapshot_id for snapshot_id, _plugin_name in rows_to_queue}
if preserve_queued and queued_rows:
queued_snapshot_ids = {snapshot_id for snapshot_id, _plugin_name, _hook_name in queued_rows}
queued_snapshot_ids = {snapshot_id for snapshot_id, _plugin_name in queued_rows}
affected_snapshot_ids.update(
str(snapshot_id)
for snapshot_id in Snapshot.objects.filter(id__in=queued_snapshot_ids)
@ -292,7 +282,7 @@ def run_plugins(
)
)
requested_plugins_by_id: dict[str, set[str]] = defaultdict(set)
for snapshot_id, plugin_name, _hook_name in requested_rows:
for snapshot_id, plugin_name in requested_pairs:
requested_plugins_by_id[snapshot_id].add(plugin_name)
for snapshot in Snapshot.objects.filter(id__in=affected_snapshot_ids).only("id", "status", "modified_at"):
# Guard the read-time status so we never bump retry_at on a
@ -328,7 +318,7 @@ def run_plugins(
refresh=False,
extra_filter={"status": snapshot.status},
)
if existing_crawl_ids and not requested_rows:
if existing_crawl_ids and not requested_pairs:
from archivebox.crawls.models import Crawl
for crawl in Crawl.objects.filter(id__in=existing_crawl_ids).only("id", "status", "retry_at", "modified_at"):

View File

@ -0,0 +1,88 @@
from django.db import migrations, models
from django.db.models import Sum
def consolidate_archiveresults_per_plugin(apps, schema_editor):
ArchiveResult = apps.get_model("core", "ArchiveResult")
Snapshot = apps.get_model("core", "Snapshot")
duplicate_groups = ArchiveResult.objects.values("snapshot_id", "plugin").annotate(count=models.Count("id")).filter(count__gt=1)
affected_snapshot_ids = set()
for group in duplicate_groups.iterator(chunk_size=200):
rows = list(
ArchiveResult.objects.filter(
snapshot_id=group["snapshot_id"],
plugin=group["plugin"],
).order_by("created_at", "id"),
)
canonical = rows[0]
winner = max(
rows,
key=lambda row: (
bool(row.output_files),
int(row.output_size or 0),
row.modified_at,
str(row.id),
),
)
output_files = {}
mimetypes = set()
for row in rows:
output_files.update(row.output_files or {})
mimetypes.update(part.strip() for part in (row.output_mimetypes or "").split(",") if part.strip())
canonical.hook_name = winner.hook_name
canonical.status = winner.status
canonical.output_str = winner.output_str
canonical.output_json = winner.output_json
canonical.output_files = output_files
canonical.output_size = max(
sum(int(metadata.get("size") or 0) for metadata in output_files.values() if isinstance(metadata, dict)),
*(int(row.output_size or 0) for row in rows),
)
canonical.output_mimetypes = ",".join(sorted(mimetypes))
canonical.start_ts = min((row.start_ts for row in rows if row.start_ts), default=None)
canonical.end_ts = max((row.end_ts for row in rows if row.end_ts), default=None)
canonical.save(
update_fields=[
"hook_name",
"status",
"output_str",
"output_json",
"output_files",
"output_size",
"output_mimetypes",
"start_ts",
"end_ts",
],
)
ArchiveResult.objects.filter(id__in=[row.id for row in rows[1:]]).delete()
affected_snapshot_ids.add(group["snapshot_id"])
for snapshot_id in affected_snapshot_ids:
total = ArchiveResult.objects.filter(snapshot_id=snapshot_id).aggregate(total=Sum("output_size"))["total"] or 0
Snapshot.objects.filter(id=snapshot_id).update(output_size=total)
class Migration(migrations.Migration):
dependencies = [
("core", "0051_postgres_url_pattern_ops_index"),
]
operations = [
migrations.RemoveConstraint(
model_name="archiveresult",
name="unique_archiveresult_per_snapshot_hook",
),
migrations.RunPython(
consolidate_archiveresults_per_plugin,
reverse_code=migrations.RunPython.noop,
),
migrations.AddConstraint(
model_name="archiveresult",
constraint=models.UniqueConstraint(
fields=("snapshot", "plugin"),
name="unique_archiveresult_per_snapshot_plugin",
),
),
]

View File

@ -1740,8 +1740,8 @@ class Snapshot(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithConfig, ModelW
self.add_tag_ids([tag.pk])
def _merge_archive_results_from_index(self, index_data: dict, update_existing: bool = True):
"""Merge ArchiveResults one row per hook; retries update the existing row."""
existing = {(ar.plugin, ar.hook_name): ar for ar in ArchiveResult.objects.filter(snapshot=self)}
"""Merge ArchiveResults one row per plugin; retries update the existing row."""
existing = {ar.plugin: ar for ar in ArchiveResult.objects.filter(snapshot=self)}
if update_existing:
for archiveresult in existing.values():
normalized_status = ArchiveResult.normalize_status(archiveresult.status)
@ -1804,7 +1804,7 @@ class Snapshot(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithConfig, ModelW
output_mimetypes = result_data.get("output_mimetypes", "")
hook_name = result_data.get("hook_name", "")
existing_result = existing.get((plugin, hook_name))
existing_result = existing.get(plugin)
if existing_result:
if not update_existing:
return
@ -1813,6 +1813,9 @@ class Snapshot(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithConfig, ModelW
if existing_result.status != status:
existing_result.status = status
update_fields.append("status")
if hook_name and existing_result.hook_name != hook_name:
existing_result.hook_name = hook_name
update_fields.append("hook_name")
if output_str and existing_result.output_str != output_str:
existing_result.output_str = output_str
update_fields.append("output_str")
@ -1869,7 +1872,7 @@ class Snapshot(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithConfig, ModelW
end_ts=end_ts,
process=process,
)
existing[(plugin, hook_name)] = archiveresult
existing[plugin] = archiveresult
def write_index_json(self):
"""Write index.json in 0.9.x format (deprecated, use write_index_jsonl)."""
@ -2199,11 +2202,42 @@ class Snapshot(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithConfig, ModelW
for tag in dup.tags.all():
keeper.add_tag_ids([tag.pk])
# Move ArchiveResults
ArchiveResult.objects.filter(snapshot=dup).update(
snapshot=keeper,
modified_at=timezone.now(),
)
# Move ArchiveResults while preserving the one-row-per-plugin invariant.
for result in ArchiveResult.objects.filter(snapshot=dup):
existing = ArchiveResult.objects.filter(snapshot=keeper, plugin=result.plugin).first()
if existing is None:
result.snapshot = keeper
result.save(update_fields=["snapshot", "modified_at"])
continue
prefer_result = result.output_size > existing.output_size
prior_output_sizes = (existing.output_size, result.output_size)
output_files = {**(existing.output_files or {}), **(result.output_files or {})}
existing.output_files = output_files
existing.output_size = max(
sum(
ArchiveResult._coerce_output_file_size(metadata.get("size"))
for metadata in output_files.values()
if isinstance(metadata, dict)
),
*prior_output_sizes,
)
if prefer_result:
existing.status = result.status
existing.hook_name = result.hook_name
existing.output_str = result.output_str
existing.output_json = result.output_json
existing.output_mimetypes = ",".join(
sorted(
{
mimetype.strip()
for value in (existing.output_mimetypes, result.output_mimetypes)
for mimetype in value.split(",")
if mimetype.strip()
},
),
)
existing.save()
result.delete()
# Delete
dup.delete()
@ -2942,8 +2976,8 @@ class Snapshot(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithConfig, ModelW
- abx_plugins/plugins/*/on_Snapshot__*.{py,sh,js}
- data/custom_plugins/*/on_Snapshot__*.{py,sh,js}
Creates one ArchiveResult per hook (not per plugin), with hook_name set.
This enables step-based execution where all hooks in a step can run in parallel.
Creates one ArchiveResult per plugin. A queued plugin row runs all of that
plugin's Snapshot hooks; hook_name records the most recently completed hook.
"""
try:
self.validate_url_for_archiving()
@ -2961,17 +2995,21 @@ class Snapshot(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithConfig, ModelW
hooks = ((hook_path.parent.name, hook_path.stem) for hook_path in discover_hooks("Snapshot", config=config))
archiveresults = []
for plugin, hook_name in hooks:
# ArchiveResult output is one filesystem directory per plugin hook, so
# retries must update this row in place instead of creating siblings.
for plugin in dict.fromkeys(plugin for plugin, _hook_name in hooks):
# ArchiveResult output is one filesystem directory per plugin, so all
# hooks and retries update this row in place instead of creating siblings.
archiveresult, _created = ArchiveResult.objects.get_or_create(
snapshot=self,
plugin=plugin,
hook_name=hook_name,
defaults={
"hook_name": "",
"status": ArchiveResult.INITIAL_STATE,
},
)
if archiveresult.hook_name == self.BROWSER_EXTENSION_UPLOAD_HOOK_NAME:
archiveresult.hook_name = ""
archiveresult.status = ArchiveResult.INITIAL_STATE
archiveresult.save(update_fields=["hook_name", "status", "modified_at"])
if archiveresult.status == ArchiveResult.INITIAL_STATE:
archiveresults.append(archiveresult)
@ -3060,10 +3098,10 @@ class Snapshot(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithConfig, ModelW
snapshot=self,
status=ArchiveResult.StatusChoices.FAILED,
)
legacy_result_count = retryable_results.filter(hook_name="").count()
now = timezone.now()
count = retryable_results.exclude(hook_name="").update(
count = retryable_results.update(
status=ArchiveResult.StatusChoices.QUEUED,
hook_name="",
output_str="",
output_json=None,
output_files={},
@ -3074,11 +3112,11 @@ class Snapshot(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithConfig, ModelW
modified_at=now,
)
if count + legacy_result_count > 0:
if count > 0:
self.refresh_from_db(fields=["modified_at", "retry_at", "status"])
self.queue_for_extraction(when=now)
return count + legacy_result_count
return count
# =========================================================================
# URL Helper Properties (migrated from Link schema)
@ -3928,7 +3966,7 @@ class ArchiveResult(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithNotes):
models.Index(fields=["-start_ts", "-id"], name="archiveresult_start_idx"),
]
constraints: ClassVar[list[models.BaseConstraint]] = [
models.UniqueConstraint(fields=["snapshot", "plugin", "hook_name"], name="unique_archiveresult_per_snapshot_hook"),
models.UniqueConstraint(fields=["snapshot", "plugin"], name="unique_archiveresult_per_snapshot_plugin"),
]
def __str__(self):
@ -4039,16 +4077,16 @@ class ArchiveResult(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithNotes):
except ArchiveResult.DoesNotExist:
pass
# Get or create by snapshot_id + plugin + hook_name. The filesystem has a
# single output dir for each hook, so retries update that same DB row.
# Get or create by snapshot_id + plugin. The filesystem has one output
# directory per plugin, so every hook and retry updates that same DB row.
try:
snapshot = Snapshot.objects.get(id=snapshot_id)
result, _ = ArchiveResult.objects.get_or_create(
snapshot=snapshot,
plugin=plugin,
hook_name=record.get("hook_name", ""),
defaults={
"hook_name": record.get("hook_name", ""),
"status": record.get("status", "queued"),
"output_str": record.get("output_str", ""),
},
@ -4115,13 +4153,10 @@ class ArchiveResult(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithNotes):
def schedule_delete_cleanup(self, *, using: str | None = None) -> None:
"""Remove shared plugin output and refresh persisted Snapshot metadata after commit."""
snapshot_id = self.snapshot_id
plugin = self.plugin
paths = self.validate_output_paths_for_delete(self.output_paths_for_delete())
def cleanup() -> None:
results = type(self).objects.using(using) if using else type(self).objects
if not results.filter(snapshot_id=snapshot_id, plugin=plugin).exists():
type(self).delete_output_paths(paths)
type(self).delete_output_paths(paths)
type(self).refresh_snapshot_output_sizes({snapshot_id})
snapshot = Snapshot.objects.filter(pk=snapshot_id).first()
if snapshot:
@ -4155,6 +4190,7 @@ class ArchiveResult(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithNotes):
def reset_for_retry(self, *, save: bool = True) -> None:
self.status = self.StatusChoices.QUEUED
self.hook_name = ""
self.retry_at = None
self.output_str = ""
self.output_json = None
@ -4167,6 +4203,7 @@ class ArchiveResult(ModelWithDeleteAfter, ModelWithOutputDir, ModelWithNotes):
self.save(
update_fields=[
"status",
"hook_name",
"retry_at",
"output_str",
"output_json",

View File

@ -4,6 +4,7 @@ import asyncio
import inspect
import json
import os
import re
import signal
import sys
import time
@ -202,6 +203,11 @@ def _normalize_status(status: str) -> str:
return status or "failed"
def _snapshot_hook_order(hook_name: str) -> int:
match = re.match(r"^on_Snapshot__(\d+)", hook_name or "")
return int(match.group(1)) if match else -1
def _normalize_snapshot_title(candidate: str, *, snapshot_url: str) -> str:
title = " ".join(line.strip() for line in str(candidate or "").splitlines() if line.strip()).strip()
if not title:
@ -319,6 +325,7 @@ def _save_archiveresult_event_to_db(
start_ts = parse_event_datetime(event.start_ts)
end_ts = parse_event_datetime(event.end_ts) or timezone.now()
defaults = {
"hook_name": event.hook_name,
"status": _normalize_status(event.status),
"output_str": event.output_str,
"output_json": event.output_json,
@ -337,7 +344,6 @@ def _save_archiveresult_event_to_db(
result = ArchiveResult.objects.filter(
snapshot=snapshot,
plugin=event.plugin,
hook_name=event.hook_name,
).first()
if result is None:
try:
@ -345,7 +351,6 @@ def _save_archiveresult_event_to_db(
result = ArchiveResult.objects.create(
snapshot=snapshot,
plugin=event.plugin,
hook_name=event.hook_name,
**defaults,
)
except IntegrityError:
@ -353,9 +358,30 @@ def _save_archiveresult_event_to_db(
result = ArchiveResult.objects.get(
snapshot=snapshot,
plugin=event.plugin,
hook_name=event.hook_name,
)
if result.output_files:
merged_output_files = {**result.output_files, **defaults["output_files"]}
defaults["output_files"] = merged_output_files
defaults["output_size"], defaults["output_mimetypes"] = _summarize_output_files(merged_output_files)
defaults["output_size"] = max(defaults["output_size"], int(result.output_size or 0))
defaults["output_mimetypes"] = ",".join(
dict.fromkeys(
mimetype.strip()
for value in (defaults["output_mimetypes"], result.output_mimetypes)
for mimetype in value.split(",")
if mimetype.strip()
),
)
if result.start_ts and defaults["start_ts"]:
defaults["start_ts"] = min(result.start_ts, defaults["start_ts"])
if result.end_ts and defaults["end_ts"]:
defaults["end_ts"] = max(result.end_ts, defaults["end_ts"])
if _snapshot_hook_order(result.hook_name) > _snapshot_hook_order(event.hook_name):
for field in ("hook_name", "status", "output_str", "output_json", "process_id", "notes"):
if field in result.__dict__:
defaults[field] = result.__dict__[field]
with _perf_span("archivebox.ArchiveResultService.on_ArchiveResultEvent.diff_fields"):
update_fields = []
for field, value in defaults.items():
@ -396,7 +422,7 @@ def _save_archiveresult_event_to_db(
def mark_archiveresult_started(event: ProcessStartedEvent, *, snapshot_id: str, process_id: str) -> None:
"""Advance an existing queued hook row after its OS process is persisted."""
"""Advance an existing queued plugin row after its OS process is persisted."""
from archivebox.core.models import ArchiveResult
started_at = parse_event_datetime(event.start_ts)
@ -405,9 +431,9 @@ def mark_archiveresult_started(event: ProcessStartedEvent, *, snapshot_id: str,
ArchiveResult.objects.filter(
snapshot_id=snapshot_id,
plugin=event.plugin_name,
hook_name=event.hook_name,
status=ArchiveResult.StatusChoices.QUEUED,
).update(
hook_name=event.hook_name,
status=ArchiveResult.StatusChoices.STARTED,
start_ts=started_at,
end_ts=None,
@ -422,7 +448,7 @@ class ArchiveResultService(BaseService):
def __init__(self, bus):
self._completed_process_event_ids: set[str] = set()
self._save_locks: dict[tuple[str, str, str], asyncio.Lock] = {}
self._save_locks: dict[tuple[str, str], asyncio.Lock] = {}
super().__init__(bus)
self.bus.on(ArchiveResultEvent, self.on_ArchiveResultEvent__save_to_db)
self.bus.on(ProcessCompletedEvent, self.on_ProcessCompletedEvent__save_to_db)
@ -437,7 +463,7 @@ class ArchiveResultService(BaseService):
where=lambda candidate: self.bus.event_is_child_of(event, candidate),
)
key = (str(event.snapshot_id), event.plugin, event.hook_name)
key = (str(event.snapshot_id), event.plugin)
lock = self._save_locks.setdefault(key, asyncio.Lock())
async with lock:
await sync_to_async(_save_archiveresult_event_to_db, thread_sensitive=True)(event, process_started)

View File

@ -1038,7 +1038,7 @@ class CrawlRunner:
)
return [plugin for plugin in queued_plugins if plugin in expanded_selected_plugins]
selected_hooks_by_plugin = None
queued_plugins = None
if snapshot["status"] == "started":
_reset_count, running_count = await sync_to_async(snapshot["_snapshot"].reset_abandoned_results, thread_sensitive=True)()
if running_count:
@ -1056,8 +1056,8 @@ class CrawlRunner:
)
return
if not self.selected_plugins_from_args:
queued_plugins, selected_hooks_by_plugin = await sync_to_async(
queued_plugins_and_hooks_for_snapshot,
queued_plugins = await sync_to_async(
queued_plugins_for_snapshot,
thread_sensitive=True,
)(snapshot["id"])
if queued_plugins:
@ -1070,13 +1070,10 @@ class CrawlRunner:
snapshot["id"],
disabled_queued_plugins,
)
selected_hooks_by_plugin = {
plugin: hooks for plugin, hooks in (selected_hooks_by_plugin or {}).items() if plugin in queued_plugins
}
snapshot_selected_plugins = queued_plugins
elif not self.selected_plugins_from_args:
queued_plugins, selected_hooks_by_plugin = await sync_to_async(
queued_plugins_and_hooks_for_snapshot,
queued_plugins = await sync_to_async(
queued_plugins_for_snapshot,
thread_sensitive=True,
)(snapshot["id"])
if queued_plugins:
@ -1089,9 +1086,6 @@ class CrawlRunner:
snapshot["id"],
disabled_queued_plugins,
)
selected_hooks_by_plugin = {
plugin: hooks for plugin, hooks in (selected_hooks_by_plugin or {}).items() if plugin in queued_plugins
}
snapshot_selected_plugins = queued_plugins
if snapshot["depth"] > 0 and CrawlLimitState.from_config(snapshot["config"]).get_stop_reason() in (
"crawl_max_size",
@ -1106,10 +1100,10 @@ class CrawlRunner:
if snapshot_selected_plugins
else self.plugins
)
if selected_hooks_by_plugin is not None:
await sync_to_async(fail_unavailable_queued_hooks, thread_sensitive=True)(
if queued_plugins is not None:
await sync_to_async(fail_unavailable_queued_plugins, thread_sensitive=True)(
snapshot["id"],
selected_hooks_by_plugin,
queued_plugins,
plugins,
)
remaining_queued_plugins = await sync_to_async(
@ -1129,7 +1123,6 @@ class CrawlRunner:
return
snapshot_selected_plugins = remaining_queued_plugins
plugins = filter_plugins(self.plugins, snapshot_selected_plugins, include_providers=True)
selected_hooks_by_plugin = include_background_prerequisite_hooks(selected_hooks_by_plugin, plugins)
abx_snapshot = AbxSnapshot(
id=snapshot["id"],
url=snapshot["url"],
@ -1153,7 +1146,7 @@ class CrawlRunner:
snapshot_cleanup_enabled=True,
snapshot_cleanup_phase_timeout=snapshot_phase_timeout,
abort_requested=self.crawl_is_cancelled,
selected_hooks_by_plugin=selected_hooks_by_plugin,
selected_hooks_by_plugin=None,
)
try:
snapshot_event = SnapshotEvent(
@ -1375,37 +1368,19 @@ def run_binary(binary_id: str) -> None:
asyncio.run(_run_binary(binary_id))
def queued_plugins_and_hooks_for_snapshot(snapshot_id: str) -> tuple[list[str] | None, dict[str, set[str] | None] | None]:
def queued_plugins_for_snapshot(snapshot_id: str) -> list[str] | None:
from archivebox.core.models import ArchiveResult
queued_results = list(
queued_plugins = list(
ArchiveResult.objects.filter(
snapshot_id=snapshot_id,
status=ArchiveResult.StatusChoices.QUEUED,
)
.exclude(plugin="")
.only("id", "plugin", "hook_name"),
.order_by("plugin")
.values_list("plugin", flat=True),
)
selected_hooks_by_plugin: dict[str, set[str] | None] = {}
queued_plugins = sorted({result.plugin for result in queued_results})
for result in queued_results:
# hook_name is the modern scheduler identity. Empty hook_name rows are
# legacy plugin-level work and must keep running the whole plugin.
if not result.hook_name:
selected_hooks_by_plugin[result.plugin] = None
elif result.plugin not in selected_hooks_by_plugin:
selected_hooks_by_plugin[result.plugin] = {result.hook_name}
elif selected_hooks_by_plugin[result.plugin] is not None:
selected_hooks_by_plugin[result.plugin].add(result.hook_name)
if queued_plugins:
return queued_plugins, selected_hooks_by_plugin
return None, None
def queued_plugins_for_snapshot(snapshot_id: str) -> list[str] | None:
queued_plugins, _selected_hooks_by_plugin = queued_plugins_and_hooks_for_snapshot(snapshot_id)
return queued_plugins
return queued_plugins or None
def config_overrides_for_queued_plugins(selected_plugins: list[str], **overrides: Any) -> dict[str, Any]:
@ -1422,40 +1397,27 @@ def config_overrides_for_queued_plugins(selected_plugins: list[str], **overrides
return config_overrides
def fail_unavailable_queued_hooks(
def fail_unavailable_queued_plugins(
snapshot_id: str,
selected_hooks_by_plugin: dict[str, set[str] | None],
queued_plugins: list[str],
plugins: dict[str, Plugin],
) -> None:
from archivebox.core.models import ArchiveResult
unavailable_plugins = set(queued_plugins) - set(plugins)
if not unavailable_plugins:
return
now = timezone.now()
for plugin_name, selected_hook_names in selected_hooks_by_plugin.items():
if selected_hook_names is None:
continue
if plugin_name in plugins:
available_hook_names = {
name for hook in plugins[plugin_name].filter_hooks("Snapshot") for name in (hook.name, Path(hook.name).stem)
}
else:
available_hook_names = set()
missing_hook_names = [hook_name for hook_name in selected_hook_names if hook_name not in available_hook_names]
if not missing_hook_names:
continue
# Hook-level resume rows are durable scheduler state. If a plugin is
# installed but no longer exposes a queued hook, mark that row failed so
# the snapshot is not retried forever with no hook left to execute.
ArchiveResult.objects.filter(
snapshot_id=snapshot_id,
plugin=plugin_name,
hook_name__in=missing_hook_names,
status=ArchiveResult.StatusChoices.QUEUED,
).update(
status=ArchiveResult.StatusChoices.FAILED,
start_ts=now,
end_ts=now,
output_str="Queued hook is no longer available in the installed plugin",
)
ArchiveResult.objects.filter(
snapshot_id=snapshot_id,
plugin__in=unavailable_plugins,
status=ArchiveResult.StatusChoices.QUEUED,
).update(
status=ArchiveResult.StatusChoices.FAILED,
start_ts=now,
end_ts=now,
output_str="Queued plugin is not available in this ArchiveBox installation",
)
def skip_disabled_queued_plugins(snapshot_id: str, plugin_names: list[str]) -> None:
@ -1480,35 +1442,6 @@ def skip_disabled_queued_plugins(snapshot_id: str, plugin_names: list[str]) -> N
)
def include_background_prerequisite_hooks(
selected_hooks_by_plugin: dict[str, set[str] | None],
plugins: dict[str, Plugin],
) -> dict[str, set[str] | None]:
expanded: dict[str, set[str] | None] = {}
for plugin_name, selected_hook_names in selected_hooks_by_plugin.items():
if selected_hook_names is None or plugin_name not in plugins:
expanded[plugin_name] = selected_hook_names
continue
plugin_hooks = sorted(plugins[plugin_name].filter_hooks("Snapshot"), key=lambda hook: hook.sort_key)
selected_sort_keys = [
hook.sort_key for hook in plugin_hooks if hook.name in selected_hook_names or Path(hook.name).stem in selected_hook_names
]
if not selected_sort_keys:
expanded[plugin_name] = set(selected_hook_names)
continue
first_selected_sort_key = min(selected_sort_keys)
expanded_hook_names = set(selected_hook_names)
# Earlier background hooks publish live resources (e.g. Chrome tabs)
# needed by later foreground hooks, but completed foreground hooks stay
# final and are not rerun during hook-level resume.
for hook in plugin_hooks:
if hook.is_background and hook.sort_key < first_selected_sort_key:
expanded_hook_names.add(hook.name)
expanded_hook_names.add(Path(hook.name).stem)
expanded[plugin_name] = expanded_hook_names
return expanded
def snapshot_hooks_for_pending_archiveresults(snapshot) -> list[tuple[str, str]]:
from archivebox.config.common import get_config
from archivebox.core.models import Snapshot
@ -1772,10 +1705,10 @@ def _run_due_snapshot_locked(snapshot, *, lock_seconds: int, interactive_interru
return False
snapshot.refresh_from_db()
if snapshot.status == Snapshot.StatusChoices.QUEUED:
has_server_archiveresults = snapshot.archiveresult_set.exclude(
has_server_workset = snapshot.archiveresult_set.exclude(
hook_name=Snapshot.BROWSER_EXTENSION_UPLOAD_HOOK_NAME,
).exists()
if has_server_archiveresults and snapshot.is_finished_processing():
if has_server_workset and snapshot.is_finished_processing():
finalize_completed_snapshot(str(snapshot.id), output_dir=Path(snapshot.output_dir))
snapshot.refresh_from_db()
if snapshot.status == Snapshot.StatusChoices.SEALED:
@ -1784,10 +1717,10 @@ def _run_due_snapshot_locked(snapshot, *, lock_seconds: int, interactive_interru
snapshot.refresh_from_db()
_runner_console_line(crawl_id=snapshot.crawl_id, snapshot=snapshot, status="SEALED")
return True
# A Snapshot with no server hook rows is fresh lifecycle work; materialize
# its configured hook set. Browser-extension uploads are completed external
# outputs, not the durable server workset, and must coexist with these rows.
if not has_server_archiveresults:
if not has_server_workset:
# Materialize one durable row per configured plugin. Existing
# browser-uploaded rows are reused and queued so the server adds its
# outputs to that plugin result instead of creating a sibling row.
snapshot.create_pending_archiveresults(hooks=snapshot_hooks_for_pending_archiveresults(snapshot))
snapshot.sm.tick()
snapshot.refresh_from_db()
@ -1804,11 +1737,11 @@ def _run_due_snapshot_locked(snapshot, *, lock_seconds: int, interactive_interru
_runner_console_line(crawl_id=snapshot.crawl_id, snapshot=snapshot, status="SEALED")
return True
if snapshot.status == Snapshot.StatusChoices.STARTED:
queued_plugins, selected_hooks_by_plugin = queued_plugins_and_hooks_for_snapshot(str(snapshot.id))
if queued_plugins and selected_hooks_by_plugin:
fail_unavailable_queued_hooks(
queued_plugins = queued_plugins_for_snapshot(str(snapshot.id))
if queued_plugins:
fail_unavailable_queued_plugins(
str(snapshot.id),
selected_hooks_by_plugin,
queued_plugins,
_discover_archivebox_plugins(),
)
if not queued_plugins_for_snapshot(str(snapshot.id)):
@ -1832,10 +1765,10 @@ def _run_due_snapshot_locked(snapshot, *, lock_seconds: int, interactive_interru
)
snapshot.refresh_from_db()
if queued_plugins_for_snapshot(str(snapshot.id)):
# Hook-level resume work is tracked by queued ArchiveResult rows, not by
# Plugin-level resume work is tracked by queued ArchiveResult rows, not by
# the Snapshot lease. If a partial pass returns with rows still queued,
# wake the Snapshot immediately so takeover does not wait out a stale
# active-state lock before running the remaining hooks.
# active-state lock before running the remaining plugins.
snapshot.update_and_requeue(retry_at=timezone.now())
return True

View File

@ -1,6 +1,7 @@
from datetime import timedelta
import pytest
from django.core.files.uploadedfile import SimpleUploadedFile
from django.db import connection
from django.test.utils import CaptureQueriesContext
from django.utils import timezone
@ -13,6 +14,43 @@ from archivebox.tests.conftest import api_client_request
pytestmark = pytest.mark.django_db(transaction=True)
def test_archiveresult_upload_upserts_one_row_per_plugin(client, api_admin_user, api_headers):
crawl = Crawl.objects.create(urls="https://example.com", created_by=api_admin_user)
snapshot = Snapshot.objects.create(url="https://example.com/unified-result", crawl=crawl)
first_response = client.post(
"/api/v1/core/archiveresults",
{
"snapshot_id": str(snapshot.id),
"plugin": "screenshot",
"hook_name": "on_Snapshot__archivebox_browser_extension_upload",
"files": SimpleUploadedFile("browser.png", b"browser", content_type="image/png"),
"output_paths": "browser.png",
},
**api_headers,
)
second_response = client.post(
"/api/v1/core/archiveresults",
{
"snapshot_id": str(snapshot.id),
"plugin": "screenshot",
"hook_name": "on_Snapshot__50_screenshot.py",
"files": SimpleUploadedFile("server.png", b"server", content_type="image/png"),
"output_paths": "server.png",
},
**api_headers,
)
assert first_response.status_code == 200, first_response.content
assert second_response.status_code == 200, second_response.content
assert second_response.json()["id"] == first_response.json()["id"]
result = ArchiveResult.objects.get(snapshot=snapshot, plugin="screenshot")
assert result.hook_name == "on_Snapshot__50_screenshot.py"
assert set(result.output_files) == {"browser.png", "server.png"}
assert (result.output_dir / "browser.png").read_bytes() == b"browser"
assert (result.output_dir / "server.png").read_bytes() == b"server"
def test_archiveresult_upload_api_queues_snapshot_maintenance_without_finalizing(client, api_admin_user, api_headers):
crawl = Crawl.objects.create(
urls="https://example.com",

View File

@ -199,7 +199,7 @@ def test_process_completed_projects_inline_archiveresult(tmp_path, hermetic_lib_
_cleanup_machine_process_rows()
def test_archiveresult_event_retry_updates_existing_hook_row(tmp_path, hermetic_lib_dir):
def test_archiveresult_events_from_multiple_hooks_update_one_plugin_row(tmp_path, hermetic_lib_dir):
from archivebox.core.models import ArchiveResult
snapshot = _create_snapshot()
@ -221,16 +221,61 @@ def test_archiveresult_event_retry_updates_existing_hook_row(tmp_path, hermetic_
snapshot,
plugin="hashes",
hook_name="on_Snapshot__93_hashes.py",
event_hook_name="on_Snapshot__99_hashes_followup.py",
lib_dir=hermetic_lib_dir,
env={"HASHES_ENABLED": "True"},
)
assert retry_result.id == first_result_id
assert retry_result.status == ArchiveResult.StatusChoices.SUCCEEDED
assert ArchiveResult.objects.filter(snapshot=snapshot, plugin="hashes", hook_name="on_Snapshot__93_hashes.py").count() == 1
assert retry_result.hook_name == "on_Snapshot__99_hashes_followup.py"
assert ArchiveResult.objects.filter(snapshot=snapshot, plugin="hashes").count() == 1
_cleanup_machine_process_rows()
def test_archiveresult_duplicate_hook_rows_are_rejected():
def test_late_background_event_cannot_overwrite_plugin_result():
from django.utils import timezone
from abx_dl.events import ArchiveResultEvent
from archivebox.core.models import ArchiveResult
from archivebox.services.archive_result_service import _save_archiveresult_event_to_db
snapshot = _create_snapshot()
now = timezone.now().isoformat()
_save_archiveresult_event_to_db(
ArchiveResultEvent(
snapshot_id=str(snapshot.id),
plugin="chrome",
hook_name="on_Snapshot__30_chrome_navigate.js",
status="succeeded",
output_str="navigation complete",
output_files=[OutputFile(path="navigate.json", extension="json", mimetype="application/json", size=7)],
start_ts=now,
end_ts=now,
),
None,
)
_save_archiveresult_event_to_db(
ArchiveResultEvent(
snapshot_id=str(snapshot.id),
plugin="chrome",
hook_name="on_Snapshot__01_chrome_tab.daemon.bg.js",
status="skipped",
output_str="background cleanup",
output_files=[OutputFile(path="tab.json", extension="json", mimetype="application/json", size=3)],
start_ts=now,
end_ts=now,
),
None,
)
result = ArchiveResult.objects.get(snapshot=snapshot, plugin="chrome")
assert result.hook_name == "on_Snapshot__30_chrome_navigate.js"
assert result.status == ArchiveResult.StatusChoices.SUCCEEDED
assert result.output_str == "navigation complete"
assert set(result.output_files) == {"navigate.json", "tab.json"}
assert result.output_size == 10
def test_archiveresult_duplicate_plugin_rows_are_rejected():
from django.db import IntegrityError, transaction
from archivebox.core.models import ArchiveResult
@ -246,11 +291,61 @@ def test_archiveresult_duplicate_hook_rows_are_rejected():
ArchiveResult.objects.create(
snapshot=snapshot,
plugin="wget",
hook_name="on_Snapshot__06_wget.finite.bg",
hook_name="on_Snapshot__99_other_wget_hook",
status=ArchiveResult.StatusChoices.SUCCEEDED,
)
def test_pending_archiveresults_create_one_row_per_plugin():
from archivebox.core.models import ArchiveResult
snapshot = _create_snapshot()
results = snapshot.create_pending_archiveresults(
hooks=[
("chrome", "on_Snapshot__00_chrome_launch.daemon.bg.js"),
("chrome", "on_Snapshot__01_chrome_tab.daemon.bg.js"),
("chrome", "on_Snapshot__30_chrome_navigate.js"),
],
)
assert len(results) == 1
assert results[0].plugin == "chrome"
assert results[0].hook_name == ""
assert ArchiveResult.objects.filter(snapshot=snapshot, plugin="chrome").count() == 1
def test_snapshot_merge_preserves_one_combined_plugin_result():
from archivebox.core.models import ArchiveResult, Snapshot
keeper = _create_snapshot()
duplicate = Snapshot.objects.create(
url="https://example.com/duplicate",
crawl=keeper.crawl,
status=Snapshot.StatusChoices.STARTED,
)
ArchiveResult.objects.create(
snapshot=keeper,
plugin="responses",
status=ArchiveResult.StatusChoices.SUCCEEDED,
output_files={"browser.png": {"size": 7}},
output_size=7,
)
ArchiveResult.objects.create(
snapshot=duplicate,
plugin="responses",
status=ArchiveResult.StatusChoices.NORESULTS,
output_files={"server.json": {"size": 3}},
output_size=3,
)
Snapshot._merge_snapshots([keeper, duplicate])
result = ArchiveResult.objects.get(snapshot=keeper, plugin="responses")
assert set(result.output_files) == {"browser.png", "server.json"}
assert result.output_size == 10
assert not Snapshot.objects.filter(pk=duplicate.pk).exists()
def test_process_completed_projects_failed_archiveresult_from_shipped_hook(tmp_path, hermetic_lib_dir):
from archivebox.core.models import ArchiveResult
@ -434,12 +529,13 @@ def test_retry_failed_archiveresults_requeues_snapshot_in_queued_state():
reset_count = snapshot.retry_failed_archiveresults()
snapshot.refresh_from_db()
result = ArchiveResult.objects.get(snapshot=snapshot, plugin="chrome", hook_name="on_Snapshot__11_chrome_wait")
result = ArchiveResult.objects.get(snapshot=snapshot, plugin="chrome")
assert reset_count == 1
assert snapshot.status == Snapshot.StatusChoices.QUEUED
assert snapshot.retry_at is not None
assert snapshot.current_step == 0
assert result.status == ArchiveResult.StatusChoices.QUEUED
assert result.hook_name == ""
assert result.output_str == ""
assert result.output_json is None
assert result.output_files == {}

View File

@ -970,7 +970,7 @@ class TestRecoverOrchestratorState:
assert snapshot.retry_at is None
assert snapshot.downloaded_at is not None
def test_create_pending_archiveresults_uses_canonical_hook_names(self):
def test_create_pending_archiveresults_creates_one_plugin_row(self):
from django.utils import timezone
from archivebox.base_models.models import get_or_create_system_user_pk
@ -992,9 +992,10 @@ class TestRecoverOrchestratorState:
snapshot.create_pending_archiveresults()
hook_names = list(ArchiveResult.objects.filter(snapshot=snapshot).values_list("hook_name", flat=True))
assert hook_names
assert all(not hook_name.endswith((".py", ".js", ".sh")) for hook_name in hook_names)
results = list(ArchiveResult.objects.filter(snapshot=snapshot))
assert results
assert all(result.hook_name == "" for result in results)
assert len(results) == len({result.plugin for result in results})
def test_snapshot_hooks_for_pending_archiveresults_respects_disabled_plugins_when_plugins_empty(self):
from django.utils import timezone
@ -1610,7 +1611,7 @@ class TestRecoverOrchestratorState:
assert result.end_ts is not None
@pytest.mark.django_db(transaction=True)
def test_run_due_snapshot_fails_obsolete_queued_hook_name(self):
def test_run_due_snapshot_fails_unavailable_queued_plugin(self):
from django.utils import timezone
from archivebox.base_models.models import get_or_create_system_user_pk
@ -1632,8 +1633,8 @@ class TestRecoverOrchestratorState:
)
result = ArchiveResult.objects.create(
snapshot=snapshot,
plugin="singlefile",
hook_name="on_Snapshot__50_singlefile.py",
plugin="removed_plugin",
hook_name="",
status=ArchiveResult.StatusChoices.QUEUED,
)
@ -1642,7 +1643,7 @@ class TestRecoverOrchestratorState:
result.refresh_from_db()
snapshot.refresh_from_db()
assert result.status == ArchiveResult.StatusChoices.FAILED
assert result.output_str == "Queued hook is no longer available in the installed plugin"
assert result.output_str == "Queued plugin is not available in this ArchiveBox installation"
assert snapshot.retry_at is None
@pytest.mark.django_db(transaction=True)
@ -1715,7 +1716,7 @@ class TestRecoverOrchestratorState:
@pytest.mark.django_db(transaction=True)
@pytest.mark.timeout(300)
@pytest.mark.parametrize("chrome_isolation", ["crawl", "snapshot"])
def test_resume_queued_chrome_navigate_reruns_background_prerequisites(
def test_retry_queued_chrome_result_reruns_full_plugin(
self,
initialized_archive,
recursive_test_site,
@ -1755,21 +1756,18 @@ class TestRecoverOrchestratorState:
)
assert list_process.returncode == 0, list_process.stderr or list_process.stdout
chrome_results = parse_jsonl_output(list_process.stdout)
navigate_record = next(record for record in chrome_results if record["hook_name"] == "on_Snapshot__30_chrome_navigate")
snapshot_id = navigate_record["snapshot_id"]
assert len(chrome_results) == 1
chrome_record = chrome_results[0]
snapshot_id = chrome_record["snapshot_id"]
with use_archivebox_db(initialized_archive):
tab_result = ArchiveResult.objects.get(
snapshot_id=snapshot_id,
plugin="chrome",
hook_name="on_Snapshot__01_chrome_tab.daemon.bg",
)
first_tab_process_id = tab_result.process_id
assert first_tab_process_id is not None
chrome_result = ArchiveResult.objects.get(snapshot_id=snapshot_id, plugin="chrome")
first_process_id = chrome_result.process_id
assert first_process_id is not None
update_process = run_archivebox_cmd(
["archiveresult", "update", "--status=queued"],
stdin=next(line for line in list_process.stdout.splitlines() if navigate_record["id"] in line) + "\n",
stdin=next(line for line in list_process.stdout.splitlines() if chrome_record["id"] in line) + "\n",
cwd=initialized_archive,
env=env,
timeout=60,
@ -1796,21 +1794,12 @@ class TestRecoverOrchestratorState:
cleanup_process_group(run_process.pid)
with use_archivebox_db(initialized_archive):
navigate_result = ArchiveResult.objects.get(
snapshot_id=snapshot_id,
plugin="chrome",
hook_name="on_Snapshot__30_chrome_navigate",
)
tab_result = ArchiveResult.objects.get(
snapshot_id=snapshot_id,
plugin="chrome",
hook_name="on_Snapshot__01_chrome_tab.daemon.bg",
)
chrome_result = ArchiveResult.objects.get(snapshot_id=snapshot_id, plugin="chrome")
assert run_process.returncode == 0
assert navigate_result.status == ArchiveResult.StatusChoices.SUCCEEDED
assert tab_result.process_id is not None
assert tab_result.process_id != first_tab_process_id
assert chrome_result.status == ArchiveResult.StatusChoices.SUCCEEDED
assert chrome_result.process_id is not None
assert chrome_result.process_id != first_process_id
def test_recover_orchestrator_state_ignores_sealed_downloaded_snapshot_without_results(self):
from django.utils import timezone

View File

@ -0,0 +1,53 @@
import pytest
from django.db import connection
from django.db.migrations.executor import MigrationExecutor
pytestmark = pytest.mark.django_db(transaction=True)
def test_migration_consolidates_plugin_rows_and_enforces_uniqueness():
try:
executor = MigrationExecutor(connection)
executor.migrate([("core", "0051_postgres_url_pattern_ops_index")])
old_apps = executor.loader.project_state([("core", "0051_postgres_url_pattern_ops_index")]).apps
Crawl = old_apps.get_model("crawls", "Crawl")
Snapshot = old_apps.get_model("core", "Snapshot")
ArchiveResult = old_apps.get_model("core", "ArchiveResult")
crawl = Crawl.objects.create(urls="https://example.com")
snapshot = Snapshot.objects.create(url="https://example.com/migration", crawl=crawl, output_size=10)
ArchiveResult.objects.create(
snapshot=snapshot,
plugin="responses",
hook_name="browser-upload",
status="succeeded",
output_files={"browser.png": {"size": 7}},
output_size=7,
output_mimetypes="image/png",
)
ArchiveResult.objects.create(
snapshot=snapshot,
plugin="responses",
hook_name="server-capture",
status="noresults",
output_files={"server.json": {"size": 3}},
output_size=3,
output_mimetypes="application/json",
)
executor = MigrationExecutor(connection)
executor.migrate([("core", "0052_unique_archiveresult_per_snapshot_plugin")])
new_apps = executor.loader.project_state([("core", "0052_unique_archiveresult_per_snapshot_plugin")]).apps
Snapshot = new_apps.get_model("core", "Snapshot")
ArchiveResult = new_apps.get_model("core", "ArchiveResult")
snapshot = Snapshot.objects.get(id=snapshot.id)
result = ArchiveResult.objects.get(snapshot=snapshot, plugin="responses")
assert ArchiveResult.objects.filter(snapshot=snapshot, plugin="responses").count() == 1
assert result.status == "succeeded"
assert set(result.output_files) == {"browser.png", "server.json"}
assert result.output_size == 10
assert snapshot.output_size == 10
assert set(result.output_mimetypes.split(",")) == {"application/json", "image/png"}
finally:
MigrationExecutor(connection).migrate([("core", "0052_unique_archiveresult_per_snapshot_plugin")])

View File

@ -141,29 +141,3 @@ class TestArchiveResultAdminListView:
assert not output_dir.exists()
snapshot.refresh_from_db()
assert snapshot.output_size == 0
def test_deleting_duplicate_preserves_shared_plugin_output(self, snapshot):
from archivebox.core.models import ArchiveResult
output_dir = Path(snapshot.output_dir) / "responses"
output_dir.mkdir(parents=True, exist_ok=True)
output_file = output_dir / "index.jsonl"
output_file.write_text("captured response\n")
primary = ArchiveResult.objects.create(
snapshot=snapshot,
plugin="responses",
hook_name="on_Snapshot__24_responses.daemon.bg",
status=ArchiveResult.StatusChoices.SUCCEEDED,
output_size=18,
)
duplicate = ArchiveResult.objects.create(
snapshot=snapshot,
plugin="responses",
hook_name="on_Snapshot__24_responses.daemon.bg.replayed",
status=ArchiveResult.StatusChoices.NORESULTS,
)
duplicate.delete()
assert ArchiveResult.objects.filter(pk=primary.pk).exists()
assert output_file.read_text() == "captured response\n"