From cd8aa826e125ac54efee09b8d76471f6d3d05144 Mon Sep 17 00:00:00 2001
From: alikia2x <alikia2x@outlook.com>
Date: Mon, 17 Mar 2025 00:33:28 +0800
Subject: [PATCH] fix: prevent videos from being crawled for too long

---
 lib/db/snapshot.ts          | 19 ++++++++++++++++++-
 lib/mq/exec/snapshotTick.ts | 20 +++++++-------------
 2 files changed, 25 insertions(+), 14 deletions(-)

diff --git a/lib/db/snapshot.ts b/lib/db/snapshot.ts
index 663a628..5921059 100644
--- a/lib/db/snapshot.ts
+++ b/lib/db/snapshot.ts
@@ -71,7 +71,7 @@ export async function getSongSnapshotCount(client: Client, aid: number) {
 }
 
 export async function getShortTermEtaPrediction(client: Client, aid: number) {
-	const queryResult = await client.queryObject<{eta: number}>(
+	const queryResult = await client.queryObject<{ eta: number }>(
 		`
 		WITH old_snapshot AS (
 			SELECT created_at, views 
@@ -120,6 +120,23 @@ export async function getShortTermEtaPrediction(client: Client, aid: number) {
 	return queryResult.rows[0].eta;
 }
 
+export async function getIntervalFromLastSnapshotToNow(client: Client, aid: number) {
+	const queryResult = await client.queryObject<{ interval: number }>(
+		`
+		SELECT EXTRACT(EPOCH FROM (NOW() - created_at)) AS interval
+		FROM video_snapshot
+		WHERE aid = $1
+		ORDER BY created_at DESC
+		LIMIT 1;
+		`,
+		[aid],
+	);
+	if (queryResult.rows.length === 0) {
+		return null;
+	}
+	return queryResult.rows[0].interval;
+}
+
 export async function songEligibleForMilestoneSnapshot(client: Client, aid: number) {
 	const count = await getSongSnapshotCount(client, aid);
 	if (count < 2) {
diff --git a/lib/mq/exec/snapshotTick.ts b/lib/mq/exec/snapshotTick.ts
index 12443ff..bbc7205 100644
--- a/lib/mq/exec/snapshotTick.ts
+++ b/lib/mq/exec/snapshotTick.ts
@@ -1,8 +1,9 @@
 import { Job } from "bullmq";
-import { MINUTE, SECOND } from "$std/datetime/constants.ts";
+import { HOUR, MINUTE, SECOND } from "$std/datetime/constants.ts";
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
 import { db } from "lib/db/init.ts";
 import {
+getIntervalFromLastSnapshotToNow,
 	getShortTermEtaPrediction,
 	getSongsNearMilestone,
 	getUnsnapshotedSongs,
@@ -55,19 +56,12 @@ async function processMilestoneSnapshots(client: Client, vidoesNearMilestone: So
 	let i = 0;
 	for (const snapshot of vidoesNearMilestone) {
 		if (await snapshotScheduled(snapshot.aid)) {
-			logger.silly(
-				`Video ${snapshot.aid} is already scheduled for snapshot`,
-				"mq",
-				"fn:processMilestoneSnapshots",
-			);
 			continue;
 		}
-		if (await songEligibleForMilestoneSnapshot(client, snapshot.aid) === false) {
-			logger.silly(
-				`Video ${snapshot.aid} is not eligible for milestone snapshot`,
-				"mq",
-				"fn:processMilestoneSnapshots",
-			);
+		const timeFromLastSnapshot = await getIntervalFromLastSnapshotToNow(client, snapshot.aid);
+		const lastSnapshotLessThan8Hrs = timeFromLastSnapshot && timeFromLastSnapshot * SECOND < 8 * HOUR;
+		const notEligible = await songEligibleForMilestoneSnapshot(client, snapshot.aid);
+		if (notEligible && lastSnapshotLessThan8Hrs) {
 			continue;
 		}
 		const factor = Math.floor(i / 8);
@@ -143,7 +137,7 @@ export const takeSnapshotForMilestoneVideoWorker = async (job: Job) => {
 			eta = viewsToIncrease / (incrementSpeed + DELTA);
 		}
 		const scheduledNextSnapshotDelay = eta * SECOND / 3;
-		const maxInterval = 20 * MINUTE;
+		const maxInterval = 60 * MINUTE;
 		const minInterval = 1 * SECOND;
 		const delay = truncate(scheduledNextSnapshotDelay, minInterval, maxInterval);
 		await SnapshotQueue.add("snapshotMilestoneVideo", {