From 35e2e38811358cab1bf125dd09acb87950e2dec2 Mon Sep 17 00:00:00 2001
From: Matthieu Haineault <matthieuh@targointernet.com>
Date: Mon, 25 Aug 2025 16:42:05 -0400
Subject: [PATCH] feat(attachments): setup Garbage Collector for attachments
 module

---
 src/modules/attachments/attachments.module.ts | 12 ++-
 .../services/garbage-collector.service.ts     | 78 +++++++++++++++++++
 2 files changed, 88 insertions(+), 2 deletions(-)
 create mode 100644 src/modules/attachments/services/garbage-collector.service.ts

diff --git a/src/modules/attachments/attachments.module.ts b/src/modules/attachments/attachments.module.ts
index 75f5f02..ee8883f 100644
--- a/src/modules/attachments/attachments.module.ts
+++ b/src/modules/attachments/attachments.module.ts
@@ -2,10 +2,18 @@ import { ScheduleModule } from "@nestjs/schedule";
 import { PrismaService } from "src/prisma/prisma.service";
 import { ArchivalAttachmentService } from "./services/archival-attachment.service";
 import { Module } from "@nestjs/common";
+import { GarbargeCollectorService } from "./services/garbage-collector.service";
 
 @Module({
     imports: [ScheduleModule.forRoot()],
-    providers: [PrismaService, ArchivalAttachmentService],
-    exports: [ArchivalAttachmentService],
+    providers: [
+        PrismaService,
+        ArchivalAttachmentService,
+        GarbargeCollectorService,
+    ],
+    exports: [
+        ArchivalAttachmentService,
+        GarbargeCollectorService
+    ],
 })
 export class ArchivalAttachmentModule {}
\ No newline at end of file
diff --git a/src/modules/attachments/services/garbage-collector.service.ts b/src/modules/attachments/services/garbage-collector.service.ts
new file mode 100644
index 0000000..13e8c21
--- /dev/null
+++ b/src/modules/attachments/services/garbage-collector.service.ts
@@ -0,0 +1,78 @@
+import { Injectable, Logger } from "@nestjs/common";
+import { Cron } from "@nestjs/schedule";
+import { PrismaService } from 'src/prisma/prisma.service';
+import * as path from 'node:path';
+import { promises as fsp } from 'node:fs';
+import { resolveAttachmentsRoot } from "src/config/attachment.config";
+
+@Injectable()
+export class GarbargeCollectorService {
+    private readonly logger = new Logger(GarbargeCollectorService.name);
+
+    //.env refs
+    private readonly batch_size = Number(process.env.GC_BATCH_SIZE || 500);
+    private readonly cron_expression = process.env.GC_CRON || '15 4 * * *'; // everyday at 04:15 AM
+
+    //fetchs root of storage
+    private readonly root = resolveAttachmentsRoot();
+
+    constructor(private readonly prisma: PrismaService) {}
+
+    //planif for the Cronjob
+    @Cron(function(this:GarbargeCollectorService) { return this.cron_expression; } as any)
+    async runScheduled() {
+        await this.collect();
+    }
+
+    //Manage Garbage collecting by batch of elements until a batch != full
+    async collect() {
+        let total = 0, round = 0;
+        //infinit loop (;;) with break
+        for(;;) {
+            round++;
+            const num = await this.collectBatch();
+            total += num;
+            this.logger.log(`Garbage Collector round #${round} removed ${num}`);
+            if(num < this.batch_size) break;     //breaks if not a full batch       
+        }
+        this.logger.log(`Garbage Collecting done: total removed ${total}`);
+        return { removed:total };
+    }
+
+    //Manage a single lot of orphan blobs
+    private async collectBatch(): Promise<number> {
+        const blobs = await this.prisma.blobs.findMany({
+            where: { refcount: { lte: 0 } },
+            select: { sha256: true, storage_path: true },
+            take: this.batch_size, 
+        });
+        if(blobs.length === 0) return 0;
+
+        // delete original file and all its variants <hash> in the same file
+        await Promise.all(
+            blobs.map(async (blob)=> {
+                const absolute_path = path.join(this.root, blob.storage_path);
+                await this.deleteFileIfExists(absolute_path); //tries to delete original file if found
+
+                const dir = path.dirname(absolute_path);
+                const base = path.basename(absolute_path);
+                try {
+                    const entries = await fsp.readdir(dir, { withFileTypes: true});
+                    const targets = entries.filter(entry => entry.isFile() && entry.name.startsWith(base + '.'))
+                                           .map(entry => path.join(dir, entry.name));
+                    //deletes all variants
+                    await Promise.all(targets.map(target => this.deleteFileIfExists(target)));
+                } catch {}
+            })
+        );
+        //deletes blobs lignes if file is deleted
+        const hashes = blobs.map(blob => blob.sha256);
+        await this.prisma.blobs.deleteMany({where: { sha256: { in: hashes } } });
+        return blobs.length;
+    }
+
+    //helper: deletes path if exists and ignore errors
+    private async deleteFileIfExists(path: string) {
+        try { await fsp.unlink(path); } catch {}
+    }
+}
\ No newline at end of file