feat: flatten file tree for better performance

This commit is contained in:
tdurieux
2024-04-26 10:31:57 +01:00
parent ccdc95e4a8
commit 710f7328e7
23 changed files with 516 additions and 514 deletions
+20 -9
View File
@@ -1,4 +1,3 @@
import { Tree } from "../types";
import config from "../../config";
import * as fs from "fs";
import { Extract } from "unzip-stream";
@@ -10,6 +9,8 @@ import { promisify } from "util";
import { lookup } from "mime-types";
import { trace } from "@opentelemetry/api";
import StorageBase, { FILE_TYPE } from "./Storage";
import FileModel from "../model/files/files.model";
import { IFile } from "../model/files/files.types";
export default class FileSystem extends StorageBase {
type = "FileSystem";
@@ -138,23 +139,25 @@ export default class FileSystem extends StorageBase {
opt: {
onEntry?: (file: { path: string; size: number }) => void;
} = {}
): Promise<Tree> {
): Promise<IFile[]> {
return trace
.getTracer("ano-file")
.startActiveSpan("fs.listFiles", async (span) => {
span.setAttribute("path", dir);
const fullPath = join(config.FOLDER, this.repoPath(repoId), dir);
let files = await fs.promises.readdir(fullPath);
const output: Tree = {};
const output2: IFile[] = [];
for (let file of files) {
let filePath = join(fullPath, file);
try {
const stats = await fs.promises.stat(filePath);
if (file[0] == "$") {
file = "\\" + file;
}
if (stats.isDirectory()) {
output[file] = await this.listFiles(repoId, join(dir, file), opt);
output2.push(
new FileModel({ name: file, path: dir, repoID: repoId })
);
output2.push(
...(await this.listFiles(repoId, join(dir, file), opt))
);
} else if (stats.isFile()) {
if (opt.onEntry) {
opt.onEntry({
@@ -162,14 +165,22 @@ export default class FileSystem extends StorageBase {
size: stats.size,
});
}
output[file] = { size: stats.size, sha: stats.ino.toString() };
output2.push(
new FileModel({
name: file,
path: dir,
repoID: repoId,
size: stats.size,
sha: stats.ino.toString(),
})
);
}
} catch (error) {
span.recordException(error as Error);
}
}
span.end();
return output;
return output2;
});
}
+13 -19
View File
@@ -14,9 +14,10 @@ import { lookup } from "mime-types";
import * as archiver from "archiver";
import { trace } from "@opentelemetry/api";
import { dirname, basename, join } from "path";
import { Tree, TreeFile } from "../types";
import AnonymousError from "../AnonymousError";
import StorageBase, { FILE_TYPE } from "./Storage";
import { IFile } from "../model/files/files.types";
import FileModel from "../model/files/files.model";
export default class S3Storage extends StorageBase {
type = "AWS";
@@ -245,13 +246,13 @@ export default class S3Storage extends StorageBase {
}
/** @override */
async listFiles(repoId: string, dir: string = ""): Promise<Tree> {
async listFiles(repoId: string, dir: string = ""): Promise<IFile[]> {
const span = trace.getTracer("ano-file").startSpan("s3.listFiles");
span.setAttribute("path", dir);
try {
if (!config.S3_BUCKET) throw new Error("S3_BUCKET not set");
if (dir && dir[dir.length - 1] != "/") dir = dir + "/";
const out: Tree = {};
const out: IFile[] = [];
let req: ListObjectsV2CommandOutput;
let nextContinuationToken: string | undefined;
do {
@@ -267,22 +268,15 @@ export default class S3Storage extends StorageBase {
for (const f of req.Contents) {
if (!f.Key) continue;
f.Key = f.Key.replace(join(this.repoPath(repoId), dir), "");
const paths = f.Key.split("/");
let current: Tree = out;
for (let i = 0; i < paths.length - 1; i++) {
let p = paths[i];
if (!p) continue;
if (!(current[p] as Tree)) {
current[p] = {} as Tree;
}
current = current[p] as Tree;
}
if (f.ETag) {
const fileInfo: TreeFile = { size: f.Size || 0, sha: f.ETag };
const fileName = paths[paths.length - 1];
if (fileName) current[fileName] = fileInfo;
}
out.push(
new FileModel({
name: basename(f.Key),
path: dirname(f.Key),
repoID: repoId,
size: f.Size,
sha: f.ETag,
})
);
}
} while (req && req.Contents && req.IsTruncated);
return out;
+2 -2
View File
@@ -3,9 +3,9 @@ import { Transform, Readable } from "stream";
import * as archiver from "archiver";
import { Response } from "express";
import { Tree } from "../types";
import S3Storage from "./S3";
import FileSystem from "./FileSystem";
import { IFile } from "../model/files/files.types";
export type Storage = S3Storage | FileSystem;
@@ -62,7 +62,7 @@ export default abstract class StorageBase {
* List the files from dir
* @param dir
*/
abstract listFiles(repoId: string, dir: string): Promise<Tree>;
abstract listFiles(repoId: string, dir: string): Promise<IFile[]>;
/**
* Extract the content of tar to dir