diff --git a/index.js b/index.js index 3aa0dbc..af6444e 100644 --- a/index.js +++ b/index.js @@ -70,15 +70,13 @@ async function fetchAndHashContent(url, visitedUrls = new Set()) { } visitedUrls.add(url); - const response = await axios.get(url, { responseType: 'arraybuffer', timeout: 10000 }); const content = response.data.toString('utf-8'); - // Parse HTML content to identify linked resources const root = parseHtml(content); const linkedResources = root.querySelectorAll('link[rel="stylesheet"], script[src]'); // Fetch and hash linked resources - const linkedResourceHashes = await Promise.all(linkedResources.map(async (resource) => { + const linkedResource = await Promise.all(linkedResources.map(async (resource) => { const resourceUrl = parseUrl(resource.getAttribute('href') || resource.getAttribute('src'), true); let absoluteResourceUrl = resourceUrl.href; if (!resourceUrl.hostname) { @@ -91,10 +89,10 @@ async function fetchAndHashContent(url, visitedUrls = new Set()) { })); // Combine the content and hashes of linked resources - return `${content}_${linkedResourceHashes.join('_')}`; + return `${content}_${linkedResource.join('_')}`; } - +const hashCache = new Map(); // API endpoint to start the recursive download and hashing app.post('/hash', async (req, res) => { try { @@ -107,17 +105,31 @@ app.post('/hash', async (req, res) => { const promises = urls.map(async (url) => { const urlWithoutHashAndQuery = parseUrlWithoutHashAndQuery(url); - const hashedContent = await fetchAndHashContent(urlWithoutHashAndQuery); - const fileHash = await hashContent(Buffer.from(hashedContent, 'utf-8')); - return { url, fileHash }; + let hash; + // regex to identify owner and repo name from https://owner.github.io/repo-name + const githubRepoRegex = /https?:\/\/([\w-]+)\.github\.io\/([\w-]+)/; + if (githubRepoRegex.test(urlWithoutHashAndQuery)) { + const [, owner, repo] = githubRepoRegex.exec(urlWithoutHashAndQuery) || [null, null, null,]; + const { data } = await axios.get(`https://api.github.com/repos/${owner}/${repo}`); + const lastUpdated = new Date(data.pushed_at); + + const cached = hashCache.get(urlWithoutHashAndQuery); + if (cached && cached.lastUpdated >= lastUpdated) { + hash = cached.hash; + } else { + const hashedContent = await fetchAndHashContent(urlWithoutHashAndQuery); + hash = await hashContent(Buffer.from(hashedContent, 'utf-8')); + hashCache.set(urlWithoutHashAndQuery, { hash, lastUpdated }); + } + } else { + const hashedContent = await fetchAndHashContent(urlWithoutHashAndQuery); + hash = await hashContent(Buffer.from(hashedContent, 'utf-8')); + } + + return { url, hash }; }); let results = await Promise.all(promises); - results = results.reduce((acc, { url, fileHash }) => { - acc[url] = fileHash; - return acc; - }, {}); - res.json(results); } catch (error) { res.status(500).json({ error: error.message }); diff --git a/index.min.js b/index.min.js index ae7e2e0..ac91f98 100644 --- a/index.min.js +++ b/index.min.js @@ -1 +1 @@ -require("dotenv").config();const express=require("express"),cors=require("cors"),axios=require("axios"),{createHash:createHash}=require("crypto"),archiver=require("archiver"),rateLimit=require("express-rate-limit"),{parse:parseUrl,URL:URL}=require("url"),{parse:parseHtml}=require("node-html-parser"),allowedDomains=process.env.ALLOWED_DOMAINS.split(","),app=express();app.use(cors());const port=process.env.PORT||3e3,host=process.env.HOST||"0.0.0.0";function addProtocolToUrl(url){return url.startsWith("http://")||url.startsWith("https://")||(url="https://"+url),url}function parseUrlWithoutHashAndQuery(fullUrl){fullUrl=addProtocolToUrl(fullUrl);const parsedUrl=new URL(fullUrl);parsedUrl.hash="",parsedUrl.search="";return parsedUrl.toString()}async function hashContent(content){const hash=createHash("sha256");return hash.update(content),hash.digest("hex")}async function fetchAndHashContent(url,visitedUrls=new Set){if(visitedUrls.has(url))return"";visitedUrls.add(url);const content=(await axios.get(url,{responseType:"arraybuffer",timeout:1e4})).data.toString("utf-8"),linkedResources=parseHtml(content).querySelectorAll('link[rel="stylesheet"], script[src]');return`${content}_${(await Promise.all(linkedResources.map((async resource=>{const resourceUrl=parseUrl(resource.getAttribute("href")||resource.getAttribute("src"),!0);let absoluteResourceUrl=resourceUrl.href;resourceUrl.hostname||(resourceUrl.path.startsWith("/")||url.endsWith("/")||(url+="/"),absoluteResourceUrl=`${url}${resourceUrl.path}`);const resourceContent=await fetchAndHashContent(absoluteResourceUrl,visitedUrls);return`${resourceUrl.path}_${resourceContent}`})))).join("_")}`}async function downloadGitHubRepo(owner,repo){if(!owner||!repo)throw new Error("Missing owner or repo");const zipUrl=`https://github.com/${owner}/${repo}/archive/refs/heads/master.zip`;return(await axios.get(zipUrl,{responseType:"arraybuffer"})).data}app.use(express.json()),app.use(rateLimit({windowMs:6e4,max:10})),app.get("/",((req,res)=>{res.send("Hello There!")})),app.post("/hash",(async(req,res)=>{try{let{urls:urls}=req.body;if(!urls)return res.status(400).json({error:"Missing in the request parameters"});Array.isArray(urls)||(urls=[urls]);const promises=urls.map((async url=>{const urlWithoutHashAndQuery=parseUrlWithoutHashAndQuery(url),hashedContent=await fetchAndHashContent(urlWithoutHashAndQuery);return{url:url,fileHash:await hashContent(Buffer.from(hashedContent,"utf-8"))}}));let results=await Promise.all(promises);results=results.reduce(((acc,{url:url,fileHash:fileHash})=>(acc[url]=fileHash,acc)),{}),res.json(results)}catch(error){res.status(500).json({error:error.message})}})),app.post("/download-repos",(async(req,res)=>{try{let{urls:urls}=req.body;if(!urls)return res.status(400).json({error:"Missing in the request parameters"});Array.isArray(urls)||(urls=[urls]);const archive=archiver("zip");res.attachment("repos.zip");const downloadPromises=urls.map((async url=>{const[owner,name]=url.split("/").slice(-2);if(!owner||!name)return void console.error(`Invalid url format: ${url}`);const zipBuffer=await downloadGitHubRepo(owner,name);archive.append(zipBuffer,{name:`${owner}-${name}.zip`})}));await Promise.all(downloadPromises),archive.finalize(),archive.pipe(res)}catch(error){res.status(500).json({error:error.message})}})),app.listen(port,host,(()=>{console.log(`Server is running at http://${host}:${port}`)})),module.exports=app; \ No newline at end of file +require("dotenv").config();const express=require("express"),cors=require("cors"),axios=require("axios"),{createHash:createHash}=require("crypto"),archiver=require("archiver"),rateLimit=require("express-rate-limit"),{parse:parseUrl,URL:URL}=require("url"),{parse:parseHtml}=require("node-html-parser"),allowedDomains=process.env.ALLOWED_DOMAINS.split(","),app=express();app.use(cors());const port=process.env.PORT||3e3,host=process.env.HOST||"0.0.0.0";function addProtocolToUrl(url){return url.startsWith("http://")||url.startsWith("https://")||(url="https://"+url),url}function parseUrlWithoutHashAndQuery(fullUrl){fullUrl=addProtocolToUrl(fullUrl);const parsedUrl=new URL(fullUrl);parsedUrl.hash="",parsedUrl.search="";return parsedUrl.toString()}async function hashContent(content){const hash=createHash("sha256");return hash.update(content),hash.digest("hex")}async function fetchAndHashContent(url,visitedUrls=new Set){if(visitedUrls.has(url))return"";visitedUrls.add(url);const content=(await axios.get(url,{responseType:"arraybuffer",timeout:1e4})).data.toString("utf-8"),linkedResources=parseHtml(content).querySelectorAll('link[rel="stylesheet"], script[src]');return`${content}_${(await Promise.all(linkedResources.map((async resource=>{const resourceUrl=parseUrl(resource.getAttribute("href")||resource.getAttribute("src"),!0);let absoluteResourceUrl=resourceUrl.href;resourceUrl.hostname||(resourceUrl.path.startsWith("/")||url.endsWith("/")||(url+="/"),absoluteResourceUrl=`${url}${resourceUrl.path}`);const resourceContent=await fetchAndHashContent(absoluteResourceUrl,visitedUrls);return`${resourceUrl.path}_${resourceContent}`})))).join("_")}`}app.use(express.json()),app.use(rateLimit({windowMs:6e4,max:10})),app.get("/",((req,res)=>{res.send("Hello There!")}));const hashCache=new Map;async function downloadGitHubRepo(owner,repo){if(!owner||!repo)throw new Error("Missing owner or repo");const zipUrl=`https://github.com/${owner}/${repo}/archive/refs/heads/master.zip`;return(await axios.get(zipUrl,{responseType:"arraybuffer"})).data}app.post("/hash",(async(req,res)=>{try{let{urls:urls}=req.body;if(!urls)return res.status(400).json({error:"Missing in the request parameters"});Array.isArray(urls)||(urls=[urls]);const promises=urls.map((async url=>{const urlWithoutHashAndQuery=parseUrlWithoutHashAndQuery(url);let hash;const githubRepoRegex=/https?:\/\/([\w-]+)\.github\.io\/([\w-]+)/;if(githubRepoRegex.test(urlWithoutHashAndQuery)){const[,owner,repo]=githubRepoRegex.exec(urlWithoutHashAndQuery)||[null,null,null],{data:data}=await axios.get(`https://api.github.com/repos/${owner}/${repo}`),lastUpdated=new Date(data.pushed_at),cached=hashCache.get(urlWithoutHashAndQuery);if(cached&&cached.lastUpdated>=lastUpdated)hash=cached.hash;else{const hashedContent=await fetchAndHashContent(urlWithoutHashAndQuery);hash=await hashContent(Buffer.from(hashedContent,"utf-8")),hashCache.set(urlWithoutHashAndQuery,{hash:hash,lastUpdated:lastUpdated})}}else{const hashedContent=await fetchAndHashContent(urlWithoutHashAndQuery);hash=await hashContent(Buffer.from(hashedContent,"utf-8"))}return{url:url,hash:hash}}));let results=await Promise.all(promises);res.json(results)}catch(error){res.status(500).json({error:error.message})}})),app.post("/download-repos",(async(req,res)=>{try{let{urls:urls}=req.body;if(!urls)return res.status(400).json({error:"Missing in the request parameters"});Array.isArray(urls)||(urls=[urls]);const archive=archiver("zip");res.attachment("repos.zip");const downloadPromises=urls.map((async url=>{const[owner,name]=url.split("/").slice(-2);if(!owner||!name)return void console.error(`Invalid url format: ${url}`);const zipBuffer=await downloadGitHubRepo(owner,name);archive.append(zipBuffer,{name:`${owner}-${name}.zip`})}));await Promise.all(downloadPromises),archive.finalize(),archive.pipe(res)}catch(error){res.status(500).json({error:error.message})}})),app.listen(port,host,(()=>{console.log(`Server is running at http://${host}:${port}`)})),module.exports=app; \ No newline at end of file