prompt looks good

pwa-builder · Feb 15, 2023 · 7d5dae3 · 7d5dae3
1 parent 14ea593
commit 7d5dae3
Show file tree

Hide file tree

Showing 6 changed files with 664 additions and 59 deletions.
diff --git a/dist/api.js b/dist/api.js
@@ -1,50 +1,61 @@
 import express from 'express';
 import fetch from 'node-fetch';
+import { JSDOM } from 'jsdom';
 import dotenv from 'dotenv';
 // import { askGPT, initGPT } from './gpt.js';
-import { initOpenAI, askOpenAI } from './openai.js';
+import { initOpenAI, askOpenAI, manifestPrompt } from './openai.js';
 dotenv.config();
 const app = express();
 const port = process.env.PORT || 3000;
 app.use(express.json());
-let GPTSesion = null;
+let OpenAISesion = null;
 app.get('/initAPI', async (req, res) => {
-    GPTSesion = await initOpenAI();
-    if (GPTSesion) {
-        res.status(200).send({ message: 'GPT initialized' });
+    OpenAISesion = await initOpenAI();
+    if (OpenAISesion) {
+        res.status(200).send({ message: 'LLM initialized' });
     }
     else {
-        GPTSesion = null;
-        res.status(500).send({ error: 'GPT unavailable' });
+        OpenAISesion = null;
+        res.status(500).send({ error: 'LLM unavailable' });
     }
 });
 app.get('/generateManifest', async (req, res) => {
     if (!req.query.url) {
         res.status(400).send({ error: 'URL not specified' });
         return;
     }
-    const request = await fetch(req.query.url.toString());
+    if (!OpenAISesion) {
+        res.status(500).send({ error: 'API not initialized' });
+        return;
+    }
+    const request = await fetch(req.query.url.toString(), {
+        headers: {
+            'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36',
+        }
+    });
     let rawHTML = await request.text();
     let headerHTML;
     if (typeof rawHTML === 'string') {
         rawHTML = rawHTML.replace(/\r|\n/g, '').replace(/\s{2,}/g, '');
-        headerHTML = /<head>(.*)<\/head>/.test(rawHTML) ? rawHTML.match(/<head>(.*)<\/head>/)[0] : 'Head not found';
+        headerHTML = /<head>(.*)<\/head>/.test(rawHTML) ? rawHTML.match(/<head>(.*)<\/head>/)[0] : null;
     }
-    if (!GPTSesion) {
-        res.status(500).send({ error: 'API not initialized' });
+    if (!headerHTML) {
+        res.status(400).send({ error: '<head> HTML not found' });
         return;
     }
-    if (headerHTML) {
-        const manifest = await askOpenAI(headerHTML, GPTSesion);
-        if (manifest) {
-            res.status(200).send({ manifest });
-        }
-        else
-            res.status(400).send({ error: 'GPT unavailable or failed to generate manifest' });
-    }
-    else {
-        res.status(400).send({ error: '<head> HTML not found' });
+    console.log(`HEAD: ${headerHTML}`);
+    const document = new JSDOM(headerHTML).window.document;
+    document.querySelectorAll('script').forEach((script) => script?.remove?.());
+    document.querySelectorAll('style').forEach((style) => style?.remove?.());
+    document.querySelectorAll('meta[http-equiv="origin-trial"],meta[http-equiv="content-type"],meta[name="google-site-verification"]').forEach((meta) => meta?.remove?.());
+    document.querySelectorAll('link[rel="stylesheet"],link[rel="modulepreload"],link[rel="preload"],link[rel="dns-prefetch"],link[rel="preload"]').forEach((link) => link?.remove?.());
+    const preparedHTML = document.head.innerHTML.replace(/&amp;|&{2,}|<!--|-->/g, '');
+    const manifest = await askOpenAI(manifestPrompt(preparedHTML), OpenAISesion);
+    if (manifest) {
+        res.status(200).send({ manifest });
     }
+    else
+        res.status(400).send({ error: 'LLM unavailable or failed to generate manifest' });
 });
 app.post('/generateWinPackage', async (req, res) => {
     return;

diff --git a/dist/openai.js b/dist/openai.js
@@ -5,16 +5,27 @@ export async function initOpenAI() {
     });
     return new OpenAIApi(configuration);
 }
+export function manifestPrompt(html) {
+    const prompt = `INSTRUCTION
+parse this html and generate web manifest with this string properties: lang,name[max lenght 45],short_name[max lenght 12],description[max lenght 160],background_color[hex],dir,display,orientation,scope,start_url,theme_color[hex]. And array properties: categories[max lenght 3],icons[max lenght 1]. Take all information and icons from html, use only biggest icon in array of icons.
+HTML
+${html}
+OUTPUT FILE
+web app manifests in JSON format no new line or return symbols, compact: <code>`;
+    return prompt;
+}
 export async function askOpenAI(prompt, openai) {
-    let gptAnswer = null;
+    let davinciAnswer = null;
+    console.log(`prompt: ${prompt}`);
     try {
         const completion = await openai.createCompletion({
-            model: "text-davinci-003",
-            prompt: `parse this html and generate web manifest in JSON format, silent answer with no comments or explains. Take all information and icons from html tags, use only one biggest icon in array of icons. ${prompt}`,
-            temperature: 0.6,
+            model: "code-davinci-002",
+            prompt,
+            temperature: 0.1,
+            max_tokens: 256,
         });
-        gptAnswer = completion.data;
-        console.warn(gptAnswer);
+        davinciAnswer = completion.data;
+        console.log(`answer: ${JSON.stringify(davinciAnswer)}`);
     }
     catch (error) {
         if (error?.response) {
@@ -25,10 +36,12 @@ export async function askOpenAI(prompt, openai) {
         }
     }
     let manifest = null;
-    if (gptAnswer && gptAnswer.choices[0].text)
+    if (davinciAnswer?.choices && davinciAnswer.choices[0]?.text) {
+        const restoredResponse = `<code>${davinciAnswer.choices[0].text}`;
         try {
-            manifest = JSON.parse(gptAnswer.choices[0].text);
+            manifest = JSON.parse(restoredResponse.match(/<code>(.*)<\/code>/)[1]);
         }
         catch (error) { }
+    }
     return manifest;
 }