fix category

app.config.ts

@@ -36,6 +36,7 @@ export default {
 			firebaseStorageBucket: process.env.FIREBASE_STORAGE_BUCKET,
 			firebaseMessagingSenderId: process.env.FIREBASE_MESSAGING_SENDER_ID,
 			firebaseAppId: process.env.FIREBASE_APP_ID,
+			openAPIKey: process.env.OPENAI_API_KEY,
 			eas: {},
 		},
 	},

metro.config.js

@@ -5,3 +5,16 @@ const defaultConfig = getDefaultConfig(__dirname);
 defaultConfig.resolver.assetExts.push("cjs");
 
 module.exports = defaultConfig;
+// metro.config.js
+module.exports = {
+    server: {
+      rewriteRequestUrl: (url) => {
+        if (!url.endsWith('.bundle')) {
+          return url;
+        }
+        // https://github.com/facebook/react-native/issues/36794
+        // JavaScriptCore strips query strings, so try to re-add them with a best guess.
+        return url + '?platform=ios&dev=true&minify=false&modulesOnly=false&runModule=true';
+      }, // ...
+    }, // ...
+  }

src/types/index.ts

@@ -21,6 +21,7 @@ export interface SelfPost {
 	authorUid: string;
 	timestamp: Timestamp;
 	content: string;
+	category: string;
 }
 
 export interface SocialPost {

src/utils/hooks/Vector.tsx

@@ -0,0 +1,116 @@
+import React from "react";
+import Constants from "expo-constants";
+
+
+
+export default async function categorise_vector(prompt:string){
+    const open_api = Constants.manifest?.extra?.openAPIKey;
+    console.log(open_api);
+    const api_url = "https://api.openai.com/v1/embeddings";
+    const model = "text-embedding-ada-002";
+    const get_embedding = async () => {
+        try{
+            const response = await fetch(api_url, {
+                headers:{
+                    'Authorization': `Bearer ${open_api}`,
+                    'Content-Type': 'application/json'
+                },
+                body : JSON.stringify({
+                    'model' : model,
+                    'input' : prompt,
+                }),
+                method: 'POST',
+            })
+            const data = await response.json();
+            return (data.data[0].embedding);
+        } catch (error){
+            console.log(error);
+            return null;
+        }
+    }
+    const data = await get_embedding();
+    return data;
+}
+
+// calculate the sim of two vectors (not that fast but its fine)
+function cosineSimilarity(vectorA: Array<number>, vectorB: Array<number>) {
+    const dotProduct = vectorA.reduce((acc, val, index) => acc + val * vectorB[index], 0);
+    const normA = Math.sqrt(vectorA.reduce((acc, val) => acc + val * val, 0));
+    const normB = Math.sqrt(vectorB.reduce((acc, val) => acc + val * val, 0));
+    return dotProduct / (normA * normB);
+}
+
+// returns true if the key was sucessfully added, false if the key already exists
+export async function add_on_standard_embeddings(standard_embeddings: any, new_word: string){
+    if(standard_embeddings.hasOwnProperty(new_word)){
+        return false;
+    } else {
+        const new_embedding = categorise_vector(new_word);
+        standard_embeddings[new_word] = new_embedding;
+        return true;
+    }
+}
+
+// will retrieve the closest word from a list of standard embeddings
+export function get_closest_category(embedding: Array<number>, standard_embeddings: any){
+    let keys = Object.keys(standard_embeddings);
+    let closestWord = keys[0];
+    let highestSim = -Infinity;
+    for(let i = 0; i < keys.length; ++i){
+        const similarity = cosineSimilarity(embedding, standard_embeddings[keys[i]]);
+        if(similarity > highestSim){
+            highestSim = similarity;
+            closestWord = keys[i];
+        }
+    }
+    return closestWord;
+}
+
+function lower_token_sample(list_of_thanks: Array<string>, constant = 60){
+    if(list_of_thanks.length > constant){
+        let n = constant;
+        let result = new Array(n),
+        len = list_of_thanks.length,
+        taken = new Array(len);
+        while (n--) {
+            var x = Math.floor(Math.random() * len);
+            result[n] = list_of_thanks[x in taken ? taken[x] : x];
+            taken[x] = --len in taken ? taken[len] : len;
+        }
+        return result;
+    } else {
+        return list_of_thanks;
+    }
+}
+
+export async function get_summary(list_of_thanks: Array<string>){
+    const open_api = Constants.manifest?.extra?.openAPIKey;
+    const api_url = "https://api.openai.com/v1/chat/completions";
+    const model = "gpt-3.5-turbo";
+    const prompt = "This is a list of things I am thankful for: " 
+    + JSON.stringify(lower_token_sample(list_of_thanks)) + ". Summarise this into a short sentence capturing 4-5 key ideas. Refer to me as you.";
+
+    const get_summary = async () => {
+        try{
+            const response = await fetch(api_url, {
+                headers:{
+                    'Authorization': `Bearer ${open_api}`,
+                    'Content-Type': 'application/json'
+                },
+                body : JSON.stringify({
+                    'model' : model,
+                    'messages' : [{'role' : 'user', 'content' : prompt}],
+                }),
+                method: 'POST',
+            })
+            const data = await response.json();
+            console.log(data.choices[0].message.content);
+            return (data.choices[0].message.content);
+        } catch (error){
+            console.log(error);
+            return null;
+        }
+    }
+    const data = await get_summary();
+    return data;
+}

src/utils/hooks/usePost.tsx

@@ -1,4 +1,6 @@
 import { useEffect, useState } from "react";
+import categorise_vector, {get_closest_category} from "./Vector";
+import word_embeddings from "./word_embeddings.json";
 
 import {} from "firebase/auth";
 import { FirestoreUser, SelfPost, SocialPost } from "src/types";
@@ -44,13 +46,17 @@ export function usePost() {
 	const makeSelfPost = async (content: string) => {
 		await fetchFireUser();
 		if (!fireUser) return;
+		const embedding = await categorise_vector(content);
+		const category = await get_closest_category(embedding, word_embeddings);
 
 		const selfPost: SelfPost = {
 			uid: uuidv4(),
 			authorUid: fireUser?.uid,
 			timestamp: Timestamp.now(),
 			content: content,
+			category: category,
 		};
+		console.log(category);
 
 		// add self post to self posts collection
 		await setDoc(doc(db, "selfPosts", selfPost.uid), selfPost);