From e59cf0274c50b7da1c837ea1d02742b8f554e440 Mon Sep 17 00:00:00 2001
From: Saul <158204665+UO294936@users.noreply.github.com>
Date: Mon, 10 Feb 2025 08:18:44 +0100
Subject: [PATCH] Create LLM Comparison file in /docs

Perhaps we should move this file to another folder, to maintain a well organized storage structure.
---
 docs/LLM_Comparison.adoc | 110 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 110 insertions(+)
 create mode 100644 docs/LLM_Comparison.adoc

diff --git a/docs/LLM_Comparison.adoc b/docs/LLM_Comparison.adoc
new file mode 100644
index 00000000..78a10f95
--- /dev/null
+++ b/docs/LLM_Comparison.adoc
@@ -0,0 +1,110 @@
+= Pros and Cons List
+
+== 1. Gemini API (by Google)
+
+=== Pros:
+* *Advanced AI Capabilities:* Strong at understanding and generating human-like text, making it useful for giving hints dynamically.
+* *Context Awareness:* Can remember previous interactions in a conversation, making it helpful for progressive hints.
+* *Multimodal Support:* Can process text, images, and even voice (depending on the version).
+* *Google Integration:* If your app involves other Google services (like Firebase or Google Cloud), integration is smoother.
+
+=== Cons:
+* *Pricing:* Can become expensive depending on usage volume.
+* *Latency:* Depending on API usage and complexity, response time might not be instant.
+* *API Limits:* May have restrictions on request frequency and token limits.
+
+== 2. Empathy API
+
+=== Pros:
+* *Emotion & Sentiment Analysis:* Focuses on understanding user emotions (could be useful for adjusting hint difficulty based on frustration level).
+* *Customization:* Some versions allow fine-tuning responses for more user-friendly interactions.
+* *Ethical AI Focus:* Designed to be safe, transparent, and aligned with responsible AI practices.
+
+=== Cons:
+* *Limited AI Capabilities:* Less advanced than Gemini in complex reasoning and providing contextual hints dynamically.
+* *Less Widespread Support:* Fewer developer resources and documentation compared to Google’s AI ecosystem.
+* *Scalability:* Might not handle a high number of concurrent users as efficiently as Gemini.
+
+== Implementation Difficulty: Gemini API vs. Empathy API
+
+[options="header"]
+|===
+| Feature | Gemini API | Empathy API
+| *Ease of Setup* | 🟡 Moderate – Requires Google API Key, setting up permissions in Google Cloud | 🟢 Easier – Designed for user-friendly integration with simple REST calls
+| *Documentation & Support* | 🟢 Excellent – Extensive Google support & community | 🟡 Limited – Fewer resources, but still documented
+| *Integration Complexity* | 🔴 More complex – Uses Google AI SDKs, may require Firebase integration for persistence | 🟢 Simpler – Mostly REST API-based, easier to use directly in a web app
+| *Customization* | 🟢 Highly customizable – Can fine-tune responses | 🟡 Limited – More focused on sentiment analysis
+| *Response Speed* | 🟡 Can be slow for complex queries | 🟢 Generally faster for simple interactions
+|===
+
+== Considering Ease of Implementation
+
+* An easy-to-integrate solution → *Empathy API* is simpler to implement with basic REST API calls.
+* Most powerful AI with more control over responses → *Gemini API* is better, but requires more setup (Google Cloud configuration, API key, potential SDK use).
+
+== Usage Example
+
+=== 1. Gemini API (Google AI) Example
+Google's Gemini API requires Google Cloud setup, an API key, and uses a RESTful API or Python SDK.
+
+==== Setup for Gemini API
+. Get an API key from Google AI Studio or Google Cloud Console.
+. Enable Gemini API in Google Cloud.
+. Use a REST request to send queries.
+
+[source,javascript]
+----
+const axios = require("axios");
+
+// Replace with your actual Gemini API key
+const API_KEY = "YOUR_GEMINI_API_KEY";
+const API_URL = `https://generativelanguage.googleapis.com/v1beta/models/gemini-pro:generateText?key=${API_KEY}`;
+
+async function getHint(question) {
+    try {
+        const response = await axios.post(API_URL, {
+            prompt: { text: `Provide a hint for this question: ${question}` }
+        });
+
+        console.log("Hint:", response.data.candidates[0].output);
+    } catch (error) {
+        console.error("Error fetching hint:", error.message);
+    }
+}
+
+// Example usage
+getHint("What is the capital of France?");
+----
+
+=== 2. Empathy API Example
+Empathy API is easier to use and mainly focuses on sentiment-based interactions.
+
+==== Setup for Empathy API
+. Sign up on Empathy API’s website to get an API key.
+. Use a REST request to send queries.
+
+[source,javascript]
+----
+const axios = require("axios");
+
+// Replace with your actual Empathy API key
+const API_KEY = "YOUR_EMPATHY_API_KEY";
+const API_URL = "https://api.empathy.co/hints";
+
+async function getHint(question) {
+    try {
+        const response = await axios.post(API_URL, {
+            query: question,
+            apiKey: API_KEY
+        });
+
+        console.log("Hint:", response.data.hint);
+    } catch (error) {
+        console.error("Error fetching hint:", error.message);
+    }
+}
+
+// Example usage
+getHint("What is the capital of France?");
+----
+