feat: Bug Recommendation system using vector cosine-similarity

.gitignore

@@ -12,4 +12,7 @@ dist-ssr/
 .DS_Store
 Thumbs.db
 
-.vscode/
+.vscode/
+
+__pycache__/
+*.pyc

ai-service/app.py

@@ -0,0 +1,26 @@
+import os
+from dotenv import load_dotenv
+from flask import Flask, request, jsonify
+from model import get_embedding
+from utils import build_text
+
+load_dotenv()
+app = Flask(__name__)
+
+@app.route("/embed", methods=["POST"])
+def embed():
+    try:
+        data = request.json
+        if not data:
+            return jsonify({"error": "Invalid input"}), 400
+        text = build_text(data)
+        embedding = get_embedding(text)
+        return jsonify({"embedding": embedding}), 200
+    except Exception as e:
+        print("Error in /embed:", str(e))
+        return jsonify({"error": "Internal server error"}), 500
+
+if __name__ == "__main__":
+    PORT = int(os.getenv("PORT", 3000))
+    DEBUG = os.getenv("FLASK_DEBUG", "false").lower() in ["true", "1", "yes"]
+    app.run(port=PORT, debug=DEBUG)

ai-service/model.py

@@ -0,0 +1,6 @@
+from sentence_transformers import SentenceTransformer
+
+model = SentenceTransformer('all-MiniLM-L6-v2')
+
+def get_embedding(text):
+    return model.encode(text).tolist()

ai-service/requirements.txt

@@ -0,0 +1,5 @@
+flask
+sentence-transformers
+numpy
+scikit-learn
+python-dotenv

ai-service/utils.py

@@ -0,0 +1,21 @@
+def safe_join(value):
+    if isinstance(value, list):
+        return " ".join(value)
+    if isinstance(value, str):
+        return value
+    return ""
+def build_text(data):
+    title = data.get("title", "")
+    description = data.get("description", "")
+    tags_text = safe_join(data.get("tags"))
+    tech_text = safe_join(data.get("techStack"))
+    difficulty = data.get("difficulty", "")
+
+    text = f"""
+    BUG TITLE: {title}. {title}.
+    BUG DESCRIPTION: {description}.
+    BUG TAGS: {tags_text}.
+    TECHNOLOGIES: {tech_text}.
+    DIFFICULTY LEVEL: {difficulty}.
+    """
+    return text.lower().strip()

core-services/package.json

@@ -13,6 +13,7 @@
   "license": "ISC",
   "type": "commonjs",
   "dependencies": {
+    "axios": "^1.14.0",
     "cors": "^2.8.6",
     "dotenv": "^17.3.1",
     "express": "^5.2.1",

core-services/src/app.js

@@ -1,5 +1,7 @@
 const express = require("express");
 const cors = require("cors");
+const vectorRouter = require("./routes/vector.Routes");
+const recommendationRouter = require("./routes/recommendation.Routes");
 
 const app = express();
 
@@ -10,4 +12,7 @@ app.get("/", (req, res) => {
   res.send("Backend is running!!");
 });
 
+app.use("/api/vectors", vectorRouter);
+app.use("/api/recommendations", recommendationRouter);
+
 module.exports = app;

core-services/src/controllers/recommendation.Controller.js

@@ -0,0 +1,49 @@
+const Bug = require("../models/Bug");
+const Vector = require("../models/Vector");
+const { cosineSimilarity } = require("../utils/similarityScore");
+const mongoose = require("mongoose");
+
+exports.getSimilarBugs = async (req, res) => {
+  try {
+    const bugId = req.params.id;
+    if (!mongoose.Types.ObjectId.isValid(bugId)) {
+      return res.status(400).json({ message: "Invalid bugId" });
+    }
+
+    const currentVectorDoc = await Vector.findOne({ bugId });
+
+    if (!currentVectorDoc) {
+      return res.status(404).json({ message: "Vector not found" });
+    }
+    const currentVector = currentVectorDoc.vector;
+    const allVectors = await Vector.find({ bugId: { $ne: bugId } });
+    const similarities = allVectors.map((item) => ({
+      bugId: item.bugId,
+      score: cosineSimilarity(currentVector, item.vector)
+    }));
+    const top5 = similarities
+      .sort((a, b) => b.score - a.score)
+      .slice(0, 5);
+    const bugs = await Bug.find({
+      _id: { $in: top5.map(i => i.bugId) }
+    });
+    const bugMap = new Map();
+    bugs.forEach(b => bugMap.set(b._id.toString(), b));
+
+    const finalRecommendations = top5.map(item => {
+      const bug = bugMap.get(item.bugId.toString());
+      if (!bug) return null;
+      return {
+        ...bug.toObject(),
+        similarityScore: item.score
+      };
+    }).filter(Boolean);
+    res.json({
+      recommendations: finalRecommendations
+    });
+
+  } catch (err) {
+    console.error(err);
+    res.status(500).json({ message: "Error fetching recommendations" });
+  }
+};

core-services/src/controllers/vector.Controller.js

@@ -0,0 +1,46 @@
+const mongoose = require("mongoose");
+const Bug = require("../models/Bug");
+const Vector = require("../models/Vector");
+const { getEmbedding } = require("../services/embedding.Service");
+
+exports.createVector = async (req, res) => {
+  try {
+    const { bugId } = req.body;
+
+    if (!bugId) {
+      return res.status(400).json({ message: "bugId is required" });
+    }
+
+    if (!mongoose.Types.ObjectId.isValid(bugId)) {
+      return res.status(400).json({ message: "Invalid bugId format" });
+    }
+
+    const bug = await Bug.findById(bugId);
+
+    if (!bug) {
+      return res.status(404).json({ message: "Bug not found" });
+    }
+
+    const embedding = await getEmbedding(bug);
+
+    const vectorDoc = await Vector.findOneAndUpdate(
+      { bugId: bug._id },
+      {
+        bugId: bug._id,
+        vector: embedding,
+        modelVersion: "all-MiniLM-L6-v2",
+        updatedAt: new Date(),
+      },
+      { upsert: true, returnDocument: "after"}
+    );
+
+    return res.status(200).json({
+      message: "Vector created successfully",
+      data: vectorDoc,
+    });
+
+  } catch (error) {
+    console.error("VECTOR ERROR:", error);
+    return res.status(500).json({ message: "Error creating vector" });
+  }
+};