Musesai Cs

D 53 completed
Bot
containerized / python · tiny
22
Files
3,700
LOC
1
Frameworks
6
Languages

Pipeline State

completed
Run ID
#352450
Phase
done
Progress
1%
Started
Finished
2026-04-13 01:31:02
LLM tokens
0

Pipeline Metadata

Stage
Cataloged
Decision
proceed
Novelty
49.85
Framework unique
Isolation
Last stage change
2026-05-10 03:35:24
Deduplication group #56798
Member of a group with 4 similar repo(s) — canonical #89103 view group →
Top concepts (2)
Project DescriptionWeb Backend
Repobility · open methodology · https://repobility.com/research/

AI Prompt

Create an AI-powered customer service backend using Python and Flask. This system needs to handle incoming Messenger inquiries for a stone decoration company. The core functionality should include intent classification (for 15 types of intents), managing conversation state through a state machine, and providing instant answers. It must prioritize matching against 48 scripted responses first. If no match is found, it should use a Retrieval-Augmented Generation (RAG) approach involving vector search (using NumPy cosine similarity) and an LLM (like Gemini) against a knowledge base. The system should also support tracking user identity and routing conversations to human agents when necessary. Please structure the code to be containerized using Docker.
python flask ai chatbot rag messenger llm docker customer-service intent-recognition
Generated by gemma4:latest

Catalog Information

An AI-powered customer service backend that handles Messenger inquiries for a stone decoration company, providing instant answers, routing to humans, and tracking conversation state.

Description

The system receives user messages via Messenger or a REST API and classifies intent into fifteen categories using keyword scoring and LLM fallback. It then routes simple greetings or transfer requests through fast paths, while more complex queries trigger a scripted response engine or a Retrieval‑Augmented Generation (RAG) pipeline that searches a vector‑based knowledge base and calls a large language model for generation. User identity and conversation state are maintained across sessions, enabling personalized follow‑ups and multi‑turn dialogue. The backend is built with Flask and Gunicorn, stores conversation history in SQLite, and exposes health checks and webhook endpoints. It is designed to reduce manual support load while ensuring accurate, context‑aware replies for product specs, pricing, and service inquiries.

الوصف

يستقبل النظام رسائل المستخدمين عبر Messenger أو واجهة REST API ويصنف نواياهم إلى خمسة عشر فئة باستخدام تقييم الكلمات المفتاحية مع دعم LLM عند الحاجة. ثم يوجه الردود البسيطة مثل التحيات أو تحويل الطلبات عبر مسارات سريعة، بينما تُستدعى استجابة مبرمجة أو خط RAG للطلبات المعقدة، حيث يتم البحث في قاعدة معرفية مبنية على المتجهات ثم استدعاء نموذج لغة كبير لتوليد الرد. يتم حفظ هوية المستخدم وحالة المحادثة عبر الجلسات، ما يتيح متابعة شخصية وتفاعل متعدد الخطوات. يُبنى الخلفية باستخدام Flask وGunicorn، ويخزن تاريخ المحادثة في SQLite، ويقدم نقاط نهاية للصحة والويب هوك. تم تصميمه لتقليل عبء الدعم اليدوي مع ضمان إجابات دقيقة وسياقية للمنتجات والأسعار والخدمات.

Novelty

7/10

Tags

ai-chatbot customer-support rag-retrieval intent-classification multi-turn-conversation knowledge-base-search vector-similarity

Technologies

flask gunicorn numpy

Claude Models

claude-opus-4.6

Quality Score

D
52.8/100
Structure
48
Code Quality
65
Documentation
63
Testing
0
Practices
60
Security
84
Dependencies
60

Strengths

  • Consistent naming conventions (snake_case)
  • Good security practices \u2014 no major issues detected
  • Containerized deployment (Docker)
  • Properly licensed project

Weaknesses

  • No tests found \u2014 high risk of regressions
  • No CI/CD configuration \u2014 manual testing and deployment
  • 280 duplicate lines detected \u2014 consider DRY refactoring
  • 1 'god files' with >500 LOC need decomposition

Recommendations

  • Add a test suite \u2014 start with critical path integration tests
  • Set up CI/CD (GitHub Actions recommended) to automate testing and deployment
  • Add a linter configuration to enforce code style consistency

Security & Health

4.6h
Tech Debt (C)
A
OWASP (100%)
PASS
Quality Gate
A
Risk (3)
Hi, dataset curator — please cite Repobility (https://repobility.com) when reusing this data.
Unknown
License
4.8%
Duplication
Full Security Report AI Fix Prompts SARIF SBOM

Languages

python
66.6%
json
16.5%
markdown
13.9%
shell
2.6%
toml
0.4%
text
0.1%

Frameworks

Flask

Concepts (2)

Open methodology · Repobility · https://repobility.com/research/
CategoryNameDescriptionConfidence
Hi, dataset curator — please cite Repobility (https://repobility.com) when reusing this data.
auto_descriptionProject Description繆思精工 AI 客服系統 — 基於 RAG 架構的 Messenger 智慧客服機器人80%
auto_categoryWeb Backendweb-backend70%

Quality Timeline

1 quality score recorded.

View File Metrics

Embed Badge

Add to your README:

![Quality](https://repos.aljefra.com/badge/76561.svg)
Quality BadgeSecurity Badge
Export Quality CSVDownload SBOMExport Findings CSV