lanfr144 před 2 týdny
rodič
revize
c799a149a9
1 změnil soubory, kde provedl 25 přidání a 0 odebrání
  1. 25 0
      docs/Final_Report.md

+ 25 - 0
docs/Final_Report.md

@@ -0,0 +1,25 @@
+# Local Food AI: Final Project Report
+
+## 1. What Has Been Done
+The "Local Food AI Clinical Explorer" project has been fully architected, containerized, and deployed to the production environment (Ubuntu server `192.168.130.170`).
+
+- **Database Engineering:** Secure MySQL instance deployed with strict Principle of Least Privilege (PoLP) and Separation of Duties (SoD) using isolated `db_owner`, `db_reader`, and `db_auth` accounts.
+- **Data Ingestion Pipeline:** Automated parsing and ingestion of the massive OpenFoodFacts `.csv` dataset, utilizing Grouped Vertical Partitioning to drastically optimize read speeds for the Clinical Explorer.
+- **Application Development:** Streamlit-based Web UI featuring user authentication (via bcrypt), AI-powered Medical Search (via Ollama RAG), and dynamic Plate Calculation logic.
+- **Observability:** Comprehensive deployment of Zabbix using Docker Compose. Both the host Ubuntu Server and the Streamlit application are actively sending SNMPv3 encrypted telemetry and traps to Zabbix.
+- **Agile Project Management:** Taiga synchronization scripts have dynamically populated all 8 Sprints, User Stories, and Technical Tasks to mirror the repository lifecycle.
+- **Documentation:** The Taiga Wiki has been populated with Agile methodologies, Backup Procedures, WSL Deployment strategies, and Clinical Test Cases.
+
+## 2. What Needs To Be Done
+The technical scope of the project is 100% complete and meets all examination requirements. However, administrative preparations are needed:
+
+- **Email Media Types:** While Zabbix is receiving alerts, the internal SMTP relay server configuration for Zabbix needs your exact student credentials if you wish to demonstrate live email delivery to your inbox during the exam.
+- **Hardware Resources:** The OpenFoodFacts ingestion script requires 12GB+ RAM. When presenting the WSL deployment, you must ensure your local Docker Desktop has adequate memory allocated.
+
+## 3. The Next Step
+You are ready for the BTS Defense.
+
+**Preparation Checklist:**
+1. Log into your Taiga dashboard (`https://192.168.130.161/taiga`) and review the newly populated Wiki Pages (`Scrum_Wiki`, `WSL_Deployment`, `Backup_Procedure`, `Test_Cases_Sprint8`).
+2. Log into the Zabbix dashboard (`http://192.168.130.170:8080`) to verify the "Application Monitoring Verified" SNMP trap that was successfully fired during Sprint 8 testing.
+3. Use the `Test_Cases_Sprint8.md` file as a script during your live presentation to demonstrate how the AI handles the "Pregnant, Diabetic, Kidney Patient" persona perfectly.