بناء مجموعة بيانات NER عربية للأخبار (Label Studio + spaCy)

flowchart RL S1["1. Scrape news to JSONL"] S2["2. Import to Label Studio"] S3["3. Define labels: PERSON/ORG/LOC/EVENT"] S4["4. Assign annotators"] S5["5. Calculate inter-annotator agreement"] S6["6. Export JSON-MIN"] S7["7. Convert to spaCy format"] S8["8. Train + evaluate"] S9["9. Deploy as REST"] S1 --> S2 S2 --> S3 S3 --> S4 S4 --> S5 S5 --> S6 S6 --> S7 S7 --> S8 S8 --> S9 subgraph Tools["🔧 الأدوات المستخدمة"] T0(("Label Studio")) T1(("spaCy")) T2(("Postgres")) T3(("MinIO")) end style Tools fill:#fffbeb,stroke:#f59e0b,stroke-width:1px