Upload upload_hf.py with huggingface_hub

upload_hf.py

@@ -0,0 +1,168 @@
+"""
+upload_hf.py
+============
+Carica su HuggingFace solo i file necessari del progetto GenerAI.
+
+Uso:
+    python upload_hf.py
+    python upload_hf.py --repo amogaddy/GenerAI --model-weights
+"""
+
+import argparse
+import os
+import shutil
+import tempfile
+from pathlib import Path
+
+# Carica .env se presente
+if Path(".env").exists():
+    for line in Path(".env").read_text(encoding="utf-8").splitlines():
+        if "=" in line and not line.startswith("#"):
+            k, v = line.split("=", 1)
+            os.environ.setdefault(k.strip(), v.strip())
+
+# File del codice da includere
+SOURCE_FILES = [
+    "app.py",
+    "brain.py",
+    "knowledge_base.py",
+    "scraper.py",
+    "errors.py",
+    "seed_italian.py",
+    "export_dataset.py",
+    "finetune.py",
+    "upload_hf.py",
+    "requirements.txt",
+    "requirements-train.txt",
+]
+
+DEFAULT_REPO = "amogaddy/GenerAI"
+MODEL_DIR    = "./generai-finetuned"
+
+
+def check_deps():
+    try:
+        from huggingface_hub import login, upload_folder, HfApi
+        return login, upload_folder, HfApi
+    except ImportError:
+        print("❌ huggingface_hub non installato.")
+        print("   pip install huggingface-hub")
+        exit(1)
+
+
+def build_model_card(repo_id: str, include_weights: bool) -> str:
+    return f"""---
+license: mit
+base_model: Qwen/Qwen2.5-0.5B-Instruct
+language:
+  - it
+tags:
+  - italian
+  - generai
+  - fine-tuned
+  - rag
+---
+
+# GenerAI 🤖
+
+Assistente AI in italiano con memoria locale (ChromaDB) e ricerca web.
+
+## Come usarlo
+
+```bash
+pip install -r requirements.txt
+python app.py
+```
+
+## Fine-tuning
+
+```bash
+pip install -r requirements-train.txt
+python export_dataset.py
+python finetune.py --dataset dataset.jsonl --hf-repo {repo_id}
+```
+
+## Licenza
+
+MIT — fai quello che vuoi, basta lasciare il credito.
+"""
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Carica GenerAI su HuggingFace")
+    parser.add_argument("--repo",          default=DEFAULT_REPO, help=f"repo_id HuggingFace (default: {DEFAULT_REPO})")
+    parser.add_argument("--model-weights", action="store_true",  help="Includi anche i pesi del modello fine-tunato")
+    parser.add_argument("--no-login",      action="store_true",  help="Salta il login (usa token già salvato)")
+    args = parser.parse_args()
+
+    login, upload_folder, HfApi = check_deps()
+
+    # ── Login ──────────────────────────────────────────────────────────────────
+    if not args.no_login:
+        token = os.environ.get("HF_TOKEN", "")
+        if token:
+            print("Login HuggingFace con token da .env...")
+            login(token=token)
+        else:
+            print("Login HuggingFace...")
+            login()
+
+    # ── Crea cartella temporanea con solo i file necessari ─────────────────────
+    with tempfile.TemporaryDirectory() as tmp:
+        tmp_path = Path(tmp)
+        copied = []
+        missing = []
+
+        for fname in SOURCE_FILES:
+            src = Path(fname)
+            if src.exists():
+                shutil.copy2(src, tmp_path / src.name)
+                copied.append(fname)
+            else:
+                missing.append(fname)
+
+        # Genera README.md / model card
+        (tmp_path / "README.md").write_text(
+            build_model_card(args.repo, args.model_weights),
+            encoding="utf-8",
+        )
+        copied.append("README.md (generato)")
+
+        print(f"\nFile da caricare ({len(copied)}):")
+        for f in copied:
+            print(f"  + {f}")
+        if missing:
+            print(f"\nFile non trovati (saltati):")
+            for f in missing:
+                print(f"  - {f}")
+
+        # ── Upload codice ──────────────────────────────────────────────────────
+        print(f"\nUpload codice -> {args.repo} ...")
+        upload_folder(
+            folder_path=str(tmp_path),
+            repo_id=args.repo,
+            repo_type="model",
+            commit_message="Upload GenerAI — codice + grammatica italiana",
+        )
+        print("Codice caricato.")
+
+    # ── Upload pesi modello (opzionale) ────────────────────────────────────────
+    if args.model_weights:
+        if not Path(MODEL_DIR).exists():
+            print(f"\nCartella pesi non trovata: {MODEL_DIR}")
+            print("  Esegui prima: python finetune.py --dataset dataset.jsonl")
+        else:
+            print(f"\nUpload pesi modello -> {args.repo} ...")
+            upload_folder(
+                folder_path=MODEL_DIR,
+                repo_id=args.repo,
+                repo_type="model",
+                commit_message="Upload pesi modello fine-tunato GenerAI",
+            )
+            print("Pesi modello caricati.")
+
+    print(f"\nCompletato! -> https://huggingface.co/{args.repo}")
+
+
+if __name__ == "__main__":
+    main()