fix(async): chiamate bloccanti in asyncio.to_thread (#36 punto 1)

kaleido (pio.to_image), WeasyPrint (write_pdf), pdfplumber (extract_text_from_pdf)
e la scrittura del CSV di produzione giravano sincrone dentro handler async:
un PDF pesante o un run grande fermava il worker per tutte le altre richieste.
Ora avvolte in await asyncio.to_thread(...) in report_service.py,
ai_service.py e production_export_service.py.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

src/backend/services/ai_service.py

@@ -1,4 +1,5 @@
 """AI service for parsing technical sheets via OpenRouter."""
+import asyncio
 import json
 import logging
 from io import BytesIO
@@ -36,7 +37,9 @@ def extract_text_from_pdf(pdf_bytes: bytes) -> str:
 
 
 async def parse_technical_sheet(pdf_bytes: bytes) -> list[dict]:
-    text = extract_text_from_pdf(pdf_bytes)
+    # pdfplumber parses synchronously and can take a while on a big PDF: off
+    # the event loop, or one upload stalls every other request being served.
+    text = await asyncio.to_thread(extract_text_from_pdf, pdf_bytes)
     if not text.strip():
         return []