Fix audit findings: budget editing, dead code, logging, multi-currency

- Add budget editing: updateBudget() API, edit button on budget cards,
  BudgetFormModal adapted for create/update (category locked on edit)
- Remove permanently-broken POST /auth/totp/verify stub and its unused
  TOTPVerifyRequest schema
- Wire getHoldingTransactions() to AssetDetail page — transaction history
  table now shows above the candlestick chart, sorted newest-first
- Fix multi-currency net worth in account_service: account balances are
  now converted to base_currency via ExchangeRate table before summing
- Replace silent bare pass exception handlers with logger.warning() in
  transactions.py (OCR/AI pipeline) and price_feed_service.py (search)
  — ValueError in date/number regex parsing left silent (control flow)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/app/api/v1/transactions.py

@@ -1,11 +1,14 @@
 import csv
 import io
+import logging
 import mimetypes
 import os
 import uuid
 from pathlib import Path
 from typing import Annotated
 
+logger = logging.getLogger(__name__)
+
 from fastapi import APIRouter, Depends, File, Form, HTTPException, Query, UploadFile
 from fastapi.responses import FileResponse
 from sqlalchemy.ext.asyncio import AsyncSession
@@ -265,8 +268,8 @@ async def delete_attachment(
     path = Path(settings.upload_dir) / str(user.id) / ref["stored_name"]
     try:
         path.unlink(missing_ok=True)
-    except OSError:
-        pass
+    except OSError as e:
+        logger.warning("Could not delete attachment file %s: %s", path, e)
 
     new_refs = [r for r in refs if r["id"] != attachment_id]
     await db.execute(
@@ -307,8 +310,8 @@ def _extract_ocr_text(file_bytes: bytes, mime_type: str) -> str:
             text = "\n".join(pages_text).strip()
             if text:
                 return text
-        except Exception:
-            pass
+        except Exception as e:
+            logger.warning("pdfplumber text extraction failed: %s", e)
         # Scanned PDF — convert first page to image then OCR
         try:
             from pdf2image import convert_from_bytes
@@ -316,8 +319,8 @@ def _extract_ocr_text(file_bytes: bytes, mime_type: str) -> str:
             images = convert_from_bytes(file_bytes, first_page=1, last_page=1, dpi=200)
             if images:
                 return pytesseract.image_to_string(images[0])
-        except Exception:
-            pass
+        except Exception as e:
+            logger.warning("pdf2image/tesseract OCR failed: %s", e)
         return ""
     else:
         import io
@@ -326,7 +329,8 @@ def _extract_ocr_text(file_bytes: bytes, mime_type: str) -> str:
         try:
             img = Image.open(io.BytesIO(file_bytes))
             return pytesseract.image_to_string(img)
-        except Exception:
+        except Exception as e:
+            logger.warning("Image OCR failed: %s", e)
             return ""
 
 
@@ -488,11 +492,10 @@ async def _call_ai_parse(file_bytes: bytes, mime_type: str, user_row) -> dict:
                         "ocr_text":    ocr_text,
                     }
                 except json.JSONDecodeError:
-                    # AI returned something non-JSON — fall through to rules, keep raw for debug
-                    pass
+                    logger.warning("AI returned non-JSON response, falling back to rule-based parser")
 
-        except (httpx.HTTPStatusError, httpx.RequestError):
-            pass  # fall through to rule-based
+        except (httpx.HTTPStatusError, httpx.RequestError) as e:
+            logger.warning("AI API request failed (%s), falling back to rule-based parser", type(e).__name__)
 
     # Step 3: rule-based fallback (also used when AI is not configured)
     if ocr_text.strip():