fix: Update stats for multi-level drill-down (#62)

The transaction count and in/out totals in the top-right stats bar now
correctly reflect all drill-down filters in multi-level navigation.

Previously, drilling from e.g. "Amazon > Groceries" would still show
stats for all Amazon transactions. The bug was in get_filtered_df()
which used an elif chain that only applied the first filter. Changed to
sequential if statements to apply all dimension filters.

Also updated CLAUDE.md to explicitly prohibit git pull without
permission.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

Author: wesm

CLAUDE.md

@@ -2,18 +2,23 @@
 
 ## CRITICAL: Git Branch Management for AI Assistants
 
-**⚠️ NEVER change git branches or create new branches without explicit user permission.**
+**⚠️ NEVER change git branches, pull, or create new branches without explicit user permission.**
 
 - ✅ **ALWAYS ask before** `git checkout <branch>`
+- ✅ **ALWAYS ask before** `git pull` or `git fetch` followed by merge/rebase
 - ✅ **ALWAYS ask before** creating new branches
 - ✅ **Stay on the branch the user checked out** unless they explicitly ask you to switch
 - ❌ **NEVER run `git checkout` on your own**
+- ❌ **NEVER run `git pull` on your own**
 - ❌ **NEVER create branches autonomously**
+- ❌ **NEVER switch branches when starting a new task** - the user has already set up the branch
 
 **If you need to work on a different branch**, ask the user first:
 - "Should I switch to branch X to work on this?"
 - "Should I create a new branch for this feature?"
 
+**When the user says a PR is merged**, do NOT automatically checkout main and pull. Wait for explicit instructions.
+
 ## CRITICAL: Personal Data Protection
 
 **⚠️ NEVER include user's personal data in code, comments, or documentation.**

moneyflow/state.py

@@ -788,15 +788,15 @@ def get_filtered_df(self) -> Optional[pl.DataFrame]:
                 if self.selected_time_day is not None:
                     df = df.filter(pl.col("date").dt.day() == self.selected_time_day)
 
-        # Apply view-specific filters
+        # Apply view-specific filters (can have multiple levels in multi-level drill-down)
         if self.view_mode == ViewMode.DETAIL:
             if self.selected_merchant:
                 df = df.filter(pl.col("merchant") == self.selected_merchant)
-            elif self.selected_category:
+            if self.selected_category:
                 df = df.filter(pl.col("category") == self.selected_category)
-            elif self.selected_group:
+            if self.selected_group:
                 df = df.filter(pl.col("group") == self.selected_group)
-            elif self.selected_account:
+            if self.selected_account:
                 df = df.filter(pl.col("account") == self.selected_account)
 
         return df

tests/test_state.py

@@ -551,6 +551,100 @@ def test_get_filtered_df_combined_filters(self, app_state):
         assert all("Starbucks" in row["merchant"] for row in filtered.iter_rows(named=True))
         assert all(row["group"] != "Transfers" for row in filtered.iter_rows(named=True))
 
+    def test_get_filtered_df_multi_level_drill_down(self, app_state):
+        """Test multi-level drill-down filters all dimensions correctly.
+
+        Regression test: ensures stats (transaction count, in/out totals) are
+        calculated correctly for multi-level drill-downs like "Amazon > Groceries".
+        Previously, only the first filter was applied due to elif chain.
+        """
+        data = [
+            # Amazon Groceries transactions (should be included)
+            {
+                "id": "txn_1",
+                "date": date(2024, 1, 1),
+                "amount": -50.00,
+                "merchant": "Amazon",
+                "merchant_id": "merch_1",
+                "category": "Groceries",
+                "category_id": "cat_1",
+                "group": "Food & Dining",
+                "account": "Checking",
+                "account_id": "acc_1",
+                "notes": "",
+                "hideFromReports": False,
+                "pending": False,
+                "is_recurring": False,
+            },
+            {
+                "id": "txn_2",
+                "date": date(2024, 1, 15),
+                "amount": -75.00,
+                "merchant": "Amazon",
+                "merchant_id": "merch_1",
+                "category": "Groceries",
+                "category_id": "cat_1",
+                "group": "Food & Dining",
+                "account": "Checking",
+                "account_id": "acc_1",
+                "notes": "",
+                "hideFromReports": False,
+                "pending": False,
+                "is_recurring": False,
+            },
+            # Amazon Electronics (excluded - wrong category)
+            {
+                "id": "txn_3",
+                "date": date(2024, 1, 10),
+                "amount": -200.00,
+                "merchant": "Amazon",
+                "merchant_id": "merch_1",
+                "category": "Electronics",
+                "category_id": "cat_2",
+                "group": "Shopping",
+                "account": "Checking",
+                "account_id": "acc_1",
+                "notes": "",
+                "hideFromReports": False,
+                "pending": False,
+                "is_recurring": False,
+            },
+            # Target Groceries (excluded - wrong merchant)
+            {
+                "id": "txn_4",
+                "date": date(2024, 1, 20),
+                "amount": -100.00,
+                "merchant": "Target",
+                "merchant_id": "merch_2",
+                "category": "Groceries",
+                "category_id": "cat_1",
+                "group": "Food & Dining",
+                "account": "Checking",
+                "account_id": "acc_1",
+                "notes": "",
+                "hideFromReports": False,
+                "pending": False,
+                "is_recurring": False,
+            },
+        ]
+        app_state.transactions_df = pl.DataFrame(data)
+
+        # Simulate multi-level drill-down: Amazon > Groceries
+        app_state.view_mode = ViewMode.DETAIL
+        app_state.selected_merchant = "Amazon"
+        app_state.selected_category = "Groceries"
+
+        filtered = app_state.get_filtered_df()
+
+        # Should only get Amazon + Groceries transactions (2 out of 4)
+        assert len(filtered) == 2
+        assert all(row["merchant"] == "Amazon" for row in filtered.iter_rows(named=True))
+        assert all(row["category"] == "Groceries" for row in filtered.iter_rows(named=True))
+
+        # Verify the total matches expected (stats calculation uses this)
+        total = float(filtered["amount"].sum())
+        assert total == -125.00  # -50 + -75
+
 
 class TestNavigation:
     """Test navigation and drill-down functionality."""