Long-period spend analysis (>1 year)

# Instructions — Spend analysis over a long period (> 1 year)

## ⚠️ Strict limit: 366 days maximum

`spendesk_analyze_spend` returns an explicit error if the period exceeds 366 days:
```
"fromDate must be on or before toDate, and the range cannot exceed 366 days."
```

**Solution**: split into sub-periods ≤ 366 days, call separately, combine the results.

## ⚠️ Second gotcha: limit=10 by default

For a groupBy=month over 12 months, the 2 least significant months are aggregated into an "Others" row.
→ Always pass `limit: 12` for a complete annual month-by-month analysis.
→ For 2 years: combine two calls with `limit: 12` each.

## Step 1 — Calculate the sub-periods

Rule: each sub-period ≤ 365 days (leave a 1-day margin).

Examples:
- 2024 analysis: fromDate=2024-01-01, toDate=2024-12-31 ✅ (365 days)
- 2025 analysis: fromDate=2025-01-01, toDate=2025-12-31 ✅ (364 days)
- 2026 YTD analysis: fromDate=2026-01-01, toDate=2026-07-06 ✅ (186 days)

## Step 2 — Call each sub-period (in parallel if possible)

For each sub-period:
```json
{
  "companyId": "...",
  "fromDate": "YYYY-MM-DD",
  "toDate": "YYYY-MM-DD",
  "groupBy": ["month"],
  "limit": 12
}
```

## Step 3 — Combine the results

**Only combine after obtaining all results** (do not present a partial view).

Aggregation:
- `totalAmount`: sum the `summary.totalAmount.amount` of each sub-period (same functional currency)
- `totalCount`: sum the `summary.totalCount`
- Rows by month: concatenate the `data[]` arrays from both calls (no duplicates — distinct periods)

⚠️ If a sub-period contains an "Others" row: the month-by-month breakdown is **partial** for that period. Increase `limit` or narrow the period.

## Step 4 — Year-over-year comparison

To compare 2024 vs 2025:
- Call 1: fromDate=2024-01-01, toDate=2024-12-31
- Call 2: fromDate=2025-01-01, toDate=2025-12-31
- Change = (totalAmount_2025 - totalAmount_2024) / totalAmount_2024 × 100

Alternative: use `compareWithPreviousPeriod=true` on a period ≤ 366 days (automatically compares with the equivalent previous period).

## Live example (real data for this account)

| Period | Total (EUR) | # payables |
|---------|-------------|-------------|
| 2025 (Jan-Dec) | EUR 17,916,500.61 | 6,359 |
| 2026 (Jan-Jun) | EUR 8,174,523.65 | 3,937 |

Note: in the 2025 results, the default `limit=10` → May and Sept. (the 2 least significant) fell into "Others". With `limit=12`, all months are visible.

## Example prompt

- "Analyze our spend over the last 2 years" → 2 calls, year by year
- "Spend report 2024 vs 2025" → 2 calls + comparison
- "Monthly trend since January 2024" → split into 2 periods + limit=12