Monthly spending trend

# Instructions — Monthly spend trend

## ⚠️ Default sort = descending amount (not chronological)

Without `sortBy`, months are returned from highest to lowest:
```
2026-01 (10.7%) → 2025-12 (10.3%) → 2025-10 (9.4%) → ...
```

For a **chronological trend chart**, always specify:
```json
"sortBy": "name",
"sortOrder": "asc"
```
→ Months are then returned in order `2025-07 → 2025-08 → ... → 2026-06`.

## ⚠️ Use limit=12 for a full year

By default `limit=10` → for a groupBy=month over 12 months, the 2 least significant months fall into "Others". Always pass `limit=12` (or `limit=24` for 2 years).

## ⚠️ Partial last month

If the period includes the current month, the last month will show a lower amount (partial data).
Always state: *"Data for [current month] is partial."*

## Step 1 — Monthly call

```json
{
  "companyId": "...",
  "fromDate": "YYYY-MM-DD",
  "toDate": "YYYY-MM-DD",
  "groupBy": ["month"],
  "sortBy": "name",
  "sortOrder": "asc",
  "limit": 12
}
```

## Step 2 — Interpreting the fields

Each row:
- `key`: `"YYYY-MM"` — month in ISO format (e.g. `"2025-07"`)
- `amount`: total spend for the month (integer coefficient, divide by 10^precision)
- `count`: number of payables in the month
- `share`: % of the period total

## Step 3 — Calculating trend metrics

```
monthly_average = summary.totalAmount / number_of_months
max_month = max(data[].amount)
min_month = min(data[].amount)
month_over_month_change = (month_n.amount - month_n-1.amount) / month_n-1.amount × 100
```

## Step 4 — Detecting anomalies

Alert on:
- Month with very low `count` but high `amount` → a few large invoices
- Month with change > ±30% vs the previous month → spike or dip to explain
- Incomplete current month → mention explicitly

## Live example (real data, Jul 2025 → Jun 2026, sorted by amount — not by date)

| Rank | Month | Amount (EUR) | Count | Share |
|------|------|--------------|-------|-------|
| 1 | 2026-01 | 1,769,755 | 605 | 10.7% |
| 2 | 2025-12 | 1,697,623 | 666 | 10.3% |
| 3 | 2025-10 | 1,554,335 | 774 | 9.4% |
| 4 | 2026-02 | 1,531,954 | 742 | 9.3% |
| 5 | 2026-03 | 1,464,520 | 688 | 8.9% |
| 6 | 2025-08 | 1,457,602 | 339 | 8.8% |
| 7 | 2026-04 | 1,368,217 | 801 | 8.3% |
| 8 | 2025-11 | 1,290,375 | 652 | 7.8% |
| 9 | 2025-07 | 1,200,645 | 387 | 7.3% |
| 10 | 2025-09 | 1,149,472 | 444 | 7.0% |
| 11 | 2026-05 | 1,129,378 | 492 | 6.8% |
| 12 | 2026-06 | 910,700 | 609 | 5.5% |

→ No "Others" row ✅ (limit=12 = 12 months)
→ Total: EUR 16,524,574.27 over 12 months
→ Monthly average: **~EUR 1,377,048/month**
→ June 2026 is the lowest (partial data — month in progress at the time of the test)

With `sortBy=name, sortOrder=asc`: the same data would be returned in the order jul-25, aug-25, ..., jun-26.

## Example prompt

- "Show the trend of our spend over the last 12 months" → groupBy=month, sortBy=name, sortOrder=asc, limit=12
- "Which month did we spend the most?" → groupBy=month, sortBy=amount (default), limit=1
- "Month-by-month evolution since January 2026" → fromDate=2026-01-01, toDate=today, groupBy=month, sortBy=name