Week granularity dates are incorrect with MS SQL

**Describe the bug**
When you query a cube with a time dimension and specify the `week` granularity, the dates and data for each date are incorrect. They seem to be off by 1.

**To Reproduce**
1. Set up an MS SQL table with a date column
2. Insert 1 row per date from 2023-12-30 to 2024-01-10:

1 2023-12-30 00:00:00.0000000
2 2023-12-31 00:00:00.0000000
3 2024-01-01 00:00:00.0000000
4 2024-01-02 00:00:00.0000000
5 2024-01-03 00:00:00.0000000
6 2024-01-04 00:00:00.0000000
7 2024-01-05 00:00:00.0000000
8 2024-01-06 00:00:00.0000000
9 2024-01-07 00:00:00.0000000
10 2024-01-08 00:00:00.0000000
11 2024-01-09 00:00:00.0000000
12 2024-01-10 00:00:00.0000000

3. Set up a cube that uses this table with a time dimension and a count measure.
4. Query the cube with week granularity and observe results

**Expected behavior**
Results should show:
- Week starting Sunday 24th December 2023 has 1 value
- Week starting Sunday 31st December 2023 has 7 values
- Week starting Sunday 7th January 2024 has 4 values

But unfortunately results show:
- Week starting Monday 25th December 2023 has 1 value
- Week starting Monday 1st January 2024 has 7 values
- Week starting Monday 8th January 2024 has 4 values

Suppling a filter for the date range compounds the issue.

**Version:**
[0.34.54]

**Additional context**
Seems to be an issue with how the week is calculated with this SQL fragment:

```
dateadd(
week,
DATEDIFF(
week,
0,
TODATETIMEOFFSET(
"date",
'+00:00'
)
),
0
)
```

I am unsure what the fix would be though.