Expose detailed I/O data via magic variable

@mtauraso is already working on this.

Since Jan 29, 2026.

## Expose detailed I/O data via magic variable

### Implementation

After `%%iops` execution, inject a variable `iops_detailed_data` into the user's namespace via:

```python
self.shell.user_ns['iops_detailed_data'] = detailed_data
```

### Variable Content

**When detailed data is available** (strace/fs_usage modes):
- Type: `pandas.DataFrame`
- Columns:
- `path` (str): File path
- `operation` (str): "read" or "write"
- `syscall` (str, optional): Syscall name if available
- `size_bytes` (int): Bytes transferred

**When detailed data is NOT available** (psutil mode):
- Type: `str`
- Value: `"Detailed I/O data not available: profiling uses psutil mode which only provides aggregate metrics"`

### Documentation

Add a brief note to README.md showing users can access `iops_detailed_data` after running `%%iops`.

### Example

```python
%%iops
with open('test.txt', 'w') as f:
f.write('data')
```

Next cell:
```python
iops_detailed_data # Access the DataFrame
```

Also update readthedocs.