TruncateTransform.satisfies_order_of raises AttributeError for different widths

### Apache Iceberg version

0.11.0 (latest release)

### Please describe the bug 🐞

### Description

Calling `TruncateTransform.satisfies_order_of` with two valid truncate transforms that have different widths raises an `AttributeError` instead of returning a boolean.

Same-width comparisons work because the method returns early when the transforms are equal.

### Reproduction

```python
from pyiceberg.transforms import TruncateTransform

TruncateTransform(5).satisfies_order_of(TruncateTransform(3))
```

On `main` at commit `48e710d20ceeeaa637d5aeae7746b787410859f8`, this raises:

```text
AttributeError: 'TruncateTransform' object has no attribute '_source_type'
```

The failure reproduces consistently. The same code is also present in the 0.11.1 release.

### Expected behavior

The method should compare the truncate widths and return a boolean:

```python
assert TruncateTransform(5).satisfies_order_of(TruncateTransform(3))
assert not TruncateTransform(3).satisfies_order_of(TruncateTransform(5))
```

This matches the current Apache Iceberg Java implementation:

https://github.com/apache/iceberg/blob/25654ab4b29c8b5b5c20fc427da01cb70d94ed14/api/src/main/java/org/apache/iceberg/transforms/Truncate.java#L130-L141

### Cause

`TruncateTransform.__init__` initializes `_width` but not `_source_type`. However, `satisfies_order_of` still accesses the `source_type` property backed by `_source_type`:

https://github.com/apache/iceberg-python/blob/48e710d20ceeeaa637d5aeae7746b787410859f8/pyiceberg/transforms.py#L783-L905

The existing unit test only compares a transform with itself, so it returns before reaching the failing branch:

https://github.com/apache/iceberg-python/blob/48e710d20ceeeaa637d5aeae7746b787410859f8/tests/test_transforms.py#L501-L512

A focused fix could compare `TruncateTransform` widths directly, consistent with the Java implementation, and add regression cases for different widths.

I would be happy to contribute the fix and regression tests.

### Willingness to contribute

- [x] I can contribute a fix for this bug independently
- [ ] I would be willing to contribute a fix for this bug with guidance from the Iceberg community
- [ ] I cannot contribute a fix for this bug at this time

Start in pyiceberg/transforms.py at TruncateTransform.satisfies_order_of and review the related constructor and source_type property. Run the focused cases in tests/test_transforms.py around lines 501-512, then add coverage for different widths. Done means comparisons return the expected booleans without accessing an uninitialized attribute.

Written by the indexing model from the issue text.