Document AI v1beta3 update_dataset LRO: "Unexpected state: Long-running operation had neither response nor error set"

### Determine this is the right repository

- [x] I determined this is the correct repository in which to report this bug.

### Summary of the issue

**Context**

I am configuring a dataset for a Custom Classifier processor using the
`google-cloud-documentai` v1beta3 Python client (`DocumentServiceClient.update_dataset`).

The underlying REST `PATCH` call to
`projects/{project}/locations/{location}/processors/{processor}/dataset`
succeeds and the dataset is correctly attached to the processor (visible in
the Console UI). However, the Python long‑running operation helper raises:

`GoogleAPICallError('Unexpected state: Long-running operation had neither response nor error set.')`

**Expected Behavior**

`operation.result()` on `client.update_dataset(...)` should complete without
raising an exception when the backend successfully updates the dataset, and
the returned `Operation` should have either a `response` or an `error` set.

**Actual Behavior**

The dataset is attached correctly (Console shows the dataset GCS URI and the
dataset works for later training), but `operation.result()` raises
`GoogleAPICallError: None Unexpected state: Long-running operation had neither
response nor error set.`

This makes it difficult to distinguish real failures from this client‑side LRO quirk.

### API client name and version

google-cloud-documentai==3.7.0, google-api-core==2.28.1

### Reproduction steps: code

file: setup.py

```python
from google.cloud import documentai_v1beta3 as documentai
from google.api_core.client_options import ClientOptions

# Your settings
PROJECT_ID = "PROJECT_ID"
LOCATION = "LOCATION"
PROCESSOR_ID = "PROCESSOR_ID"
gcs_uri_prefix = "GCS_URI"

# Setup the endpoint correctly
client_options = ClientOptions(api_endpoint=f"{LOCATION}-documentai.googleapis.com")

# Initialize the Document AI client
client = documentai.DocumentServiceAsyncClient(client_options=client_options)

# Build the full dataset resource name
dataset_name = f"projects/{PROJECT_ID}/locations/{LOCATION}/processors/{PROCESSOR_ID}/dataset"

# Prepare the dataset configuration
dataset = documentai.Dataset(
name=dataset_name,
gcs_managed_config=documentai.Dataset.GCSManagedConfig(
gcs_prefix=documentai.GcsPrefix(
gcs_uri_prefix=gcs_uri_prefix
)
),
spanner_indexing_config=documentai.Dataset.SpannerIndexingConfig()
)

# Prepare the update request
update_request = documentai.UpdateDatasetRequest(
dataset=dataset
)

# Call the update_dataset API
operation = client.update_dataset(request=update_request)

response = operation.result()
```

### Reproduction steps: supporting files

### Reproduction steps: actual results

Log output:
```
Traceback:
File "setup.py", line 128, in add_processor_dataset
response = operation.result(timeout=600)
File ".../google/api_core/future/polling.py", line 261, in result
raise self._exception
google.api_core.exceptions.GoogleAPICallError: None Unexpected state:
Long-running operation had neither response nor error set.
```
Despite this, the Console shows the dataset configured correctly under the
processor.

### Reproduction steps: expected results

`operation.result()` should either:

- Return a `Dataset` (or empty response) when the update succeeds, or
- Raise an error with details from the backend (e.g. permission denied, invalid name).

It should not raise `GoogleAPICallError: ... neither response nor error set`
when the dataset is successfully attached and visible in the UI.

### OS & version + platform

Base image: python:3.11-slim (Debian-based) Platform: Cloud Run job (containerized)

### Python environment

Python 3.11.x

### Python dependencies

Package Version Editable project location
------------------------------- ----------- ----------------------------------------
aiohappyeyeballs 2.6.1
aiohttp 3.13.3
aiosignal 1.4.0
annotated-types 0.7.0
async-timeout 5.0.1
asyncpg 0.31.0
attrs 25.4.0
backports-asyncio-runner 1.2.0
bottleneck 1.6.0
certifi 2026.1.4
cffi 2.0.0
charset-normalizer 3.4.4
cryptography 46.0.4
db-dtypes 1.5.0
decorator 5.2.1
deprecated 1.3.1
exceptiongroup 1.3.1
frozenlist 1.8.0
fsspec 2026.2.0
gcsfs 2026.2.0
google-api-core 2.29.0
google-api-python-client 2.189.0
google-auth 2.48.0
google-auth-httplib2 0.3.0
google-auth-oauthlib 1.2.4
google-cloud-bigquery 3.40.0
google-cloud-core 2.5.0
google-cloud-documentai 3.9.0
google-cloud-documentai-toolbox 0.15.0a0
google-cloud-secret-manager 2.26.0
google-cloud-storage 3.9.0
google-cloud-storage-control 1.9.0
google-cloud-vision 3.12.1
google-crc32c 1.8.0
google-resumable-media 2.8.0
googleapis-common-protos 1.72.0
greenlet 3.3.1
grpc-google-iam-v1 0.14.3
grpcio 1.78.0
grpcio-status 1.78.0
httplib2 0.31.2
idna 3.11
immutabledict 4.2.2
iniconfig 2.3.0
intervaltree 3.2.1
jinja2 3.1.6
joblib 1.5.3
llvmlite 0.46.0
lxml 6.0.2
markupsafe 3.0.3
multidict 6.7.1
numba 0.63.1
numexpr 2.14.1
numpy 2.2.6
oauthlib 3.3.1
packaging 26.0
pandas 2.3.3
pandas-gbq 0.33.0
pikepdf 10.3.0
pillow 11.3.0
pluggy 1.6.0
propcache 0.4.1
proto-plus 1.27.1
protobuf 6.33.5
psutil 7.2.2
pyarrow 22.0.0
pyasn1 0.6.2
pyasn1-modules 0.4.2
pycparser 3.0
pydantic 2.12.5
pydantic-core 2.41.5
pydantic-settings 2.12.0
pydata-google-auth 1.9.1
pygments 2.19.2
pyparsing 3.3.2
pytest 9.0.2
pytest-asyncio 1.3.0
python-dateutil 2.9.0.post0
python-dotenv 1.2.1
pytz 2025.2
requests 2.32.5
requests-oauthlib 2.0.0
rsa 4.9.1
scikit-learn 1.7.2
scipy 1.15.3
setuptools 82.0.0
six 1.17.0
sortedcontainers 2.4.0
sqlalchemy 2.0.46
tabulate 0.9.0
threadpoolctl 3.6.0
tifffile 2025.5.10
tomli 2.4.0
tqdm 4.67.3
typing-extensions 4.15.0
typing-inspection 0.4.2
tzdata 2025.3
uritemplate 4.2.0
urllib3 2.6.3
wrapt 2.1.1
yarl 1.22.0

### Additional context

If I call the REST API directly with:
```
curl -X PATCH \
-H "Authorization: Bearer $(gcloud auth print-access-token)" \
-H "Content-Type: application/json; charset=utf-8" \
-d @request.json \
"https://eu-documentai.googleapis.com/v1beta3/projects/PROJECT_ID/locations/LOCATION/processors/PROCESSOR_ID/dataset"
```
and the JSON body:
```
{
"name": "projects/PROJECT_ID/locations/LOCATION/processors/PROCESSOR_ID/dataset",
"gcs_managed_config": {
"gcs_prefix": {
"gcs_uri_prefix": "gs://..."
}
},
"spanner_indexing_config": {}
}
```
the operation succeeds and the dataset is attached, with no error.
Only the Python client’s `operation.result()` reports the invalid LRO state.