Getting OperationFailure error when trying to find using the $near operator

# Bug

I'm getting a `pymongo.errors.OperationFailure: $geoNear, $near, and $nearSphere are not allowed in this context` error when trying to find using the [$near](https://www.mongodb.com/docs/manual/reference/operator/query/near/#mongodb-query-op.-near) operator.

### Current Behavior

If I have defined the following classes:

```python
class GeoJSON(BaseModel):
type: str = "Point"
coordinates: tuple[float, float]

class Instruction(Model):
name: str
geo_json: GeoJSON
```

And execute this query:
```python
engine.find(Instruction, {
"geo_json": {
"$near": {
"$geometry": {
"type": "Point",
"coordinates": [
-65.49250524795319,
-24.732399067642003
]
},
"$maxDistance": 500
}
}
})
```

The following exception is raised:
```python
pymongo.errors.OperationFailure: $geoNear, $near, and $nearSphere are not allowed in this context, full error: {'ok': 0.0, 'errmsg': '$geoNear, $near, and $nearSphere are not allowed in this context', 'code': 2, 'codeName': 'BadValue', '$clusterTime': {'clusterTime': Timestamp(1676945612, 1), 'signature': {'hash': b'\x8b2\xd80ipA\x82,R\x1bN\xa7a\xad\x7f\xcb\xaeI\x84', 'keyId': 7155114180230512642}}, 'operationTime': Timestamp(1676945612, 1)}
```

### Expected behavior

The query should return a list of Instructions.

### Environment

- ODMantic version: `0.9.2`
- MongoDB version: `5.0.14 (Mongo Atlas)`
- Pydantic infos (output of `python -c "import pydantic.utils; print(pydantic.utils.version_info())`):

```
pydantic version: 1.10.5
pydantic compiled: True
install path: /usr/local/lib/python3.11/site-packages/pydantic
python version: 3.11.1 (main, Dec 23 2022, 09:39:26) [Clang 14.0.0 (clang-1400.0.29.202)]
platform: macOS-13.1-x86_64-i386-64bit
optional deps. installed: ['dotenv', 'email-validator', 'typing-extensions']

```

- Version of additional modules (if relevant):
- motor: `3.1.1`
- Python version: `Docker python:3.10-buster`

**Additional context**

In my DB i have a collection "instruction" with objects like:
```json
{
"_id": { "$oid": "63f4162924a0a5952a0904ef" },
"name": "Some name",
"geo_json":
{
"type": "Point",
"coordinates":
[
{ "$numberDouble": "-65.49250524795319" },
{ "$numberDouble": "-24.732399067642003" }
]
}
}
```
And I've added the index:
```json
{ "geo_json": "2dsphere" }
```

When I try to use the Find on Mongo Atlas, if I add this on the search field:
```json
{
"geo_json": {
"$near": {
"$geometry": {
"type": "Point",
"coordinates": [
-65.49250524795319,
-24.732399067642003
]
},
"$maxDistance": 500
}
}
}
```
It returns the object correctly.

Also, if I change the python code to this:

```python
collection = engine.get_collection(Instruction)
response = collection.find({
"geo_json": {
"$near": {
"$geometry": {
"type": "Point",
"coordinates": [
-65.49250524795319,
-24.732399067642003
]
},
"$maxDistance": 500
}
}
})
instructions = await response.to_list(length=None)
return [Instruction.parse_doc(instruction) for instruction in instructions]
```

Then the search works.

Is the find functionality not working fine or I'm just using it wrong?

Reproduce the query through engine.find(Instruction, ...) and compare it with engine.get_collection(Instruction).find(...), which the report says succeeds. Trace how the engine path handles the $near filter and verify that it preserves the MongoDB query and returns Instruction objects; done when the reported query works without bypassing the engine.

Written by the indexing model from the issue text.