Support dashed arguments for benchmark args (#441)

## Summary

<!--
Include a short paragraph of the changes introduced in this PR.
If this PR requires additional context or rationale, explain why
the changes are necessary.
-->
For all arguments in `BenchmarkGenerativeTextArgs` support dashed
versions of the name during validation to match the CLI. Additionally
alias the argument `data_request_formatter` to `request_type`.


## Details

<!--
Provide a detailed list of all changes introduced in this pull request.
-->
Alias are configured so that if both options are specified, the original
name takes precedence. E.g. the order of precedence is `max_seconds >
max-seconds`.

~~For `data_request_formatter` the request type alias takes precedence.
E.g. `request_type > request-type > data_request_formatter`.~~ Changed
this because it caused precedence issues with CLI arg. Order is now
`data_request_formatter > data-request-formatter > request_type >
request-type`.

## Test Plan

The following scenario file can be used to test. Modify which fields are
uncommented to test precedence.

```yaml
target: <hostname>
#request_type: text_completions
#request-type: text_completions
data_request_formatter: text_completions
profile: concurrent
rate: 2
max_seconds: 30
data:
  prompt_tokens: 256
  output_tokens: 128
```

## Related Issues

<!--
Link any relevant issues that this PR addresses.
-->
- Resolves #

---

- [x] "I certify that all code in this PR is my own, except as noted
below."

## Use of AI

- [ ] Includes AI-assisted code completion
- [ ] Includes code generated by an AI application
- [ ] Includes AI-generated tests (NOTE: AI written tests should have a
docstring that includes `## WRITTEN BY AI ##`)

Author: sjmonson

src/guidellm/benchmark/schemas.py

@@ -24,6 +24,8 @@
 
 import yaml
 from pydantic import (
+    AliasChoices,
+    AliasGenerator,
     ConfigDict,
     Field,
     ValidationError,
@@ -1796,9 +1798,8 @@ def create(
                 scenario_data = scenario_data["args"]
             constructor_kwargs.update(scenario_data)
 
-        for key, value in kwargs.items():
-            if value != cls.get_default(key):
-                constructor_kwargs[key] = value
+        # Apply overrides from kwargs
+        constructor_kwargs.update(kwargs)
 
         return cls.model_validate(constructor_kwargs)
 
@@ -1832,6 +1833,14 @@ def get_default(cls: type[BenchmarkGenerativeTextArgs], field: str) -> Any:
         use_enum_values=True,
         from_attributes=True,
         arbitrary_types_allowed=True,
+        validate_by_alias=True,
+        validate_by_name=True,
+        alias_generator=AliasGenerator(
+            # Support field names with hyphens
+            validation_alias=lambda field_name: AliasChoices(
+                field_name, field_name.replace("_", "-")
+            ),
+        ),
     )
 
     # Required
@@ -1878,6 +1887,12 @@ def get_default(cls: type[BenchmarkGenerativeTextArgs], field: str) -> Any:
     data_request_formatter: DatasetPreprocessor | dict[str, str] | str = Field(
         default="chat_completions",
         description="Request formatting preprocessor or template name",
+        validation_alias=AliasChoices(
+            "data_request_formatter",
+            "data-request-formatter",
+            "request_type",
+            "request-type",
+        ),
     )
     data_collator: Callable | Literal["generative"] | None = Field(
         default="generative", description="Data collator for batch processing"