[RNE Rewrite] Deduplicate tokenizer result-unwrapping with a local unwrap helper

[msluszniak]

Description

The five TokenizerHostObject methods (encode, decode, idToToken, tokenToId, and the encode path) each hand-rolled the same block around a tokenizers::Result:

auto result = self->tokenizer_->encode(...);
if (!result.ok()) {
    throw jsi::JSError(rt, std::format("encode: Failed to encode input: {}", toString(result.error())));
}
return conversions::toJsiArray(rt, result.get());

Replaced with a local unwrap helper in the file's anonymous namespace, mirroring the one already in core/model.cpp. That existing helper is file-local and typed to executorch::runtime::Result, so it cannot be reused here — the tokenizer returns tokenizers::Result carrying a distinct tokenizers::Error (which is why this file already has its own toString). Following the established per-file convention rather than hoisting a shared helper into core keeps the change scoped and leaves core untouched.

Behaviour is identical: same error messages, same exceptions.

Split out of #1319 so the std::span refactor and this cleanup can be reviewed independently. Part of the broader tokenizer host-object cleanup tracked in #1316.

Introduces a breaking change?

• Yes
• No

Type of change

• Bug fix (change which fixes an issue)
• New feature (change which adds functionality)
• Documentation update (improves or adds clarity to existing documentation)
• Other (chores, tests, code style improvements etc.)

Tested on

• iOS
• Android

Testing instructions

Behaviour-preserving; verification is compile-only, not yet run on a device.

tokenizer.cpp compiles clean under the project's strict warning set from .clangd:

cd packages/react-native-executorch
clang++ -fsyntax-only $(tr '\n' ' ' < compile_flags.txt) \
  -Wall -Wextra -Wpedantic -Wconversion -Wsign-conversion cpp/extensions/nlp/tokenizer.cpp

To exercise at runtime: the tokenizer screen in apps/nlp.

Related issues

Part of #1316

Checklist

• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have updated the documentation accordingly
• My changes generate no new warnings