Enhancement: Translations via OpenAI's gpt-3.5-turbo API

[jeffturcotte]

@blafferty

We'd like to create a custom translation connector that utilizes OpenAI's gpt-3.5-turbo API in order to facilitate translations.

At a high-level, the implementation would include:

• A new ChatGPTApiClient, closely aligned to the existing Acclaro and Google clients.
• A new ChatGPTTranslationService implementing TranslationServiceInterface that uses the aforementioned GPTApiClient to perform the necessary translations. This service would closely mirror the existing Google service but would utilize a separate client.

OpenAI API Docs: https://platform.openai.com/docs/guides/chat

The GPTApiClient would allow customization via settings of the following:

• API Key. required
• Organization Id: optional
• ChatGPT Model: optional, defaulting to gpt-3.5-turbo
• System Prompt: optional, defaulting to something like: You are an expert language translator. You repeat exactly what is said in {sourceLanguage} in {targetLanguage}. Only return the translated text. Our vision is that this works as a very simple template and sourceLanguage and targetLanguage are injected.

We have a PoC that is already functioning, but needs some refactoring and clean-up.