enhance: add support for azure deployment name mapping

[keyallis]

#43

Add a configurable env variable to set the azure deployment name to be used in lieu of the model name. Migrate away from using the snapshotted fork of go-openai, and add a hack for passing temperature.

[danielhass]

@keyallis thank you for this PR!

I tried use your changes already locally however I ran into the situation that gptscript is waiting for the model response "for ever":

[...]
{
  "completionID": "1",
  "id": "1",
  "level": "debug",
  "logger": "pkg/monitor",
  "msg": "messages",
  "parentID": "",
  "request": {
    "model": "gpt-4-turbo-preview",
    "messages": [
      {
        "role": "system",
        "content": "\nYou are task oriented system.\nYou receive input from a user, process the input from the given instructions, and then output the result.\nYour objective is to provide consistent and correct results.\nYou do not need to explain the steps taken, only provide the result to the given instructions.\nYou are referred to as a tool.\n"
      },
      {
        "role": "system",
        "content": "Say hello world"
      }
    ],
    "temperature": 1e-8
  },
  "time": "2024-02-27T17:02:22+01:00",
  "toolID": "examples/helloworld.gpt:1"
}
         content  [1] content | Waiting for model response...

Have you ran into this to sometimes?

[keyallis]

Hey @danielhass, thanks for taking a look. I'm curious what variables you're setting when testing this? These are the ones I have set:

OPENAI_API_KEY
OPENAI_URL
OPENAI_API_TYPE
OPENAI_AZURE_DEPLOYMENT

If you don't have all of these set then it might not be picking up the changes included

[drpebcak]

Why is this different than OPENAI_BASE_URL?

[drpebcak]

I guess we already have both, but this version does not match the help text

[keyallis]

That's just the env var I was using. I can swap it.

[danielhass]

Hey @danielhass, thanks for taking a look. I'm curious what variables you're setting when testing this? These are the ones I have set:

OPENAI_API_KEY
OPENAI_URL
OPENAI_API_TYPE
OPENAI_AZURE_DEPLOYMENT

If you don't have all of these set then it might not be picking up the changes included

@keyallis I used the following set of env vars:

export OPENAI_API_VERSION="2024-02-15-preview"
export OPENAI_AZURE_DEPLOYMENT="gpt-4"
export OPENAI_API_KEY="xxx"
export OPENAI_BASE_URL="https://xxx.openai.azure.com/"
export OPENAI_API_TYPE="AZURE"

[keyallis]

@danielhass I believe your issue might be with the OPENAI_API_VERSION, you might try either using the default setting for that, or some other version values. Azure has some inconsistent functionality support between api versions from what I could gather.

[danielhass]

@keyallis 💯 - changing to OPENAI_API_VERSION="2023-05-15" resolved the issue!

Shall we add this to the docs for the Azure variant?

[keyallis]

@danielhass I've omitted mention of the OPENAI_API_VERSION in the getting started flow to keep it more general. If a value is not set for that field then it defaults to the value 2023-05-15.

[danielhass]

@danielhass I've omitted mention of the OPENAI_API_VERSION in the getting started flow to keep it more general. If a value is not set for that field then it defaults to the value 2023-05-15.

@keyallis didn't know about the default. Makes total sense not to mention it in that case 👍

Thanks for bringing this to gptscript!