reduce number of arguments in openai_completion and openai_completion

Several functions passes trough six argument verbatim to OpenAI client creation.
It causes to very long argument lists and increases technological debt.

Pass OpenAI client object instead of those six arguments to reduce
duplicated code and increase repeatability.

TODO: perform the same optimization with generate_data and cli too.

Signed-off-by: Costa Shulyupin <[email protected]>

src/instructlab/sdg/generate_data.py

@@ -22,8 +22,10 @@
     num_chars_from_tokens,
 )
 from jinja2 import Template
+from openai import OpenAI
 from rouge_score import rouge_scorer
 import click
+import httpx
 import instructlab.utils
 import tqdm
 
@@ -266,18 +268,13 @@ def get_instructions_from_model(
     request_idx,
     instruction_data_pool,
     prompt_template,
-    api_base,
-    api_key,
+    client,
     model_name,
     num_prompt_instructions,
     request_batch_size,
     temperature,
     top_p,
     output_file_discarded,
-    tls_insecure,
-    tls_client_cert,
-    tls_client_key,
-    tls_client_passwd,
 ):
     batch_inputs = []
     for _ in range(request_batch_size):
@@ -306,14 +303,9 @@ def get_instructions_from_model(
     request_start = time.time()
     try:
         results = utils.openai_completion(
-            api_base=api_base,
-            api_key=api_key,
+            client,
             prompts=batch_inputs,
             model_name=model_name,
-            tls_insecure=tls_insecure,
-            tls_client_cert=tls_client_cert,
-            tls_client_key=tls_client_key,
-            tls_client_passwd=tls_client_passwd,
             batch_size=request_batch_size,
             decoding_args=decoding_args,
         )
@@ -473,6 +465,15 @@ def generate_data(
     tls_client_key: Optional[str] = None,
     tls_client_passwd: Optional[str] = None,
 ):
+    cert = tuple(
+        item for item in (tls_client_cert, tls_client_key, tls_client_passwd) if item
+    )
+    client = OpenAI(
+        base_url=api_base,
+        api_key=api_key,
+        http_client=httpx.Client(cert=cert, verify=not tls_insecure),
+    )
+
     seed_instruction_data = []
     machine_seed_instruction_data = []
     generate_start = time.time()
@@ -566,18 +567,13 @@ def generate_data(
             request_idx,
             instruction_data_pool,
             prompt_template,
-            api_base,
-            api_key,
+            client,
             model_name,
             num_prompt_instructions,
             request_batch_size,
             temperature,
             top_p,
             output_file_discarded,
-            tls_insecure,
-            tls_client_cert,
-            tls_client_key,
-            tls_client_passwd,
         )
         total_discarded += discarded
         total = len(instruction_data)

src/instructlab/sdg/utils.py

@@ -13,10 +13,9 @@
 
 # Third Party
 # instructlab - TODO these need to go away, issue #6
-from instructlab.configuration import DEFAULT_API_KEY, DEFAULT_MODEL_OLD
+from instructlab.configuration import DEFAULT_MODEL_OLD
 from instructlab.utils import get_sysprompt
-from openai import OpenAI, OpenAIError
-import httpx
+from openai import OpenAIError
 
 StrOrOpenAIObject = Union[str, object]
 
@@ -40,19 +39,14 @@ class OpenAIDecodingArguments:
 
 
 def openai_completion(
-    api_base,
-    tls_insecure,
-    tls_client_cert,
-    tls_client_key,
-    tls_client_passwd,
+    client,
     prompts: Union[str, Sequence[str], Sequence[dict[str, str]], dict[str, str]],
     decoding_args: OpenAIDecodingArguments,
     model_name="ggml-merlinite-7b-lab-Q4_K_M",
     batch_size=1,
     max_instances=sys.maxsize,
     max_batches=sys.maxsize,
     return_text=False,
-    api_key=DEFAULT_API_KEY,
     **decoding_kwargs,
 ) -> Union[
     Union[StrOrOpenAIObject],
@@ -62,11 +56,6 @@ def openai_completion(
     """Decode with OpenAI API.
 
     Args:
-        api_base: Endpoint URL where model is hosted
-        tls_insecure: Disable TLS verification
-        tls_client_cert: Path to the TLS client certificate to use
-        tls_client_key: Path to the TLS client key to use
-        tls_client_passwd: TLS client certificate password
         prompts: A string or a list of strings to complete. If it is a chat model the strings
             should be formatted as explained here:
             https://github.com/openai/openai-python/blob/main/chatml.md.
@@ -78,7 +67,6 @@ def openai_completion(
         max_instances: Maximum number of prompts to decode.
         max_batches: Maximum number of batches to decode. This will be deprecated in the future.
         return_text: If True, return text instead of full completion object (e.g. includes logprob).
-        api_key: API key API key for API endpoint where model is hosted
         decoding_kwargs: Extra decoding arguments. Pass in `best_of` and `logit_bias` if needed.
 
     Returns:
@@ -116,22 +104,8 @@ def openai_completion(
             **decoding_kwargs,
         }
 
-        if not api_key:
-            # we need to explicitly set non-empty api-key, to ensure generate
-            # connects to our local server
-            api_key = "no_api_key"
-
         # do not pass a lower timeout to this client since generating a dataset takes some time
         # pylint: disable=R0801
-        orig_cert = (tls_client_cert, tls_client_key, tls_client_passwd)
-        cert = tuple(item for item in orig_cert if item)
-        verify = not tls_insecure
-        client = OpenAI(
-            base_url=api_base,
-            api_key=api_key,
-            http_client=httpx.Client(cert=cert, verify=verify),
-        )
-
         # ensure the model specified exists on the server. with backends like vllm, this is crucial.
         model_list = client.models.list().data
         model_ids = []