From 9f24b7d43f60eb29afa5bc31e0c27b2b2693c16b Mon Sep 17 00:00:00 2001
From: Braelyn Boynton <bboynton97@gmail.com>
Date: Tue, 13 Aug 2024 15:23:59 -0700
Subject: [PATCH] remove support for openai v0

---
 agentops/llms/__init__.py  |  46 +-------
 agentops/llms/openai_v0.py | 234 -------------------------------------
 2 files changed, 4 insertions(+), 276 deletions(-)
 delete mode 100644 agentops/llms/openai_v0.py

diff --git a/agentops/llms/__init__.py b/agentops/llms/__init__.py
index 229cd51d..fa7df2c9 100644
--- a/agentops/llms/__init__.py
+++ b/agentops/llms/__init__.py
@@ -43,43 +43,6 @@ def __init__(self, client):
         self.client = client
         self.completion = ""
 
-    # def _override_openai_v0_method(self, api, method_path, module):
-    # def handle_response(result, kwargs, init_timestamp):
-    #     if api == "openai":
-    #         return handle_response_v0_openai(self, result, kwargs, init_timestamp)
-    #     return result
-    #
-    # def wrap_method(original_method):
-    #     if inspect.iscoroutinefunction(original_method):
-    #
-    #         @functools.wraps(original_method)
-    #         async def async_method(*args, **kwargs):
-    #             init_timestamp = get_ISO_time()
-    #             response = await original_method(*args, **kwargs)
-    #             return handle_response(response, kwargs, init_timestamp)
-    #
-    #         return async_method
-    #
-    #     else:
-    #
-    #         @functools.wraps(original_method)
-    #         def sync_method(*args, **kwargs):
-    #             init_timestamp = get_ISO_time()
-    #             response = original_method(*args, **kwargs)
-    #             return handle_response(response, kwargs, init_timestamp)
-    #
-    #         return sync_method
-    #
-    # method_parts = method_path.split(".")
-    # original_method = functools.reduce(getattr, method_parts, module)
-    # new_method = wrap_method(original_method)
-    #
-    # if len(method_parts) == 1:
-    #     setattr(module, method_parts[0], new_method)
-    # else:
-    #     parent = functools.reduce(getattr, method_parts[:-1], module)
-    #     setattr(parent, method_parts[-1], new_method)
-
     def override_api(self):
         """
         Overrides key methods of the specified API to record events.
@@ -112,11 +75,10 @@ def override_api(self):
                             provider = OpenAiProvider(self.client)
                             provider.override()
                         else:
-                            # Patch openai <v1.0.0 methods
-                            for method_path in self.SUPPORTED_APIS["openai"]["0.0.0"]:
-                                self._override_openai_v0_method(
-                                    api, method_path, module
-                                )
+                            raise DeprecationWarning(
+                                "OpenAI versions < 0.1 are no longer supported by AgentOps. Please upgrade OpenAI or "
+                                "downgrade AgentOps."
+                            )
 
                 if api == "cohere":
                     # Patch cohere v5.4.0+ methods
diff --git a/agentops/llms/openai_v0.py b/agentops/llms/openai_v0.py
deleted file mode 100644
index 94fcd448..00000000
--- a/agentops/llms/openai_v0.py
+++ /dev/null
@@ -1,234 +0,0 @@
-import inspect
-import pprint
-from typing import Optional
-
-from agentops.llms.instrumented_provider import InstrumentedProvider
-from agentops.time_travel import fetch_completion_override_from_time_travel_cache
-
-from ..event import ActionEvent, ErrorEvent, LLMEvent
-from ..session import Session
-from ..log_config import logger
-from agentops.helpers import check_call_stack_for_agent_id, get_ISO_time
-from ..client import safe_record
-
-
-class OpenAiInstrumentedProvider(InstrumentedProvider):
-    original_create = None
-    original_create_async = None
-
-    def __init__(self):
-        self._provider_name = "OpenAI"
-
-    def handle_response(
-        self, response, kwargs, init_timestamp, session: Optional[Session] = None
-    ) -> dict:
-        """Handle responses for OpenAI versions <v1.0.0"""
-
-        tracker.llm_event = LLMEvent(init_timestamp=init_timestamp, params=kwargs)
-        if session is not None:
-            tracker.llm_event.session_id = session.session_id
-
-        def handle_stream_chunk(chunk):
-            # NOTE: prompt/completion usage not returned in response when streaming
-            # We take the first ChatCompletionChunk and accumulate the deltas from all subsequent chunks to build one full chat completion
-            if tracker.llm_event.returns == None:
-                tracker.llm_event.returns = chunk
-
-            try:
-                accumulated_delta = tracker.llm_event.returns["choices"][0]["delta"]
-                tracker.llm_event.agent_id = check_call_stack_for_agent_id()
-                tracker.llm_event.model = chunk["model"]
-                tracker.llm_event.prompt = kwargs["messages"]
-                choice = chunk["choices"][
-                    0
-                ]  # NOTE: We assume for completion only choices[0] is relevant
-
-                if choice["delta"].get("content"):
-                    accumulated_delta["content"] += choice["delta"].content
-
-                if choice["delta"].get("role"):
-                    accumulated_delta["role"] = choice["delta"].get("role")
-
-                if choice["finish_reason"]:
-                    # Streaming is done. Record LLMEvent
-                    tracker.llm_event.returns.choices[0]["finish_reason"] = choice[
-                        "finish_reason"
-                    ]
-                    tracker.llm_event.completion = {
-                        "role": accumulated_delta["role"],
-                        "content": accumulated_delta["content"],
-                    }
-                    tracker.llm_event.end_timestamp = get_ISO_time()
-
-                    safe_record(session, tracker.llm_event)
-            except Exception as e:
-                safe_record(
-                    session, ErrorEvent(trigger_event=tracker.llm_event, exception=e)
-                )
-
-                kwargs_str = pprint.pformat(kwargs)
-                chunk = pprint.pformat(chunk)
-                logger.warning(
-                    f"Unable to parse a chunk for LLM call. Skipping upload to AgentOps\n"
-                    f"chunk:\n {chunk}\n"
-                    f"kwargs:\n {kwargs_str}\n"
-                )
-
-        # if the response is a generator, decorate the generator
-        if inspect.isasyncgen(response):
-
-            async def async_generator():
-                async for chunk in response:
-                    handle_stream_chunk(chunk)
-
-                    yield chunk
-
-            return async_generator()
-
-        elif inspect.isgenerator(response):
-
-            def generator():
-                for chunk in response:
-                    handle_stream_chunk(chunk)
-
-                    yield chunk
-
-            return generator()
-
-        # v0.0.0 responses are dicts
-        try:
-            tracker.llm_event.returns = response
-            tracker.llm_event.agent_id = check_call_stack_for_agent_id()
-            tracker.llm_event.prompt = kwargs["messages"]
-            tracker.llm_event.prompt_tokens = response["usage"]["prompt_tokens"]
-            tracker.llm_event.completion = {
-                "role": "assistant",
-                "content": response["choices"][0]["message"]["content"],
-            }
-            tracker.llm_event.completion_tokens = response["usage"]["completion_tokens"]
-            tracker.llm_event.model = response["model"]
-            tracker.llm_event.end_timestamp = get_ISO_time()
-
-            safe_record(session, tracker.llm_event)
-        except Exception as e:
-            safe_record(
-                session, ErrorEvent(trigger_event=tracker.llm_event, exception=e)
-            )
-
-            kwargs_str = pprint.pformat(kwargs)
-            response = pprint.pformat(response)
-            logger.warning(
-                f"Unable to parse response for LLM call. Skipping upload to AgentOps\n"
-                f"response:\n {response}\n"
-                f"kwargs:\n {kwargs_str}\n"
-            )
-
-        return response
-
-    def override(self):
-        pass
-
-
-def handle_response_v0_openai(
-    tracker, response, kwargs, init_timestamp, session: Optional[Session] = None
-):
-    """Handle responses for OpenAI versions <v1.0.0"""
-
-    tracker.llm_event = LLMEvent(init_timestamp=init_timestamp, params=kwargs)
-    if session is not None:
-        tracker.llm_event.session_id = session.session_id
-
-    def handle_stream_chunk(chunk):
-        # NOTE: prompt/completion usage not returned in response when streaming
-        # We take the first ChatCompletionChunk and accumulate the deltas from all subsequent chunks to build one full chat completion
-        if tracker.llm_event.returns == None:
-            tracker.llm_event.returns = chunk
-
-        try:
-            accumulated_delta = tracker.llm_event.returns["choices"][0]["delta"]
-            tracker.llm_event.agent_id = check_call_stack_for_agent_id()
-            tracker.llm_event.model = chunk["model"]
-            tracker.llm_event.prompt = kwargs["messages"]
-            choice = chunk["choices"][
-                0
-            ]  # NOTE: We assume for completion only choices[0] is relevant
-
-            if choice["delta"].get("content"):
-                accumulated_delta["content"] += choice["delta"].content
-
-            if choice["delta"].get("role"):
-                accumulated_delta["role"] = choice["delta"].get("role")
-
-            if choice["finish_reason"]:
-                # Streaming is done. Record LLMEvent
-                tracker.llm_event.returns.choices[0]["finish_reason"] = choice[
-                    "finish_reason"
-                ]
-                tracker.llm_event.completion = {
-                    "role": accumulated_delta["role"],
-                    "content": accumulated_delta["content"],
-                }
-                tracker.llm_event.end_timestamp = get_ISO_time()
-
-                safe_record(session, tracker.llm_event)
-        except Exception as e:
-            safe_record(
-                session, ErrorEvent(trigger_event=tracker.llm_event, exception=e)
-            )
-
-            kwargs_str = pprint.pformat(kwargs)
-            chunk = pprint.pformat(chunk)
-            logger.warning(
-                f"Unable to parse a chunk for LLM call. Skipping upload to AgentOps\n"
-                f"chunk:\n {chunk}\n"
-                f"kwargs:\n {kwargs_str}\n"
-            )
-
-    # if the response is a generator, decorate the generator
-    if inspect.isasyncgen(response):
-
-        async def async_generator():
-            async for chunk in response:
-                handle_stream_chunk(chunk)
-
-                yield chunk
-
-        return async_generator()
-
-    elif inspect.isgenerator(response):
-
-        def generator():
-            for chunk in response:
-                handle_stream_chunk(chunk)
-
-                yield chunk
-
-        return generator()
-
-    # v0.0.0 responses are dicts
-    try:
-        tracker.llm_event.returns = response
-        tracker.llm_event.agent_id = check_call_stack_for_agent_id()
-        tracker.llm_event.prompt = kwargs["messages"]
-        tracker.llm_event.prompt_tokens = response["usage"]["prompt_tokens"]
-        tracker.llm_event.completion = {
-            "role": "assistant",
-            "content": response["choices"][0]["message"]["content"],
-        }
-        tracker.llm_event.completion_tokens = response["usage"]["completion_tokens"]
-        tracker.llm_event.model = response["model"]
-        tracker.llm_event.end_timestamp = get_ISO_time()
-
-        safe_record(session, tracker.llm_event)
-    except Exception as e:
-        safe_record(session, ErrorEvent(trigger_event=tracker.llm_event, exception=e))
-
-        kwargs_str = pprint.pformat(kwargs)
-        response = pprint.pformat(response)
-        logger.warning(
-            f"Unable to parse response for LLM call. Skipping upload to AgentOps\n"
-            f"response:\n {response}\n"
-            f"kwargs:\n {kwargs_str}\n"
-        )
-
-    return response