[FEAT] Support for stream_completion in LLM Wrapper (#86)

* Exception handling for Prompt Service

* Supporting stream completion in LLM wrapper

* Update src/unstract/sdk/__init__.py

Signed-off-by: Chandrasekharan M <[email protected]>

---------

Signed-off-by: Chandrasekharan M <[email protected]>
Co-authored-by: Chandrasekharan M <[email protected]>

Author: harini-venkataraman

src/unstract/sdk/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.42.0"
+__version__ = "0.43.0"
 
 
 def get_sdk_version():

src/unstract/sdk/llm.py

@@ -2,6 +2,7 @@
 import re
 from typing import Any, Optional
 
+from llama_index.core.base.llms.types import CompletionResponseGen
 from llama_index.core.llms import LLM as LlamaIndexLLM
 from llama_index.core.llms import CompletionResponse
 from openai import APIError as OpenAIAPIError
@@ -80,6 +81,19 @@ def complete(
         except Exception as e:
             raise parse_llm_err(e) from e
 
+    def stream_complete(
+        self,
+        prompt: str,
+        **kwargs: Any,
+    ) -> CompletionResponseGen:
+        try:
+            response: CompletionResponseGen = self._llm_instance.stream_complete(
+                prompt, **kwargs
+            )
+            return response
+        except Exception as e:
+            raise parse_llm_err(e) from e
+
     def _get_llm(self, adapter_instance_id: str) -> LlamaIndexLLM:
         """Returns the LLM object for the tool.