Harrison/banana fix (#1311)

Co-authored-by: Erik Dunteman <44653944+erik-dunteman@users.noreply.github.com>

docs/ecosystem/bananadev.md

@@ -4,8 +4,9 @@ This page covers how to use the Banana ecosystem within LangChain.
 It is broken into two parts: installation and setup, and then references to specific Banana wrappers.
 
 ## Installation and Setup
+
 - Install with `pip3 install banana-dev`
-- Get an CerebriumAI api key and set it as an environment variable (`BANANA_API_KEY`)
+- Get an Banana api key and set it as an environment variable (`BANANA_API_KEY`)
 
 ## Define your Banana Template
 
@@ -15,13 +16,16 @@ You can check out an example Banana repository [here](https://github.com/concept
 
 ## Build the Banana app
 
-You must include a output in the result. There is a rigid response structure.
+Banana Apps must include the "output" key in the return json. 
+There is a rigid response structure.
+
 ```python
 # Return the results as a dictionary
 result = {'output': result}
 ```
 
 An example inference function would be:
+
 ```python
 def inference(model_inputs:dict) -> dict:
     global model
@@ -58,17 +62,18 @@ def inference(model_inputs:dict) -> dict:
 
 You can find a full example of a Banana app [here](https://github.com/conceptofmind/serverless-template-palmyra-base/blob/main/app.py).
 
-
 ## Wrappers
 
 ### LLM
 
 There exists an Banana LLM wrapper, which you can access with
+
 ```python
 from langchain.llms import Banana
 ```
 
 You need to provide a model key located in the dashboard:
+
 ```python
 llm = Banana(model_key="YOUR_MODEL_KEY")
 ```

langchain/llms/bananadev.py

@@ -100,10 +100,15 @@ class Banana(LLM, BaseModel):
         response = banana.run(api_key, model_key, model_inputs)
         try:
             text = response["modelOutputs"][0]["output"]
-        except KeyError:
+        except (KeyError, TypeError):
+            returned = response["modelOutputs"][0]
             raise ValueError(
-                f"Response should be {'modelOutputs': [{'output': 'text'}]}."
-                f"Response was: {response}"
+                "Response should be of schema: {'output': 'text'}."
+                f"\nResponse was: {returned}"
+                "\nTo fix this:"
+                "\n- fork the source repo of the Banana model"
+                "\n- modify app.py to return the above schema"
+                "\n- deploy that as a custom repo"
             )
         if stop is not None:
             # I believe this is required since the stop tokens