adds some tests. Improve README

Author: epinault

README.md

@@ -240,6 +240,16 @@ See the [`LangChain.ChatModels.ChatBumblebee` documentation](https://hexdocs.pm/
 
 ## Testing
 
+Before you can run the tests, make sure you have the environment variables set.
+
+You can do this by running:
+
+```
+source .envrc_template
+```
+
+Or you can copy it to `.envrc` and populate it with your private API values.
+
 To run all the tests including the ones that perform live calls against the OpenAI API, use the following command:
 
 ```

lib/chat_models/chat_bumblebee.ex

@@ -484,11 +484,11 @@ defmodule LangChain.ChatModels.ChatBumblebee do
 
     # Track non-streaming response completion
     LangChain.Telemetry.emit_event(
-      [:langchain, :llm, :response, streaming: false],
+      [:langchain, :llm, :response, :non_streaming],
       %{system_time: System.system_time()},
       %{
         model: inspect(model.serving),
-        response_size: byte_size(content)
+        response_size: byte_size(inspect(content))
       }
     )
 

lib/chat_models/chat_mistral_ai.ex

@@ -341,14 +341,14 @@ defmodule LangChain.ChatModels.ChatMistralAI do
           result ->
             # Track non-streaming response completion
             LangChain.Telemetry.emit_event(
-              [:langchain, :llm, :response, streaming: false],
+              [:langchain, :llm, :response, :non_streaming],
               %{system_time: System.system_time()},
               %{
                 model: mistralai.model,
                 response_size: byte_size(inspect(result))
               }
             )
-            
+
             Callbacks.fire(mistralai.callbacks, :on_llm_new_message, [result])
             result
         end
@@ -432,10 +432,7 @@ defmodule LangChain.ChatModels.ChatMistralAI do
   def do_process_response(model, %{"choices" => [], "usage" => %{} = _usage} = data) do
     case get_token_usage(data) do
       %TokenUsage{} = token_usage ->
-        Callbacks.fire(model.callbacks, :on_llm_token_usage, [
-          get_token_usage(data)
-        ])
-
+        Callbacks.fire(model.callbacks, :on_llm_token_usage, [token_usage])
         :ok
 
       nil ->

lib/chat_models/chat_ollama_ai.ex

@@ -433,7 +433,7 @@ defmodule LangChain.ChatModels.ChatOllamaAI do
           result ->
             # Track non-streaming response completion
             LangChain.Telemetry.emit_event(
-              [:langchain, :llm, :response, streaming: false],
+              [:langchain, :llm, :response, :non_streaming],
               %{system_time: System.system_time()},
               %{
                 model: ollama_ai.model,

lib/chat_models/chat_open_ai.ex

@@ -705,7 +705,7 @@ defmodule LangChain.ChatModels.ChatOpenAI do
 
             # Track non-streaming response completion
             LangChain.Telemetry.emit_event(
-              [:langchain, :llm, :response, streaming: false],
+              [:langchain, :llm, :response, :non_streaming],
               %{system_time: System.system_time()},
               %{
                 model: openai.model,

lib/chat_models/chat_perplexity.ex

@@ -374,17 +374,17 @@ defmodule LangChain.ChatModels.ChatPerplexity do
 
           result ->
             Callbacks.fire(perplexity.callbacks, :on_llm_new_message, [result])
-            
+
             # Track non-streaming response completion
             Telemetry.emit_event(
-              [:langchain, :llm, :response, streaming: false],
+              [:langchain, :llm, :response, :non_streaming],
               %{system_time: System.system_time()},
               %{
                 model: perplexity.model,
                 response_size: byte_size(inspect(result))
               }
             )
-            
+
             result
         end
 
@@ -651,14 +651,15 @@ defmodule LangChain.ChatModels.ChatPerplexity do
     end
   end
 
-  def do_process_response(_model, %{"choices" => [
-    %{
-      "delta" => %{"role" => role, "content" => content},
-      "finish_reason" => finish,
-      "index" => index
-    } = _choice
-  ]
-  }) do
+  def do_process_response(_model, %{
+         "choices" => [
+           %{
+             "delta" => %{"role" => role, "content" => content},
+             "finish_reason" => finish,
+             "index" => index
+           } = _choice
+         ]
+       }) do
     status = finish_reason_to_status(finish)
 
     data =
@@ -686,7 +687,8 @@ defmodule LangChain.ChatModels.ChatPerplexity do
               "delta" => %{"content" => content},
               "finish_reason" => finish,
               "index" => index
-            } = _choice
+            }
+            | _
           ]
         }
       ) do

lib/chat_models/chat_vertex_ai.ex

@@ -335,7 +335,7 @@ defmodule LangChain.ChatModels.ChatVertexAI do
 
             # Track non-streaming response completion
             LangChain.Telemetry.emit_event(
-              [:langchain, :llm, :response, streaming: false],
+              [:langchain, :llm, :response, :non_streaming],
               %{system_time: System.system_time()},
               %{
                 model: vertex_ai.model,