Update tracing and deployment docs

Author: YoanSallami

docs/Deployment/Building a REST API.md

@@ -54,11 +54,21 @@ from fastapi import FastAPI
 
 import synalinks
 
+# Import open-telemetry dependencies
+from arize.otel import register
+from openinference.instrumentation.litellm import LiteLLMInstrumentor
+
+# Load the environment variables
 load_dotenv()
 
-mlflow.litellm.autolog()
-# Set this to your MLflow server
-mlflow.set_tracking_uri(os.getenv("MLFLOW_URL"))
+# Setup OTel via Arize Phoenix convenience function
+tracer_provider = register(
+    space_id = os.environ["ARIZE_SPACE_ID"], # in app space settings page
+    api_key = os.environ["ARIZE_API_KEY"], # in app space settings page
+    project_name = os.environ["ARIZE_PROJECT_NAME"], # name this to whatever you would like
+)
+
+LiteLLMInstrumentor().instrument(tracer_provider=tracer_provider)
 
 # Set up logging
 logger = logging.getLogger(__name__)
@@ -125,10 +135,11 @@ And finally your docker compose file.
 
 ```yml title="docker-compose.yml"
 services:
-  mlflow:
-    image: ghcr.io/mlflow/mlflow:latest
+  arizephoenix:
+    image: arizephoenix/phoenix:latest
     ports:
-      - "5000:5000"
+      - "6006:6006"
+      - "4317:4317"
   backend:
     build:
       context: ./backend
@@ -138,7 +149,7 @@ services:
     env_file:
       - .env.backend
     depends_on:
-      - mlflow
+      - arizephoenix
 ```
 
 ## Launching your backend

docs/Deployment/Enabling LM Tracing.md

@@ -1,29 +1,42 @@
-# Enabling LM Tracing
+# Enabling LM Tracing with Arize Phoenix
 
 Tracing is important for several reasons, especially in the context of machine learning and software development. It helps identify issues and bugs in your Synalinks programs by providing a detailed log of events and operations. This makes it easier to pinpoint where things went wrong and why.
 
-To install MLflow, an open-source tracing software, use the following:
+In this guide we are going to setup the tracing locally, for more information on how to setup in the cloud, refer to [Arize Phoenix documentation](https://docs.arize.com/phoenix)
 
 ```shell
-pip install mlflow
+uv pip install openinference-instrumentation-litellm arize-otel
 ```
 
 To activate the LM tracing, add the following lines to the top of your script
 
 ```python
-import mlflow
-
-mlflow.litellm.autolog()
-# Set this to your MLflow server
-mlflow.set_tracking_uri("http://localhost:5000")
+# Import open-telemetry dependencies
+from arize.otel import register
+from openinference.instrumentation.litellm import LiteLLMInstrumentor
+
+# Setup OTel via Arize Phoenix convenience function
+tracer_provider = register(
+    space_id = "your-space-id", # in app space settings page
+    api_key = "your-api-key", # in app space settings page
+    project_name = "your-project-name", # name this to whatever you would like
+)
+
+LiteLLMInstrumentor().instrument(tracer_provider=tracer_provider)
 ```
 
 You are done, MLflow is now configured.
 
-To launch MLflow server, use the following command in a shell
+To launch Arize Phoenix server, first pull the docker image with the following command.
+
+```shell
+docker pull arizephoenix/phoenix
+```
+
+Then use the following command in a shell
 
 ```shell
-mlflow server --host 127.0.0.1 --port 5000
+docker run -p 6006:6006 -p 4317:4317 -i -t arizephoenix/phoenix:latest
 ```
 
-You can find more information about the server configuration [here](https://mlflow.org/docs/latest/tracking/server.html).
+Finally go to `http://0.0.0.0:6006` to monitor your application.