From 0afebb95d13b2c64097a29900ce06eb185648e97 Mon Sep 17 00:00:00 2001
From: Russell Jimmies <wlemire.wl@gmail.com>
Date: Fri, 10 Nov 2023 19:43:27 +0000
Subject: [PATCH] Issue #30: model_metadata.py to retrieve model metadata from
 deployed endpoints

---
 .env.template     |  5 +++-
 model_metadata.py | 62 +++++++++++++++++++++++++++++++++++++++++++++++
 requirements.txt  |  4 ++-
 3 files changed, 69 insertions(+), 2 deletions(-)
 create mode 100644 model_metadata.py

diff --git a/.env.template b/.env.template
index 9b4bc3f9..e59a8d43 100644
--- a/.env.template
+++ b/.env.template
@@ -1,4 +1,7 @@
 NACHET_AZURE_STORAGE_CONNECTION_STRING=
 NACHET_MODEL_ENDPOINT_REST_URL=
 NACHET_MODEL_ENDPOINT_ACCESS_KEY=
-NACHET_DATA=
\ No newline at end of file
+NACHET_DATA=
+NACHET_SUBSCRIPTION_ID=
+NACHET_RESOURCE_GROUP=
+NACHET_WORKSPACE=
diff --git a/model_metadata.py b/model_metadata.py
new file mode 100644
index 00000000..6387aafb
--- /dev/null
+++ b/model_metadata.py
@@ -0,0 +1,62 @@
+import os
+import re
+import json
+import yaml
+from dotenv import load_dotenv
+from azure.ai.ml import MLClient, Input
+from azure.ai.ml.entities import Model
+from azure.ai.ml.constants import AssetTypes
+from azure.identity import DefaultAzureCredential
+
+load_dotenv()
+
+NACHET_SUBSCRIPTION_ID = os.getenv("NACHET_SUBSCRIPTION_ID")
+NACHET_RESOURCE_GROUP = os.getenv("NACHET_RESOURCE_GROUP")
+NACHET_WORKSPACE = os.getenv("NACHET_WORKSPACE")
+
+def generate_model_metadata():
+    """
+    Retrieves deployed online_endpoints and generates metadata json 
+    """
+
+    model_metadata = []
+    model_json = {}
+
+    ml_client = MLClient(DefaultAzureCredential(), NACHET_SUBSCRIPTION_ID, NACHET_RESOURCE_GROUP, NACHET_WORKSPACE)
+
+    # Retrieve all endpoints containing "nachet"
+    endpoints = ml_client.online_endpoints.list()
+    nachet_endpoints = [endpoint for endpoint in endpoints if 'nachet' in endpoint.name.lower()]
+
+    ep = nachet_endpoints[0]
+
+    # Retrieve online_deployment
+    deployment = ml_client.online_deployments.get(endpoint_name=ep.name, name=list(ep.traffic.keys())[0])
+
+    # Retrieve deployment's model (from filePath)
+    model_filepath = deployment.model
+    pattern = re.compile(r"models/([^/]+)/versions/(\d+)")
+    match = pattern.search(model_filepath)
+    if match:
+        model_name = match.group(1)  
+        model_version = match.group(2)  
+    else:
+        print("No match found")
+
+    # Retrieve the job object from model
+    model = ml_client.models.get(name=model_name, version=model_version)
+    job = ml_client.jobs.get(name=model.job_name)
+
+    # Because json.dumps(job) returns an empty json, use the dump() method from the Job class to write contents of job object to YAML file
+    job.dump(dest='output.yaml')
+
+    # Read the YAML file and convert it to a dictionary
+    with open('output.yaml', 'r') as file:
+        job_data = yaml.safe_load(file)
+
+    return job_data
+        
+
+if __name__ == "__main__":
+    job_json = generate_model_metadata()
+    print(job_json["component"]) 
\ No newline at end of file
diff --git a/requirements.txt b/requirements.txt
index 584dc092..db3853bb 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -4,4 +4,6 @@ azure-identity
 quart
 quart-cors
 python-dotenv
-hypercorn
\ No newline at end of file
+hypercorn
+azure-ai-ml 
+azure-identity