11# TrustyAI Service  
22
3- 👋 The TrustyAI Service is intended to a hub for all kinds of Responsible AI workflows, such as  
3+ 👋 The TrustyAI Service is intended to be  a hub for all kinds of Responsible AI workflows, such as
44explainability, drift, and Large Language Model (LLM) evaluation. Designed as a REST server wrapping
55a core Python library, the TrustyAI service is intended to be a tool that can operate in a local
66environment, a Jupyter Notebook, or in Kubernetes.
@@ -15,7 +15,7 @@ environment, a Jupyter Notebook, or in Kubernetes.
1515-  Meanshift
1616
1717### ⚖️ Fairness ⚖️  
18- -  Statistical Parity Difference  
18+ -  Statistical Parity Difference
1919-  Disparate Impact Ratio
2020-  Average Odds Ratio (WIP)
2121-  Average Predictive Value Difference (WIP)
@@ -52,6 +52,11 @@ podman build -t $IMAGE_NAME .
5252podman build -t $IMAGE_NAME  --build-arg EXTRAS=eval .  
5353` ` ` ` 
5454
55+ # ## Locally (With ModelMesh/Protobuf Support)
56+ ` ` ` 
57+ uv pip install .[protobuf]
58+ ````
59+ 
5560
5661# # 🏃Running 🏃♀️
5762# ## Locally
@@ -64,7 +69,50 @@ uv run uvicorn src.main --host 0.0.0.0 --port 8080
6469podman run -t $IMAGE_NAME  -p 8080:8080 .  
6570` ` ` 
6671
72+ # # 🧪 Testing 🧪
73+ # ## Running All Tests
74+ To run all tests in  the project:
75+ ` ` ` bash
76+ python -m pytest 
77+ ` ` ` 
78+ 
79+ Or with more verbose output:
80+ ` ` ` bash
81+ python -m pytest -v 
82+ ` ` ` 
83+ 
84+ # ## Running with Coverage
85+ To run tests with coverage reporting:
86+ ` ` ` bash
87+ python -m pytest --cov=src 
88+ ` ` ` 
89+ 
90+ ---
91+ # # 🔄 Protobuf Support 🔄
92+ To process model inference data from ModelMesh models, you can install protobuf support. Otherwise, only KServe models will be supported.
93+ 
94+ # ## Installing Dependencies
95+ Install the required dependencies for  protobuf support:
96+ ` ` ` bash
97+ uv pip install -e " .[protobuf]"  
98+ ` ` ` 
99+ 
100+ # ## Generating Protobuf Code
101+ After installing dependencies, generate Python code from the protobuf definitions:
102+ 
103+ ` ` ` bash
104+ #  From the project root
105+ bash scripts/generate_protos.sh 
106+ ` ` ` 
107+ 
108+ # ## Testing Protobuf Functionality
109+ Run the tests for  the protobuf implementation:
110+ 
111+ ` ` ` bash
112+ #  From the project root
113+ python -m pytest tests/service/data/test_modelmesh_parser.py -v 
114+ ` ` ` 
115+ 
67116---
68117# # ☎️ API ☎️
69118When the service is running, visit ` localhost:8080/docs` ! 
70- 
0 commit comments