File size: 2,816 Bytes
74df718
adea8c3
 
 
 
 
 
4b66647
adea8c3
 
 
 
 
 
 
 
 
 
 
 
 
4b66647
 
adea8c3
 
 
 
 
 
 
 
 
4b66647
adea8c3
 
 
 
 
 
 
 
 
 
4b66647
adea8c3
 
74df718
adea8c3
4b66647
adea8c3
 
 
 
4b66647
adea8c3
 
 
 
 
 
 
 
 
 
 
 
74df718
4b66647
adea8c3
 
 
 
 
 
 
 
74df718
adea8c3
 
 
 
 
 
 
 
 
 
 
4b66647
 
 
 
74df718
4b66647
 
 
 
adea8c3
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
#  Getting Started with CodeLens.

Welcome to **CodeLens.**, a production-grade AI agent evaluation environment. This guide will help you get up and running in less than 2 minutes.

---

## 1. Setup your Environment

First, create a virtual environment and install the required Python dependencies.

```bash
# Create and activate virtual environment
python3 -m venv venv && source venv/bin/activate

# Install dependencies
pip install -r requirements.txt
```

---

## 2. Initialize the Database

CodeLens uses SQLite for persistent episode and leaderboard data. You must initialize the database before running the server for the first time.

```bash
# Initialize the codelens.db with 30 baseline scenarios
python scripts/migrate.py init
```

---

## 3. Launch the System

Start the FastAPI server. This serves both the **Agent API** and the **Interactive Dashboard**.

```bash
# Run the server
PYTHONPATH=. python app.py
```

---

## 4. Open the Dashboard

Once the server is running, you can access the CodeLens Dashboard at:

 **[http://localhost:7860/dashboard](http://localhost:7860/dashboard)**

From here, you can see the top-10 leaderboard and monitor real-time agent evaluations via the live event feed.

---

## 5. Run your First Evaluation

While keeping the server running in one terminal, open a **new terminal** and run the built-in Keyword agent to see results populated on the dashboard.

```bash
# Activate venv in the new terminal first!
source venv/bin/activate

# Run evaluation
python scripts/evaluate.py --agent keyword
```

---

##  Running Tests

To verify everything is working perfectly, you can run the full 155-test suite:

```bash
PYTHONPATH=. pytest tests/ -v
```

---

##  Troubleshooting

### 1. `ModuleNotFoundError: No module named 'requests'`
This happens if you haven't activated the virtual environment in your current terminal tab.
- **Fix**: Run `source venv/bin/activate` in every new terminal window.

### 2. `Usage: python3 scripts/migrate.py [init|reset]`
The migration script requires an argument to proceed.
- **Fix**: Run `python scripts/migrate.py init` specifically.

### 3. Logo not appearing in Dashboard
If the logo shows a broken image placeholder:
- **Fix**: Re-run the server with `PYTHONPATH=. python app.py`. The backend has optimized routing to serve the brand iconography from the root.

---

##  Next Steps

- **Add Scenarios**: Learn how to author new code review benchmarks in **[CONTRIBUTING.md](CONTRIBUTING.md)**.
- **Batch Evaluation**: Scale up from single evaluations to full 30-scenario reports using `scripts/evaluate.py`.
- **Docker Deployment**: Deploy a production-ready container with `docker compose up`.

---

> [!TIP]
> If you ever want to reset the database and start fresh with original scenarios, run:
> `python scripts/migrate.py reset`