# test-case

**Repository Path**: allmyteacher/test-case

## Basic Information

- **Project Name**: test-case
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-11-20
- **Last Updated**: 2025-11-20

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Agent Interface Tester

This project provides a small test harness that sends predefined human prompts to an agent backend via LangChain and records the responses in an Excel report.

## Setup

1. Install dependencies (preferably in a virtual environment):
   ```
   pip install -r requirements.txt
   ```

2. Prepare environment variables as needed:
   - `AGENT_BASE_URL` (default `http://172.16.1.239:2024`)
   - `AGENT_THREAD_ENDPOINT` (default `/threads`)
   - `AGENT_DATA_ENDPOINT` (default `/threads/{thread_id}/runs/stream`)
   - `AGENT_API_KEY` if the service requires authentication
   - `AGENT_THREAD_ID` to reuse an existing thread id (optional)
   - `AGENT_NUM_THREADS` number of threads for concurrent execution (default: 3)
   - `AGENT_ASSISTANT_ID` assistant identifier passed to the backend (default: `multi_agent`)

## CLI Run (批量命令行模式)

The tester reads prompts from all JSON files in the `test_cases` directory and writes an Excel report (`agent_execution_result.xlsx`) that includes row number, source file name, query payload, returned `Data` value, whether `Data` is present, and latency. The final row summarizes success ratio and timing statistics.

**Test Cases Directory Structure:**
- Place all your test case JSON files in the `test_cases` directory
- Each JSON file should follow the same format:
  ```json
  [
    {
      "messages": [
        {
          "role": "human",
          "content": "Your query here"
        }
      ]
    }
  ]
  ```

```powershell
python agent_interface_tester.py `
  --test-dir test_cases `
  --output agent_execution_result.xlsx `
  --base-url http://172.16.1.239:2024 `
  --thread-endpoint /threads `
  --data-endpoint /threads/{thread_id}/runs/stream `
  --num-threads 3
```

All command-line options have environment-variable defaults, so in the simplest case you can just run `python agent_interface_tester.py` (it will look for the `test_cases` directory by default).

## Web Run (可视化网页模式)

除了命令行模式，还提供了一个 Flask + H5 的网页模式，便于给领导/业务演示：

1. 启动 Web Server：
   ```powershell
   $env:WEB_HOST="127.0.0.1"
   $env:WEB_PORT="5050"
   python web_server.py
   ```

   默认监听 `http://127.0.0.1:5050/`（如需修改，可设置环境变量 `WEB_HOST` / `WEB_PORT`）。

2. 在浏览器打开：
   - 单条查询模式：输入自然语言 `content`，点击“执行查询”，流式输出会实时滚动显示。
   - 批量用例模式：选择本地 `test_cases` 文件夹（或其他包含 JSON 的文件夹），点击“开始批量执行”：
     - 网页上会显示任务进度条和动态日志区域（来源文件名、已完成条数等）。
     - 执行完成后，可在页面上点击按钮下载生成的 Excel 结果文件。
   - 顶部配置栏可随时调整 Base URL，并展示当前 `assistant_id` 以便确认调用对象。

## Features

- **Batch processing**: Process multiple JSON files from a directory automatically
- **Source file tracking**: Excel report includes source file name for each query
- **Multi-threaded execution (CLI)**: Run multiple queries concurrently (default: 3 threads)
- **Streaming response handling**: Real-time display of streaming responses (CLI & Web)
- **Web UI for demos**: Modern dark-themed page with tabs, progress bar, and downloadable results
- **Configurable endpoints**: Web UI exposes Base URL + assistant_id for quick switching
- **Progress tracking**: Visual progress bar showing test execution status with current file
- **Excel reporting**: Comprehensive results with statistics summary