# XYParser 数据流与接口时序说明

## 接口时序图

### 3.1 64导初始化连接阶段

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Dev as 64导EEG采集设备
    participant Host as 上位机
    participant Lib as XYParser库

    Dev-->>Host: 设备连接成功
    Host->>Lib: parser64 = XYParser_CreateParser(64)
    Host->>Lib: XYParser_SetAdcParams(parser64, 4.5, 6.0)
    Host->>Lib: XYParser_SetSampleRate(parser64, 250)
    Host->>Lib: XYParser_SetBypassChecksum(parser64, 1)
    Host->>Lib: gain_cmd_size = XYParser_Get64GainSampleRateCommandSize()
    Lib-->>Host: gain_cmd_size
    Host->>Lib: gain_cmd_bytes = XYParser_Serialize64GainSampleRateCommand(6, 250, gain_cmd_buf, gain_cmd_size)
    Lib-->>Host: gain_cmd_bytes
    Host->>Dev: 下发采样率250、增益6命令
```

### 3.2 64导阻抗阶段

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Dev as 64导EEG采集设备
    participant Host as 上位机
    participant Lib as XYParser库

    Host->>Lib: XYParser_SetImpedanceDetection(parser64, 1)
    Host->>Lib: impedance_gain_cmd_size = XYParser_Get64GainSampleRateCommandSize()
    Lib-->>Host: impedance_gain_cmd_size
    Host->>Lib: impedance_gain_cmd_bytes = XYParser_Serialize64GainSampleRateCommand(24, 250, impedance_gain_cmd_buf, impedance_gain_cmd_size)
    Lib-->>Host: impedance_gain_cmd_bytes
    Host->>Dev: 下发采样率250、增益24命令
    Host->>Lib: impedance_cmd_size = XYParser_Get64ImpedanceCommandSize()
    Lib-->>Host: impedance_cmd_size
    Host->>Lib: open_impedance_bytes = XYParser_Serialize64ImpedanceCommand(1, impedance_cmd_buf, impedance_cmd_size)
    Lib-->>Host: open_impedance_bytes
    Host->>Dev: 下发阻抗开启命令

    loop 持续获取阻抗
        Dev-->>Host: 原始EEG字节流
        Host->>Lib: frame_count = XYParser_Feed(parser64, raw_data, raw_size, frame_summaries, max_frames)
        Lib-->>Host: frame_count + frame_summaries
        Host->>Lib: impedance_count = XYParser_ReadImpedance(parser64, impedance_summaries, max_impedance)
        Lib-->>Host: impedance_count + impedance_summaries
    end

    Host->>Lib: close_impedance_bytes = XYParser_Serialize64ImpedanceCommand(0, impedance_cmd_buf, impedance_cmd_size)
    Lib-->>Host: close_impedance_bytes
    Host->>Dev: 下发阻抗关闭命令
    Host->>Lib: restore_gain_cmd_size = XYParser_Get64GainSampleRateCommandSize()
    Lib-->>Host: restore_gain_cmd_size
    Host->>Lib: restore_gain_cmd_bytes = XYParser_Serialize64GainSampleRateCommand(6, 250, restore_gain_cmd_buf, restore_gain_cmd_size)
    Lib-->>Host: restore_gain_cmd_bytes
    Host->>Dev: 下发采样率250、增益6命令
    Host->>Lib: XYParser_SetImpedanceDetection(parser64, 0)
```

### 3.3 64导算法阶段

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Dev as 64导EEG采集设备
    participant Host as 上位机
    participant Lib as XYParser库
    participant Algo as 算法

    Host->>Lib: XYParser_SetWelchDetection(parser64, 1)

    loop 持续采集
        Dev-->>Host: 原始EEG字节流
        Host->>Lib: frame_count = XYParser_Feed(parser64, raw_data, raw_size, frame_summaries, max_frames)
        Lib-->>Host: frame_count + frame_summaries
        Host->>Lib: value_count = XYParser_GetAlgorithmDataValueCount()
        Lib-->>Host: value_count
        Host->>Lib: ok = XYParser_ConvertSampleFramesToAlgorithmData(frame_summary, algorithm_input_data)
        Lib-->>Host: ok + algorithm_input_data
        Host->>Algo: 输入算法数据 algorithm_input_data
        Algo-->>Host: 算法输出数据 algorithm_output_bytes
        Host->>Lib: alg_frame_count = XYParser_FeedAlgorithmData(parser64, algorithm_output_bytes, algorithm_output_size, algorithm_frames, max_algorithm_frames)
        Lib-->>Host: alg_frame_count + algorithm_frames
        Host->>Lib: welch_count = XYParser_ReadWelch(parser64, welch_summaries, max_welch)
        Lib-->>Host: welch_count + welch_summaries
    end

    opt 结束时处理尾数据
        Host->>Lib: flushed = XYParser_FlushAlgorithmData(parser64, tail_frame_summary)
        Lib-->>Host: flushed + tail_frame_summary
        Host->>Lib: welch_count = XYParser_ReadWelch(parser64, welch_summaries, max_welch)
        Lib-->>Host: welch_count + welch_summaries
    end

    Host->>Lib: XYParser_DestroyParser(parser64)
```

### 3.4 64导训练打标阶段

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Port as 64导EEG设备串口
    participant Host as 上位机
    participant Lib as XYParser库
    participant Time as 时间

    Host->>Lib: trigger_cmd_size = XYParser_GetTriggerCommandSize()
    Lib-->>Host: trigger_cmd_size
    Host->>Lib: train0_bytes = XYParser_SerializeTriggerCommand(XYPARSER_TRIGGER_TRAIN_0, trigger_cmd_buf, trigger_cmd_size)
    Lib-->>Host: train0_bytes
    Host->>Port: 发送 TRAIN_0 打标
    Host->>Time: 开始训练计时
    Time-->>Host: 训练时间到
    Host->>Lib: train1_bytes = XYParser_SerializeTriggerCommand(XYPARSER_TRIGGER_TRAIN_1, trigger_cmd_buf, trigger_cmd_size)
    Lib-->>Host: train1_bytes
    Host->>Port: 发送 TRAIN_1 打标
```

### 3.5 8导初始化连接阶段

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Dev as 8导EEG采集设备
    participant Host as 上位机
    participant Lib as XYParser库

    Dev-->>Host: 设备连接成功
    Host->>Lib: parser8 = XYParser_CreateParser(8)
    Host->>Lib: XYParser_SetAdcParams(parser8, vref, gain)
    Host->>Lib: XYParser_SetSampleRate(parser8, sample_rate)
    Host->>Lib: XYParser_SetBypassChecksum(parser8, 1)
```

### 3.6 8导阻抗阶段

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Dev as 8导EEG采集设备
    participant Host as 上位机
    participant Lib as XYParser库

    Host->>Lib: XYParser_SetImpedanceDetection(parser8, 1)
    Host->>Lib: impedance_cmd_size = XYParser_Get8ChImpedanceCommandSize()
    Lib-->>Host: impedance_cmd_size
    Host->>Lib: open_impedance_bytes = XYParser_Serialize8ChImpedanceCommand(1, impedance_cmd_buf, impedance_cmd_size)
    Lib-->>Host: open_impedance_bytes
    Host->>Dev: 下发阻抗开启命令

    loop 持续获取阻抗
        Dev-->>Host: 原始EEG字节流
        Host->>Lib: frame_count = XYParser_Feed(parser8, raw_data, raw_size, frame8_summaries, max_frames)
        Lib-->>Host: frame_count + frame8_summaries
        Host->>Lib: impedance_count = XYParser_ReadImpedance(parser8, impedance_summaries, max_impedance)
        Lib-->>Host: impedance_count + impedance_summaries
    end

    Host->>Lib: close_impedance_bytes = XYParser_Serialize8ChImpedanceCommand(0, impedance_cmd_buf, impedance_cmd_size)
    Lib-->>Host: close_impedance_bytes
    Host->>Dev: 下发阻抗关闭命令
    Host->>Lib: XYParser_SetImpedanceDetection(parser8, 0)
```

### 3.7 8导算法阶段

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Dev as 8导EEG采集设备
    participant Host as 上位机
    participant Lib as XYParser库
    participant Algo as 算法

    Host->>Lib: XYParser_SetWelchDetection(parser8, 1)

    loop 持续采集
        Dev-->>Host: 原始EEG字节流
        Host->>Lib: frame_count = XYParser_Feed(parser8, raw_data, raw_size, frame8_summaries, max_frames)
        Lib-->>Host: frame_count + frame8_summaries
        Host->>Lib: converted = XYParser_Convert8ChFramesTo64Ch(frame8_summary, 1, frame64_summary, 1)
        Lib-->>Host: converted + frame64_summary
        Host->>Lib: value_count = XYParser_GetAlgorithmDataValueCount()
        Lib-->>Host: value_count
        Host->>Lib: ok = XYParser_ConvertSampleFramesToAlgorithmData(frame64_summary, algorithm_input_data)
        Lib-->>Host: ok + algorithm_input_data
        Host->>Algo: 输入算法数据 algorithm_input_data
        Algo-->>Host: 算法输出数据 algorithm_output_bytes
        Host->>Lib: alg_frame_count = XYParser_FeedAlgorithmData(parser8, algorithm_output_bytes, algorithm_output_size, algorithm_frames, max_algorithm_frames)
        Lib-->>Host: alg_frame_count + algorithm_frames
        Host->>Lib: welch_count = XYParser_ReadWelch(parser8, welch_summaries, max_welch)
        Lib-->>Host: welch_count + welch_summaries
    end

    opt 结束时处理尾数据
        Host->>Lib: flushed = XYParser_FlushAlgorithmData(parser8, tail_frame_summary)
        Lib-->>Host: flushed + tail_frame_summary
        Host->>Lib: welch_count = XYParser_ReadWelch(parser8, welch_summaries, max_welch)
        Lib-->>Host: welch_count + welch_summaries
    end

    Host->>Lib: XYParser_DestroyParser(parser8)
```

### 3.8 8导转64导导联映射关系

8导 workflow 在送入算法前，会先调用 `XYParser_Convert8ChFramesTo64Ch` 将 8 导帧扩展为 64 导帧。

- 8 个输入通道按固定导联位置写入 64 导 summary
- 未覆盖到的其余 56 个 64 导导联全部补 `0`
- `trigger type` 和 `trigger index` 原样透传

映射图如下：

```text
8ch[0] -> PO5
8ch[1] -> POZ
8ch[2] -> PO6
8ch[3] -> PO7
8ch[4] -> O1
8ch[5] -> OZ
8ch[6] -> O2
8ch[7] -> PO8
others -> 0
```

也可以理解为下面这张对应表：

| 8导索引 | 8导写入到的64导导联 |
| --- | --- |
| 0 | PO5 |
| 1 | POZ |
| 2 | PO6 |
| 3 | PO7 |
| 4 | O1 |
| 5 | OZ |
| 6 | O2 |
| 7 | PO8 |

转换过程示意：

```text
XYParser_Feed(8导原始数据)
    -> frame8_summary
    -> XYParser_Convert8ChFramesTo64Ch
    -> frame64_summary
    -> XYParser_ConvertSampleFramesToAlgorithmData
    -> algorithm_input_data
```

代码依据：

- `XYParser_Convert8ChFramesTo64Ch`

### 3.9 2导初始化连接阶段

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Dev as 2导EEG采集设备
    participant Host as 上位机
    participant Lib as XYParser库

    Dev-->>Host: 设备连接成功
    Host->>Lib: parser2 = XYParser_CreateParser(2)
    Host->>Lib: XYParser_SetAdcParams(parser2, 2.42, 6.0)
    Note over Host,Lib: 2导默认 vref = 2.42，与 8/64 导不同
    Host->>Lib: XYParser_SetSampleRate(parser2, 250)
    Host->>Lib: XYParser_SetBypassChecksum(parser2, 1)
    Host->>Lib: gain_cmd_size = XYParser_Get2ChGainSampleRateCommandSize()
    Lib-->>Host: gain_cmd_size
    Host->>Lib: gain_cmd_bytes = XYParser_Serialize2ChGainSampleRateCommand(6, 250, gain_cmd_buf, gain_cmd_size)
    Lib-->>Host: gain_cmd_bytes
    Host->>Dev: 下发采样率250、增益6命令
```

### 3.10 2导阻抗阶段

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Dev as 2导EEG采集设备
    participant Host as 上位机
    participant Lib as XYParser库

    Host->>Lib: XYParser_SetImpedanceDetection(parser2, 1)
    Host->>Lib: XYParser_SetAdcParams(parser2, 2.42, 24.0)
    Note over Host,Lib: 2导阻抗阶段仍使用 vref = 2.42
    Host->>Lib: impedance_gain_cmd_size = XYParser_Get2ChGainSampleRateCommandSize()
    Lib-->>Host: impedance_gain_cmd_size
    Host->>Lib: impedance_gain_cmd_bytes = XYParser_Serialize2ChGainSampleRateCommand(24, 250, impedance_gain_cmd_buf, impedance_gain_cmd_size)
    Lib-->>Host: impedance_gain_cmd_bytes
    Host->>Dev: 下发采样率250、增益24命令
    Host->>Lib: impedance_cmd_size = XYParser_Get2ChImpedanceCommandSize()
    Lib-->>Host: impedance_cmd_size
    Host->>Lib: open_impedance_bytes = XYParser_Serialize2ChImpedanceCommand(1, impedance_cmd_buf, impedance_cmd_size)
    Lib-->>Host: open_impedance_bytes
    Host->>Dev: 下发阻抗开启命令

    loop 持续获取阻抗
        Dev-->>Host: 原始EEG字节流
        Host->>Lib: frame_count = XYParser_Feed(parser2, raw_data, raw_size, frame2_summaries, max_frames)
        Lib-->>Host: frame_count + frame2_summaries
        Host->>Lib: impedance_count = XYParser_ReadImpedance(parser2, impedance_summaries, max_impedance)
        Lib-->>Host: impedance_count + impedance_summaries
    end

    Host->>Lib: close_impedance_bytes = XYParser_Serialize2ChImpedanceCommand(0, impedance_cmd_buf, impedance_cmd_size)
    Lib-->>Host: close_impedance_bytes
    Host->>Dev: 下发阻抗关闭命令
    Host->>Lib: restore_gain_cmd_size = XYParser_Get2ChGainSampleRateCommandSize()
    Lib-->>Host: restore_gain_cmd_size
    Host->>Lib: restore_gain_cmd_bytes = XYParser_Serialize2ChGainSampleRateCommand(6, 250, restore_gain_cmd_buf, restore_gain_cmd_size)
    Lib-->>Host: restore_gain_cmd_bytes
    Host->>Dev: 下发采样率250、增益6命令
    Host->>Lib: XYParser_SetAdcParams(parser2, 2.42, 6.0)
    Host->>Lib: XYParser_SetImpedanceDetection(parser2, 0)
```

### 3.11 2导转64导导联映射关系

2导 workflow 在送入算法前，会先调用 `XYParser_Convert2ChFramesTo64Ch` 将 2 导帧扩展为 64 导帧。

- 2 个输入通道按固定导联位置写入 64 导 summary
- 未覆盖到的其余 62 个 64 导导联全部补 `0`
- `trigger type` 和 `trigger index` 原样透传

映射图如下：

```text
2ch[0] -> FP1
2ch[1] -> FP2
others -> 0
```

也可以理解为下面这张对应表：

| 2导索引 | 2导写入到的64导导联 |
| --- | --- |
| 0 | FP1 |
| 1 | FP2 |

转换过程示意：

```text
XYParser_Feed(2导原始数据)
    -> frame2_summary
    -> XYParser_Convert2ChFramesTo64Ch
    -> frame64_summary
    -> XYParser_ConvertSampleFramesToAlgorithmData
    -> algorithm_input_data
```

代码依据：

- `XYParser_Convert2ChFramesTo64Ch`
- `Convert2ChSummaryTo64ChSummary`
- `k2ChLeadMap = { FP1, FP2 }`

### 3.12 2导 workflow 时序图

2导 demo 的完整链路与 64 导 workflow 保持一致，只是前端输入是 2 导，送算法前会先补成 64 导。

```mermaid
%%{init: {'theme': 'default', 'sequence': {'diagramMarginX': 80, 'diagramMarginY': 30, 'actorMargin': 80, 'width': 220, 'height': 80, 'messageMargin': 35}}}%%
sequenceDiagram
    participant Dev as 2导EEG采集设备
    participant Host as 上位机
    participant Lib as XYParser库
    participant Algo as 算法

    Host->>Lib: XYParser_SetWelchDetection(parser2, 1)

    loop 持续采集
        Dev-->>Host: 原始EEG字节流
        Host->>Lib: frame_count = XYParser_Feed(parser2, raw_data, raw_size, frame2_summaries, max_frames)
        Lib-->>Host: frame_count + frame2_summaries
        Host->>Lib: converted = XYParser_Convert2ChFramesTo64Ch(frame2_summary, 1, frame64_summary, 1)
        Lib-->>Host: converted + frame64_summary
        Host->>Lib: value_count = XYParser_GetAlgorithmDataValueCount()
        Lib-->>Host: value_count
        Host->>Lib: ok = XYParser_ConvertSampleFramesToAlgorithmData(frame64_summary, algorithm_input_data)
        Lib-->>Host: ok + algorithm_input_data
        Host->>Algo: 输入算法数据 algorithm_input_data
        Algo-->>Host: 算法输出数据 algorithm_output_bytes
        Host->>Lib: alg_frame_count = XYParser_FeedAlgorithmData(parser2, algorithm_output_bytes, algorithm_output_size, algorithm_frames, max_algorithm_frames)
        Lib-->>Host: alg_frame_count + algorithm_frames
        Host->>Lib: welch_count = XYParser_ReadWelch(parser2, welch_summaries, max_welch)
        Lib-->>Host: welch_count + welch_summaries
    end

    opt 结束时处理尾数据
        Host->>Lib: flushed = XYParser_FlushAlgorithmData(parser2, tail_frame_summary)
        Lib-->>Host: flushed + tail_frame_summary
        Host->>Lib: welch_count = XYParser_ReadWelch(parser2, welch_summaries, max_welch)
        Lib-->>Host: welch_count + welch_summaries
    end

    Host->>Lib: XYParser_DestroyParser(parser2)
```

时序步骤如下：

1. `XYParser2Demo` 通过 TCP 接收 2 导原始数据。
2. `XYParser_Feed(handle=2)` 解析出 `frame2_summary`。
3. `XYParser_Convert2ChFramesTo64Ch` 将 `FP1/FP2` 写入 64 导 summary，其余导联补 `0`。
4. `XYParser_ConvertSampleFramesToAlgorithmData` 将 64 导 summary 打平成算法输入。
5. ZMQ 将 64 通道 payload 发给算法服务端。
6. 算法服务端回 64 通道结果。
7. `XYParser_FeedAlgorithmData` 将算法回包喂回 parser。
8. Welch/PSD 输出 `peakHz`、`peakPsd`、各 band 能量等结果。

如果打开阻抗流程，则 2 导还会额外穿插以下控制阶段：

- 发送 2 导阻抗开启命令
- 发送 `250Hz / 24增益`
- 打开 parser 阻抗开关
- 读取并打印阻抗
- 阻抗结束后发送 `250Hz / 6增益`
- 发送 2 导阻抗关闭命令
- 关闭 parser 阻抗开关
- 恢复 2->64->算法->Welch 主链路
- `Convert2ChSummaryTo64ChSummary`
- 2导映射表 `k2ChLeadMap`