Miyamizu-MitsuhaSang
parent
2a96ce0a3d
commit
ff9b8d0f4e
529
README.md
529
README.md
|
|
@ -63,7 +63,7 @@ Authorization: Bearer <your_jwt_token>
|
|||
|
||||
```json
|
||||
{
|
||||
"user_email" : "string"
|
||||
"email" : "string"
|
||||
}
|
||||
```
|
||||
|
||||
|
|
@ -221,53 +221,106 @@ Authorization: Bearer <your_jwt_token>
|
|||
- `200`: 密码重置成功
|
||||
- `400`: 密码不合法或令牌无效
|
||||
|
||||
#### 1.8 手机找回密码（已废弃）
|
||||
|
||||
> **说明**: 该接口仍在服务端保留，但已不再推荐使用，后续版本可能会移除。
|
||||
|
||||
- **接口**: `POST /users/auth/forget-password/phone`
|
||||
- **描述**: 通过手机号码请求验证码以找回密码
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
{
|
||||
"phone_number": "string"
|
||||
}
|
||||
```
|
||||
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"message": "验证码已发送"
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 发送成功
|
||||
- `404`: 手机号未注册
|
||||
|
||||
#### 1.9 手机验证码验证（已废弃）
|
||||
|
||||
> **说明**: 该接口与 1.8 配合使用，已不再推荐使用。
|
||||
|
||||
- **接口**: `POST /users/auth/varify_code`
|
||||
- **描述**: 校验短信验证码是否有效
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
{
|
||||
"phone": "string",
|
||||
"code": "string"
|
||||
}
|
||||
```
|
||||
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"message": "验证成功，可以重置密码"
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 验证成功
|
||||
- `400`: 验证码错误或已过期
|
||||
|
||||
---
|
||||
|
||||
### 2. 词典搜索模块
|
||||
### 2. 词典搜索模块 (`Dictionary Search API`)
|
||||
|
||||
#### 2.1 词典搜索
|
||||
#### 2.1 单词精确搜索
|
||||
|
||||
- **接口**: `POST /search`
|
||||
- **描述**: 根据关键词搜索词典内容
|
||||
- **接口**: `POST /search/word`
|
||||
- **描述**: 根据语言精确查询词条，自动累计词频并返回按词性分组的释义。
|
||||
- **需要认证**: 是
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
{
|
||||
"query": "string",
|
||||
"language": "fr" | "jp",
|
||||
"sort": "relevance" | "date",
|
||||
"order": "asc" | "des"
|
||||
"query": "bonjour",
|
||||
"language": "fr",
|
||||
"sort": "relevance",
|
||||
"order": "des"
|
||||
}
|
||||
```
|
||||
|
||||
- **法语响应示例**:
|
||||
- **法语响应示例** (`language = fr`):
|
||||
|
||||
```json
|
||||
{
|
||||
"query": "string",
|
||||
"pos": ["n.m.", "v.t."],
|
||||
"query": "bonjour",
|
||||
"pos": ["n.m."],
|
||||
"contents": [
|
||||
{
|
||||
"pos": "n.m.",
|
||||
"chi_exp": "中文解释",
|
||||
"eng_explanation": "English explanation",
|
||||
"example": "例句"
|
||||
"chi_exp": "问候语；用于见面时打招呼",
|
||||
"eng_explanation": "greeting; hello",
|
||||
"example": "Bonjour, comment ça va ?"
|
||||
}
|
||||
]
|
||||
}
|
||||
```
|
||||
|
||||
- **日语响应示例**:
|
||||
- **日语响应示例** (`language = jp`):
|
||||
|
||||
```json
|
||||
{
|
||||
"query": "string",
|
||||
"pos": ["名词", "动词"],
|
||||
"query": "日本語",
|
||||
"pos": ["名词"],
|
||||
"contents": [
|
||||
{
|
||||
"chi_exp": "中文解释",
|
||||
"example": "例句"
|
||||
"chi_exp": "日语；日本的语言",
|
||||
"example": "日本語を勉強しています。"
|
||||
}
|
||||
]
|
||||
}
|
||||
|
|
@ -278,17 +331,77 @@ Authorization: Bearer <your_jwt_token>
|
|||
- `404`: 未找到词条
|
||||
- `401`: 未授权
|
||||
|
||||
#### 2.2 搜索建议
|
||||
#### 2.2 法语谚语详情
|
||||
|
||||
- **接口**: `POST /search/list`
|
||||
- **描述**: 获取搜索自动完成建议
|
||||
- **接口**: `POST /search/proverb`
|
||||
- **描述**: 通过谚语ID获取法语谚语原文与中文解释。
|
||||
- **需要认证**: 是
|
||||
- **查询参数**:
|
||||
- `proverb_id`: 谚语ID (integer)
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"proverb_text": "Petit à petit, l'oiseau fait son nid.",
|
||||
"chi_exp": "循序渐进才能取得成功。"
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 查询成功
|
||||
- `404`: 谚语不存在
|
||||
|
||||
#### 2.3 单词联想建议
|
||||
|
||||
- **接口**: `POST /search/word/list`
|
||||
- **描述**: 根据用户输入返回单词联想列表，含前缀匹配与包含匹配。
|
||||
- **需要认证**: 是
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
{
|
||||
"query": "string",
|
||||
"language": "fr" | "jp"
|
||||
"query": "bon",
|
||||
"language": "fr",
|
||||
"sort": "relevance",
|
||||
"order": "des"
|
||||
}
|
||||
```
|
||||
|
||||
- **响应示例**:
|
||||
|
||||
```json
|
||||
{
|
||||
"list": ["bonjour", "bonsoir", "bonheur"]
|
||||
}
|
||||
```
|
||||
|
||||
> **说明**: `language = "jp"` 时返回形如 `[["愛", "あい"], ["愛する", "あいする"]]` 的二维数组，第二列为假名读音。
|
||||
|
||||
#### 2.4 谚语联想建议
|
||||
|
||||
- **接口**: `POST /search/proverb/list`
|
||||
- **描述**: 按输入内容（自动识别法语或中文）返回谚语候选列表。
|
||||
- **需要认证**: 是
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
{
|
||||
"query": "慢",
|
||||
"language": "fr"
|
||||
}
|
||||
```
|
||||
|
||||
- **响应示例**:
|
||||
|
||||
```json
|
||||
{
|
||||
"list": [
|
||||
{
|
||||
"id": 12,
|
||||
"proverb": "Rien ne sert de courir, il faut partir à point.",
|
||||
"chi_exp": "做事要循序渐进，贵在及时出发。"
|
||||
}
|
||||
]
|
||||
}
|
||||
```
|
||||
|
||||
|
|
@ -358,7 +471,12 @@ Authorization: Bearer <your_jwt_token>
|
|||
|
||||
```json
|
||||
{
|
||||
"pong": true
|
||||
"pong": true,
|
||||
"redis": {
|
||||
"host": "127.0.0.1",
|
||||
"port": 6379,
|
||||
"db": 0
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
|
|
@ -508,9 +626,291 @@ Authorization: Bearer <your_jwt_token>
|
|||
|
||||
---
|
||||
|
||||
### 6. 数据模型
|
||||
### 6. 用户反馈模块 (`/improvements`)
|
||||
|
||||
#### 6.1 法语词性枚举
|
||||
#### 6.1 提交用户反馈
|
||||
|
||||
- **接口**: `POST /improvements`
|
||||
- **描述**: 登录用户提交产品改进或问题反馈，系统会向预设邮箱发送通知。
|
||||
- **需要认证**: 是
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
{
|
||||
"report_part": "string",
|
||||
"text": "string"
|
||||
}
|
||||
```
|
||||
|
||||
- **字段说明**:
|
||||
- `report_part`: 反馈类别，可选值 `ui_design`, `dict_fr`, `dict_jp`, `user`, `translate`, `writting`, `ai_assist`, `pronounce`（`comment_api_test` 仅用于内部测试）
|
||||
- `text`: 反馈正文，不能为空
|
||||
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"massages": "feedback succeed"
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 提交成功
|
||||
- `422`: 字段校验失败（不合法的类别或空文本）
|
||||
|
||||
---
|
||||
|
||||
### 7. 词条评论模块 (`/comment/word`)
|
||||
|
||||
#### 7.1 新增词条评论
|
||||
|
||||
- **接口**: `POST /comment/word/{lang}`
|
||||
- **描述**: 为指定语言的词条添加用户评论
|
||||
- **需要认证**: 是
|
||||
- **路径参数**:
|
||||
- `lang`: `fr` 或 `jp`
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
{
|
||||
"comment_word": "string",
|
||||
"comment_content": "string"
|
||||
}
|
||||
```
|
||||
|
||||
- **响应**: 创建成功时返回 `200`，响应体为空。
|
||||
- **状态码**:
|
||||
- `200`: 创建成功
|
||||
- `422`: 字段校验失败
|
||||
|
||||
---
|
||||
|
||||
### 8. 作文指导模块 (`/article-director`)
|
||||
|
||||
#### 8.1 作文批改会话
|
||||
|
||||
- **接口**: `POST /article-director/article`
|
||||
- **描述**: 将学生作文（文本形式）提交给 EduChat 模型获取结构化点评，会话上下文保存在 Redis 中。
|
||||
- **需要认证**: 是
|
||||
- **查询参数**:
|
||||
- `lang`: 作文语种，默认 `fr-FR`，可选值 `fr-FR`（法语）、`ja-JP`（日语）、`en-US`（英文）
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
{
|
||||
"title_content": "我的作文全文......",
|
||||
"article_type": "议论文"
|
||||
}
|
||||
```
|
||||
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"reply": "整体点评内容……",
|
||||
"tokens": 1834,
|
||||
"conversation_length": 3
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 批改成功
|
||||
- `401`: 未授权
|
||||
|
||||
> **提示**: 每次调用批改/追问接口之后，前端应根据需要调用重置接口清空 Redis 中的上下文。
|
||||
|
||||
#### 8.2 作文追问
|
||||
|
||||
- **接口**: `POST /article-director/question`
|
||||
- **描述**: 在现有作文会话上追加提问，获取针对性回复。
|
||||
- **需要认证**: 是
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
{
|
||||
"query": "请给出第三段的改写示例"
|
||||
}
|
||||
```
|
||||
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"reply": "改写建议……",
|
||||
"tokens": 924,
|
||||
"conversation_length": 5
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 追问成功
|
||||
- `401`: 未授权
|
||||
|
||||
#### 8.3 重置作文会话
|
||||
|
||||
- **接口**: `POST /article-director/reset`
|
||||
- **描述**: 清除当前用户在 Redis 中的作文指导上下文，确保下一次批改从头开始。
|
||||
- **需要认证**: 是
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"message": "已重置用户 <user_id> 的作文对话记录"
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 9. 发音测评模块 (`/test/pron`)
|
||||
|
||||
#### 9.1 开始/恢复测评
|
||||
|
||||
- **接口**: `GET /test/pron/start`
|
||||
- **描述**: 为当前用户新建或恢复发音测评会话，默认随机抽取20句目标语言的测评文本。
|
||||
- **需要认证**: 是
|
||||
- **查询参数**:
|
||||
- `count`: 抽题数量 (integer，默认 `20`)
|
||||
- `lang`: 语种代码，支持 `fr-FR`（法语）、`ja-JP`（日语），默认 `fr-FR`
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"ok": true,
|
||||
"resumed": false,
|
||||
"message": "New fr-FR test started",
|
||||
"session": {
|
||||
"lang": "fr-FR",
|
||||
"current_index": 0,
|
||||
"sentence_ids": [12, 45, 87],
|
||||
"total": 3
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 成功创建或恢复会话
|
||||
- `400`: 不支持的语言参数
|
||||
- `404`: 题库为空
|
||||
|
||||
#### 9.2 提交语音测评
|
||||
|
||||
- **接口**: `POST /test/pron/sentence_test`
|
||||
- **描述**: 上传 `.wav` 录音进行发音测评，服务端自动转换格式并调用 Azure Speech 评分。
|
||||
- **需要认证**: 是
|
||||
- **请求类型**: `multipart/form-data`
|
||||
- **表单字段**:
|
||||
- `record`: 上传的音频文件（仅支持 `.wav`）
|
||||
- `lang`: 语种代码，默认 `fr-FR`
|
||||
- **响应示例**:
|
||||
|
||||
```json
|
||||
{
|
||||
"ok": true,
|
||||
"data": {
|
||||
"ok": true,
|
||||
"recognized_text": "Bonjour tout le monde",
|
||||
"overall_score": 84.5,
|
||||
"accuracy": 82.0,
|
||||
"fluency": 86.0,
|
||||
"completeness": 85.0,
|
||||
"progress": "3/10"
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 评分成功（若全部句子完成，会自动结束会话）
|
||||
- `400`: 会话不存在或音频转换失败
|
||||
- `404`: 对应题目不存在
|
||||
- `415`: 音频格式不符合要求
|
||||
|
||||
#### 9.3 查询当前题目
|
||||
|
||||
- **接口**: `GET /test/pron/current_sentence`
|
||||
- **描述**: 返回当前需要朗读的句子。
|
||||
- **需要认证**: 是
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"ok": true,
|
||||
"index": 2,
|
||||
"current_sentence": "Bonjour tout le monde"
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 查询成功
|
||||
- `404`: 会话不存在
|
||||
|
||||
#### 9.4 查看本次题目列表
|
||||
|
||||
- **接口**: `POST /test/pron/testlist`
|
||||
- **描述**: 返回本次测评抽取的所有句子列表及其 ID。
|
||||
- **需要认证**: 是
|
||||
- **响应示例**:
|
||||
|
||||
```json
|
||||
[
|
||||
{"id": 12, "text": "Bonjour tout le monde"},
|
||||
{"id": 45, "text": "Je m'appelle Léa"}
|
||||
]
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 查询成功
|
||||
- `404`: 会话不存在
|
||||
|
||||
#### 9.5 结束测评
|
||||
|
||||
- **接口**: `POST /test/pron/finish`
|
||||
- **描述**: 结束当前测评会话，并返回成绩。若测评未完成，需要携带 `confirm=true` 强制结束。
|
||||
- **需要认证**: 是
|
||||
- **请求体**: `application/x-www-form-urlencoded`
|
||||
- `confirm`: boolean，默认 `false`
|
||||
- **响应示例（强制结束）**:
|
||||
|
||||
```json
|
||||
{
|
||||
"ok": true,
|
||||
"forced_end": true,
|
||||
"message": "⚠️ Test forcefully ended. 3/10 sentences completed.",
|
||||
"data": {
|
||||
"ok": true,
|
||||
"average_score": 82.3,
|
||||
"records": [
|
||||
{
|
||||
"sentence_id": 12,
|
||||
"overall_score": 84.5
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
- **状态码**:
|
||||
- `200`: 成功结束会话
|
||||
- `404`: 会话或结果不存在
|
||||
|
||||
#### 9.6 清除测评会话
|
||||
|
||||
- **接口**: `POST /test/pron/clear_session`
|
||||
- **描述**: 主动清除 Redis 中的测评会话（用户放弃测评时使用）。
|
||||
- **需要认证**: 是
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
{
|
||||
"ok": true,
|
||||
"message": "Session cleared"
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 10. 数据模型
|
||||
|
||||
#### 10.1 法语词性枚举
|
||||
|
||||
```text
|
||||
n. - 名词
|
||||
|
|
@ -528,7 +928,7 @@ interj. - 感叹词
|
|||
art. - 冠词
|
||||
```
|
||||
|
||||
#### 6.2 日语词性枚举
|
||||
#### 10.2 日语词性枚举
|
||||
|
||||
```text
|
||||
名词, 形容词, 形容动词, 连用, 一段动词, 五段动词,
|
||||
|
|
@ -538,7 +938,7 @@ art. - 冠词
|
|||
|
||||
---
|
||||
|
||||
### 7. 错误码说明
|
||||
### 11. 错误码说明
|
||||
|
||||
| 状态码 | 说明 |
|
||||
|--------|------|
|
||||
|
|
@ -554,13 +954,13 @@ art. - 冠词
|
|||
|
||||
---
|
||||
|
||||
### 8. AI助手模块 (`/ai_assist`)
|
||||
### 12. AI助手模块 (`/ai_assist`)
|
||||
|
||||
#### 8.1 词语智能问答
|
||||
#### 12.1 词语智能问答
|
||||
|
||||
- **接口**: `POST /ai_assist/exp`
|
||||
- **描述**: 针对指定词语，向AI助手提问相关问题，获取简洁自然的答案，适合初学者。
|
||||
- **需要认证**: 是
|
||||
- **接口**: `POST /ai_assist/word/exp`
|
||||
- **描述**: 针对指定词语向AI助手提问，服务端会基于Redis保存的上下文历史给出简洁、贴合学习者的回答。
|
||||
- **需要认证**: 是（`Bearer` Token）
|
||||
- **请求体**:
|
||||
|
||||
```json
|
||||
|
|
@ -570,6 +970,10 @@ art. - 冠词
|
|||
}
|
||||
```
|
||||
|
||||
- **限制**:
|
||||
- 普通用户调用次数超过100次时会返回 `400 本月API使用量已超`
|
||||
- 每个 `word` 独立维护聊天上下文，历史保存于Redis
|
||||
|
||||
- **响应**:
|
||||
|
||||
```json
|
||||
|
|
@ -584,15 +988,24 @@ art. - 冠词
|
|||
- **状态码**:
|
||||
- `200`: 问答成功
|
||||
- `400`: 本月API使用量已超
|
||||
- `401`: 未授权
|
||||
- `500`: AI调用失败
|
||||
|
||||
#### 8.2 清除词语聊天记录
|
||||
#### 12.2 通用AI对话（预留）
|
||||
|
||||
- **接口**: `POST /ai_assist/univer`
|
||||
- **描述**: 预留的通用AI对话接口，当前版本尚未实现业务逻辑，调用将返回空响应。
|
||||
- **需要认证**: 是
|
||||
- **状态码**:
|
||||
- `200`: 请求成功（响应体为空）
|
||||
|
||||
#### 12.3 清除词语聊天记录
|
||||
|
||||
- **接口**: `POST /ai_assist/clear`
|
||||
- **描述**: 清除指定词语的AI助手聊天记录
|
||||
- **需要认证**: 是
|
||||
- **请求参数**:
|
||||
- `word`: 词语 (string)
|
||||
- `word`: 词语 (query 参数，string)
|
||||
|
||||
- **响应**:
|
||||
|
||||
|
|
@ -607,7 +1020,7 @@ art. - 冠词
|
|||
|
||||
---
|
||||
|
||||
### 9. 使用示例
|
||||
### 13. 使用示例
|
||||
|
||||
#### 完整的API调用流程示例
|
||||
|
||||
|
|
@ -630,15 +1043,26 @@ curl -X POST "http://127.0.0.1:8000/users/login" \
|
|||
}'
|
||||
|
||||
# 3. 使用返回的token进行词典搜索
|
||||
curl -X POST "http://127.0.0.1:8000/search" \
|
||||
curl -X POST "http://127.0.0.1:8000/search/word" \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "Authorization: Bearer <your_token_here>" \
|
||||
-d '{
|
||||
"query": "bonjour",
|
||||
"language": "fr",
|
||||
"sort": "relevance",
|
||||
"order": "des"
|
||||
}'
|
||||
|
||||
# 4. 获取单词联想列表
|
||||
curl -X POST "http://127.0.0.1:8000/search/word/list" \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "Authorization: Bearer <your_token_here>" \
|
||||
-d '{
|
||||
"query": "bon",
|
||||
"language": "fr"
|
||||
}'
|
||||
|
||||
# 4. 使用翻译API
|
||||
# 5. 使用翻译API
|
||||
curl -X POST "http://127.0.0.1:8000/translate" \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "Authorization: Bearer <your_token_here>" \
|
||||
|
|
@ -648,11 +1072,11 @@ curl -X POST "http://127.0.0.1:8000/translate" \
|
|||
"to_lang": "zh"
|
||||
}'
|
||||
|
||||
# 5. 测试Redis连接
|
||||
# 6. 测试Redis连接
|
||||
curl -X GET "http://127.0.0.1:8000/ping-redis"
|
||||
|
||||
# 6. 词语智能问答
|
||||
curl -X POST "http://127.0.0.1:8000/ai_assist/exp" \
|
||||
# 7. 词语智能问答
|
||||
curl -X POST "http://127.0.0.1:8000/ai_assist/word/exp" \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "Authorization: Bearer <your_token_here>" \
|
||||
-d '{
|
||||
|
|
@ -660,18 +1084,18 @@ curl -X POST "http://127.0.0.1:8000/ai_assist/exp" \
|
|||
"question": "什么是法语?"
|
||||
}'
|
||||
|
||||
# 7. 清除词语聊天记录
|
||||
curl -X POST "http://127.0.0.1:8000/ai_assist/clear" \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "Authorization: Bearer <your_token_here>" \
|
||||
-d '{
|
||||
"word": "法语"
|
||||
}'
|
||||
# 8. 清除词语聊天记录
|
||||
curl -X POST "http://127.0.0.1:8000/ai_assist/clear?word=法语" \
|
||||
-H "Authorization: Bearer <your_token_here>"
|
||||
|
||||
# 9. 开启发音测评
|
||||
curl -X GET "http://127.0.0.1:8000/test/pron/start?count=5&lang=fr-FR" \
|
||||
-H "Authorization: Bearer <your_token_here>"
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 9. 开发者说明
|
||||
### 14. 开发者说明
|
||||
|
||||
- **数据库**: 使用MySQL存储词典数据和用户信息
|
||||
- **缓存**: 使用Redis进行token黑名单管理和API限流
|
||||
|
|
@ -681,10 +1105,11 @@ curl -X POST "http://127.0.0.1:8000/ai_assist/clear" \
|
|||
- **文件上传**: 支持Excel格式的批量词典导入
|
||||
- **CORS**: 支持本地开发环境跨域访问
|
||||
- **API文档**: 启动服务后访问 `http://127.0.0.1:8000/docs` 查看Swagger文档
|
||||
- **发音评测**: `/test/pron` 路由已预留，当前版本尚未提供具体接口
|
||||
|
||||
---
|
||||
|
||||
### 10. 部署说明
|
||||
### 15. 部署说明
|
||||
|
||||
1. 安装依赖: `pip install -r requirements.txt`
|
||||
2. 配置数据库连接 (settings.py)
|
||||
|
|
|
|||
|
|
@ -33,7 +33,7 @@ async def dict_exp(
|
|||
:param user:
|
||||
:return:
|
||||
"""
|
||||
if user[0].token_usage > CHAT_TTL and not user[0].is_admin:
|
||||
if user[0].token_usage > MAX_USAGE_PER and not user[0].is_admin:
|
||||
raise HTTPException(status_code=400, detail="本月API使用量已超")
|
||||
|
||||
redis = request.app.state.redis
|
||||
|
|
@ -104,6 +104,6 @@ async def universal_main():
|
|||
@ai_router.post("/clear")
|
||||
async def clear_history(word: str, request: Request, user: Tuple[User, Dict] = Depends(get_current_user)):
|
||||
redis = request.app.state.redis
|
||||
user_id = user[0].id
|
||||
user_id = str(user[0].id)
|
||||
await clear_chat_history(redis, user_id, word)
|
||||
return {"msg": f"已清除 {word} 的聊天记录"}
|
||||
|
|
|
|||
|
|
@ -0,0 +1,10 @@
|
|||
from pydantic import BaseModel
|
||||
|
||||
|
||||
class UserArticleRequest(BaseModel):
|
||||
# theme: Optional[str]
|
||||
title_content: str
|
||||
article_type: str
|
||||
|
||||
class UserQuery(BaseModel):
|
||||
query: str
|
||||
|
|
@ -0,0 +1,103 @@
|
|||
"""
|
||||
每次调用 article-director/article 接口时都要同时调用reset以清空 redis 中的上下文
|
||||
"""
|
||||
from typing import Literal, Dict, Tuple
|
||||
|
||||
from fastapi import APIRouter, Depends
|
||||
from starlette.requests import Request
|
||||
|
||||
from app.api.article_director import service
|
||||
from app.api.article_director.article_schemas import UserArticleRequest, UserQuery
|
||||
from app.models import User
|
||||
from app.utils.security import get_current_user
|
||||
|
||||
article_router = APIRouter()
|
||||
|
||||
|
||||
@article_router.post("/article-director/article")
|
||||
async def article_director(
|
||||
request: Request,
|
||||
upload_article: UserArticleRequest,
|
||||
lang: Literal["en-US", "fr-FR", "ja-JP"] = "fr-FR",
|
||||
user: Tuple[User, Dict] = Depends(get_current_user)
|
||||
):
|
||||
"""
|
||||
文本形式接口，即直接从文本框中获取
|
||||
每次调用本接口的同时都要同时调用reset接口
|
||||
:param upload_article:
|
||||
:param lang:
|
||||
:return:
|
||||
"""
|
||||
redis = request.app.state.redis
|
||||
|
||||
article_lang = "法语" if lang == "fr-FR" else "日语"
|
||||
|
||||
user_id = user[0].id
|
||||
article = upload_article.title_content
|
||||
|
||||
# 读取历史对话
|
||||
session = await service.get_session(redis_client=redis, user_id=user_id)
|
||||
|
||||
# 追加用户输入
|
||||
user_prompt = service.set_user_prompt(upload_article, article_lang=article_lang)
|
||||
session.append({"role": "user", "content": user_prompt})
|
||||
|
||||
# 调用 EduChat 模型
|
||||
completion = service.chat_ecnu_request(session)
|
||||
|
||||
# 取出回答内容
|
||||
assistant_reply = completion.choices[0].message.content
|
||||
|
||||
# 保存模型回复
|
||||
session.append({"role": "assistant", "content": assistant_reply})
|
||||
|
||||
# 存入 Redis
|
||||
await service.save_session(redis, user_id, session)
|
||||
|
||||
return {
|
||||
"reply": assistant_reply,
|
||||
"tokens": completion.usage.total_tokens,
|
||||
"conversation_length": len(session),
|
||||
}
|
||||
|
||||
|
||||
@article_router.post("/article-director/question", description="用户进一步询问")
|
||||
async def further_question(
|
||||
request: Request,
|
||||
user_prompt: UserQuery,
|
||||
user: Tuple[User, Dict] = Depends(get_current_user)
|
||||
):
|
||||
redis = request.app.state.redis
|
||||
|
||||
user_id = user[0].id
|
||||
|
||||
# 读取历史对话
|
||||
session = await service.get_session(redis_client=redis, user_id=user_id)
|
||||
|
||||
# 追加用户输入
|
||||
session.append({"role": "user", "content": user_prompt.query})
|
||||
|
||||
# 调用 EduChat 模型
|
||||
completion = service.chat_ecnu_request(session)
|
||||
|
||||
# 取出回答内容
|
||||
assistant_reply = completion.choices[0].message.content
|
||||
|
||||
# 保存模型回复
|
||||
session.append({"role": "assistant", "content": assistant_reply})
|
||||
|
||||
# 存入 Redis
|
||||
await service.save_session(redis, user_id, session)
|
||||
|
||||
return {
|
||||
"reply": assistant_reply,
|
||||
"tokens": completion.usage.total_tokens,
|
||||
"conversation_length": len(session),
|
||||
}
|
||||
|
||||
@article_router.post("/article-director/reset", description="重置上下文")
|
||||
async def reset_conversation(request: Request, user: Tuple[User, Dict] = Depends(get_current_user)):
|
||||
user_id = user[0].id
|
||||
redis = request.app.state.redis
|
||||
await service.reset_session(redis, user_id)
|
||||
return {"message": f"已重置用户 {user_id} 的作文对话记录"}
|
||||
|
|
@ -0,0 +1,64 @@
|
|||
import json
|
||||
from typing import List, Dict
|
||||
|
||||
from openai import OpenAI
|
||||
from redis import Redis
|
||||
|
||||
from app.api.article_director.article_schemas import UserArticleRequest
|
||||
from settings import settings
|
||||
|
||||
SYSTEM_PROMPT = """
|
||||
# 背景
|
||||
你是一个人工智能助手，名字叫EduChat,是一个由华东师范大学开发的教育领域大语言模型。
|
||||
# 对话主题:作文指导
|
||||
## 作文指导主题的要求：
|
||||
EduChat你需要扮演一位经验丰富的语文老师，现在需要帮助一位学生审阅作文并给出修改建议。请按照以下步骤进行：
|
||||
整体评价：先对作文的整体质量进行简要评价，指出主要优点和需要改进的方向。
|
||||
亮点分析：具体指出作文中的亮点（如结构、描写、情感表达等方面的优点）。
|
||||
具体修改建议：针对作文中的不足，从以下几个方面提出具体修改建议，并给出修改后的示例：
|
||||
语言表达：是否生动、准确？有无冗余或重复？可以如何优化？
|
||||
细节描写：是否足够具体？能否加入更多感官描写（视觉、听觉、嗅觉、触觉等）使画面更立体？
|
||||
情感表达：情感是否自然？能否更深入或升华？
|
||||
结构布局：段落衔接是否自然？开头结尾是否呼应？ （注意：每个建议点都要结合原文具体句子进行分析，并给出修改后的句子或段落作为示例）
|
||||
写作技巧提示：提供2-3条实用的写作技巧（如动态描写公式、感官交织法等），帮助学生举一反三。
|
||||
修改效果总结：简要说明按照建议修改后，作文会有哪些方面的提升（如文学性、情感层次、场景沉浸感等）。
|
||||
请用亲切、鼓励的语气进行点评，保持专业性同时让学生易于接受。
|
||||
"""
|
||||
|
||||
|
||||
def chat_ecnu_request(
|
||||
session: List[Dict[str, str]],
|
||||
):
|
||||
client = OpenAI(
|
||||
api_key=settings.ECNU_TEACH_AI_KEY,
|
||||
base_url="https://chat.ecnu.edu.cn/open/api/v1"
|
||||
)
|
||||
completion = client.chat.completions.create(
|
||||
model="educhat-r1",
|
||||
messages=session,
|
||||
temperature=0.8, # 保持创造性
|
||||
top_p=0.9, # 保持多样性
|
||||
)
|
||||
|
||||
return completion
|
||||
|
||||
def set_user_prompt(user_article: UserArticleRequest, article_lang: str):
|
||||
user_prompt = f"以下是我的{article_lang}作文，作文体裁为{user_article.article_type}，请帮我修改：{user_article.title_content}"
|
||||
return user_prompt
|
||||
|
||||
async def get_session(redis_client: Redis, user_id: str) -> List[Dict[str, str]]:
|
||||
"""从 Redis 读取对话上下文"""
|
||||
data = await redis_client.get(f"session:{user_id}")
|
||||
if data:
|
||||
return json.loads(data)
|
||||
else:
|
||||
# 如果没有记录，创建带 system prompt 的初始会话
|
||||
return [{"role": "system", "content": SYSTEM_PROMPT}]
|
||||
|
||||
async def save_session(redis_client: Redis, user_id: str, session: List[Dict[str, str]]):
|
||||
"""保存对话上下文到 Redis"""
|
||||
await redis_client.setex(f"session:{user_id}", 86400, json.dumps(session))
|
||||
|
||||
async def reset_session(redis_client: Redis, user_id: str):
|
||||
"""清空用户上下文"""
|
||||
await redis_client.delete(f"session:{user_id}")
|
||||
|
|
@ -1,3 +1,308 @@
|
|||
from fastapi import APIRouter
|
||||
import json
|
||||
import os
|
||||
import random
|
||||
import tempfile
|
||||
from typing import Literal, Tuple, Dict
|
||||
|
||||
import azure.cognitiveservices.speech as speechsdk
|
||||
from fastapi import APIRouter, Depends, UploadFile, File, HTTPException, Form
|
||||
from starlette.requests import Request
|
||||
|
||||
from app.api.pronounciation_test import service
|
||||
from app.models import PronunciationTestFr, User, PronunciationTestJp
|
||||
from app.utils.security import get_current_user
|
||||
from settings import settings
|
||||
|
||||
pron_test_router = APIRouter()
|
||||
|
||||
AZURE_KEY = settings.AZURE_SUBSCRIPTION_KEY
|
||||
SERVICE_REGION = "eastasia"
|
||||
|
||||
speech_config = speechsdk.SpeechConfig(subscription=AZURE_KEY, region=SERVICE_REGION)
|
||||
audio_config = speechsdk.audio.AudioConfig(filename="test.wav")
|
||||
|
||||
|
||||
@pron_test_router.get("/start")
|
||||
async def start_test(
|
||||
request: Request,
|
||||
count: int = 20,
|
||||
lang: Literal["fr-FR", "ja-JP"] = Form("fr-FR"),
|
||||
user: Tuple[User, Dict] = Depends(get_current_user)
|
||||
):
|
||||
"""
|
||||
开始新的发音测评会话：
|
||||
- 若存在未完成测试，则自动恢复；
|
||||
- 若无会话，则随机选取句子并创建新的 session；
|
||||
- 支持多语言（法语/日语）。
|
||||
"""
|
||||
redis = request.app.state.redis
|
||||
user_id = user[0].id
|
||||
|
||||
key = f"test_session:{user_id}"
|
||||
data = await redis.get(key)
|
||||
|
||||
# === 若存在未完成的测试会话 ===
|
||||
if data:
|
||||
session = json.loads(data)
|
||||
return {
|
||||
"ok": True,
|
||||
"resumed": True,
|
||||
"message": "Resumed existing test",
|
||||
"session": session
|
||||
}
|
||||
|
||||
# === 根据语言选择对应题库 ===
|
||||
if lang == "fr-FR":
|
||||
total_count = await PronunciationTestFr.all().count()
|
||||
table = PronunciationTestFr
|
||||
elif lang == "ja-JP":
|
||||
total_count = await PronunciationTestJp.all().count()
|
||||
table = PronunciationTestJp
|
||||
else:
|
||||
raise HTTPException(status_code=400, detail="Unsupported language code")
|
||||
|
||||
# === 随机抽取句子 ID ===
|
||||
if total_count == 0:
|
||||
raise HTTPException(status_code=404, detail=f"No test sentences found for {lang}")
|
||||
|
||||
selected = random.sample(range(1, total_count + 1), k=min(count, total_count))
|
||||
|
||||
# === 构建并保存会话 ===
|
||||
session = {
|
||||
"lang": lang, # ← 新增语言字段
|
||||
"current_index": 0,
|
||||
"sentence_ids": selected,
|
||||
"total": len(selected),
|
||||
}
|
||||
|
||||
await redis.set(key, json.dumps(session), ex=3600)
|
||||
|
||||
return {
|
||||
"ok": True,
|
||||
"resumed": False,
|
||||
"message": f"New {lang} test started",
|
||||
"session": session
|
||||
}
|
||||
|
||||
|
||||
@pron_test_router.post("/sentence_test")
|
||||
async def pron_sentence_test(
|
||||
request: Request,
|
||||
record: UploadFile = File(...),
|
||||
lang: Literal["fr-FR", "ja-JP"] = Form("fr-FR"),
|
||||
user: Tuple[User, Dict] = Depends(get_current_user)
|
||||
):
|
||||
"""
|
||||
目前暂时只提供打分服务，不支持回听录音
|
||||
:param request:
|
||||
:param record:
|
||||
:param lang:
|
||||
:param user:
|
||||
:return:
|
||||
"""
|
||||
redis = request.app.state.redis
|
||||
user_id = user[0].id
|
||||
|
||||
key = f"test_session:{user_id}"
|
||||
data = await redis.get(key)
|
||||
if not data:
|
||||
return {"ok": False, "error": "No active test session"}
|
||||
|
||||
session = json.loads(data)
|
||||
sentence_ids = session["sentence_ids"]
|
||||
index = session["current_index"]
|
||||
|
||||
if index >= len(sentence_ids):
|
||||
await redis.delete(key)
|
||||
return {"ok": True, "finished": True, "message": "All sentences tested"}
|
||||
|
||||
sentence_id = sentence_ids[index]
|
||||
sentence = await PronunciationTestFr.get(id=sentence_id)
|
||||
if not sentence:
|
||||
raise HTTPException(status_code=404, detail=f"Sentence {sentence_id} not found")
|
||||
text = sentence.text
|
||||
|
||||
if not record.filename.endswith(".wav"):
|
||||
raise HTTPException(status_code=415, detail="Invalid file suffix, only '.wav' supported")
|
||||
|
||||
with tempfile.NamedTemporaryFile(delete=False, suffix=os.path.splitext(record.filename)[1]) as tmp:
|
||||
tmp.write(await record.read())
|
||||
tmp.flush()
|
||||
src_path = tmp.name
|
||||
|
||||
# 调用转换函数
|
||||
norm_path = src_path + "_norm.wav"
|
||||
result = service.convert_to_pcm16_mono_wav(src_path, norm_path)
|
||||
if not result["ok"]:
|
||||
raise HTTPException(status_code=400, detail=result["message"])
|
||||
|
||||
# 再验证格式
|
||||
if not service.verify_audio_format(norm_path):
|
||||
raise HTTPException(status_code=415, detail="Invalid audio format")
|
||||
|
||||
try:
|
||||
result = service.assess_pronunciation(norm_path, text, lang)
|
||||
if not result["ok"]:
|
||||
raise HTTPException(status_code=400, detail=result)
|
||||
except HTTPException as e:
|
||||
return result
|
||||
except Exception as e:
|
||||
return {"ok": False, "error": str(e)}
|
||||
finally:
|
||||
os.remove(norm_path)
|
||||
|
||||
await service.save_pron_result(
|
||||
redis=redis,
|
||||
user_id=user[0].id,
|
||||
sentence_id=sentence_id,
|
||||
text=text,
|
||||
scores=result,
|
||||
expire=3600
|
||||
)
|
||||
|
||||
session["current_index"] += 1
|
||||
await redis.set(key, json.dumps(session), ex=3600)
|
||||
|
||||
result["progress"] = f"{session['current_index']}/{len(sentence_ids)}"
|
||||
|
||||
return {"ok": True, "data": result}
|
||||
|
||||
|
||||
@pron_test_router.get("/current_sentence")
|
||||
async def get_current_sentence(
|
||||
request: Request,
|
||||
user: Tuple[User, Dict] = Depends(get_current_user),
|
||||
):
|
||||
redis = request.app.state.redis
|
||||
user_id = user[0].id
|
||||
|
||||
key = f"test_session:{user_id}"
|
||||
data = await redis.get(key)
|
||||
if not data:
|
||||
return {"ok": False, "error": "No active test session"}
|
||||
|
||||
session = json.loads(data)
|
||||
sentence_ids = session["sentence_ids"]
|
||||
index = session["current_index"]
|
||||
if index >= len(sentence_ids):
|
||||
return {"ok": True, "finished": True, "message": "All sentences tested"}
|
||||
sentence_id = sentence_ids[index]
|
||||
sentence = await PronunciationTestFr.get(id=sentence_id)
|
||||
if not sentence:
|
||||
return {"ok": False, "error": "Sentence not found"}
|
||||
text = sentence.text
|
||||
|
||||
return {
|
||||
"ok": True,
|
||||
"index": index,
|
||||
"current_sentence": text,
|
||||
}
|
||||
|
||||
|
||||
@pron_test_router.post("/testlist")
|
||||
async def get_testlist(
|
||||
request: Request,
|
||||
user: Tuple[User, Dict] = Depends(get_current_user),
|
||||
):
|
||||
redis = request.app.state.redis
|
||||
user_id = user[0].id
|
||||
|
||||
key = f"test_session:{user_id}"
|
||||
data = await redis.get(key)
|
||||
if not data:
|
||||
return {"ok": False, "error": "No active test session"}
|
||||
|
||||
session = json.loads(data)
|
||||
sentence_ids = session["sentence_ids"]
|
||||
sentences = []
|
||||
|
||||
for sentence_id in sentence_ids:
|
||||
sentence = await PronunciationTestFr.get(id=sentence_id)
|
||||
if not sentence:
|
||||
raise HTTPException(status_code=404, detail=f"Sentence {sentence_id} not found")
|
||||
text = sentence.text
|
||||
sentences.append({"id": sentence_id, "text": text})
|
||||
|
||||
return sentences
|
||||
|
||||
|
||||
@pron_test_router.post("/finish")
|
||||
async def finish_test(
|
||||
request: Request,
|
||||
confirm: bool = Form(False),
|
||||
user: Tuple[User, Dict] = Depends(get_current_user),
|
||||
):
|
||||
"""
|
||||
结束测试：
|
||||
- 若用户未开始测试 → 返回提示；
|
||||
- 若测试未完成且 confirm=False → 返回提示；
|
||||
- 若测试未完成但 confirm=True → 强制结束，返回已完成部分结果；
|
||||
- 若测试已完成 → 返回完整成绩并清除缓存。
|
||||
"""
|
||||
redis = request.app.state.redis
|
||||
user_id = user[0].id
|
||||
session_key = f"test_session:{user_id}"
|
||||
|
||||
session_data = await redis.get(session_key)
|
||||
if not session_data:
|
||||
return {"ok": False, "message": "No active test session to finish"}
|
||||
|
||||
session = json.loads(session_data)
|
||||
current_index = session.get("current_index", 0)
|
||||
sentence_ids = session.get("sentence_ids", [])
|
||||
total = len(sentence_ids)
|
||||
lang = session["lang"]
|
||||
|
||||
if current_index < len(sentence_ids):
|
||||
remaining = total - current_index
|
||||
# 如果没有确认，则提醒用户
|
||||
if not confirm:
|
||||
return {
|
||||
"ok": False,
|
||||
"unfinished": True,
|
||||
"message": f"Test not finished. {remaining} sentence(s) remaining. "
|
||||
"Resend with confirm=true to force end and view partial results."
|
||||
}
|
||||
|
||||
# 如果用户确认强制结束，则读取已完成部分成绩
|
||||
result = await service.get_pron_result(redis, user_id, delete_after=True)
|
||||
await redis.delete(session_key)
|
||||
|
||||
return {
|
||||
"ok": True,
|
||||
"forced_end": True,
|
||||
"message": f"⚠️ Test forcefully ended. {current_index}/{total} sentences completed.",
|
||||
"data": result
|
||||
}
|
||||
|
||||
# === 已完成测试 ===
|
||||
result = await service.get_pron_result(redis, user_id, delete_after=True)
|
||||
if not result["ok"]:
|
||||
raise HTTPException(status_code=404, detail=result.get("error", "Unknown error"))
|
||||
# 删除 Redis session
|
||||
await redis.delete(session_key)
|
||||
|
||||
# 存入数据库
|
||||
record = await service.record_test_result(user=user[0], result=result, lang=lang)
|
||||
|
||||
return {
|
||||
"ok": True,
|
||||
"message": "Test session cleared",
|
||||
"data": result
|
||||
}
|
||||
|
||||
|
||||
@pron_test_router.post("/clear_session")
|
||||
async def clear_session(request: Request, user: Tuple[User, Dict] = Depends(get_current_user)):
|
||||
"""
|
||||
用户在未完成测试的情况下选择退出，询问是否保存进度，如果不保存则调用本接口清除 Redis
|
||||
"""
|
||||
redis = request.app.state.redis
|
||||
user_id = user[0].id
|
||||
|
||||
key = f"test_session:{user_id}"
|
||||
await redis.delete(key)
|
||||
return {
|
||||
"ok": True,
|
||||
"message": "Session cleared",
|
||||
}
|
||||
|
|
|
|||
|
|
@ -0,0 +1,345 @@
|
|||
import contextlib
|
||||
import json
|
||||
import os
|
||||
import wave
|
||||
from io import BytesIO
|
||||
from typing import Literal, Dict, Any, List
|
||||
|
||||
import azure.cognitiveservices.speech as speechsdk
|
||||
from fastapi import HTTPException
|
||||
from pydub import AudioSegment
|
||||
from redis.asyncio import Redis
|
||||
|
||||
from app.models import User
|
||||
from app.models.base import UserTestRecord
|
||||
from settings import settings
|
||||
|
||||
|
||||
# from imageio_ffmpeg import get_ffmpeg_exe
|
||||
# AudioSegment.converter = get_ffmpeg_exe()
|
||||
|
||||
|
||||
def verify_audio_format(path: str) -> bool:
|
||||
"""
|
||||
检测音频文件是否符合 Azure Speech 要求:
|
||||
采样率 16000Hz, 16-bit, 单声道 (PCM).
|
||||
返回字典包含格式信息和布尔结果。
|
||||
"""
|
||||
if not os.path.exists(path):
|
||||
raise FileNotFoundError(f"Audio file not found: {path}")
|
||||
|
||||
try:
|
||||
with contextlib.closing(wave.open(path, 'rb')) as wf:
|
||||
rate = wf.getframerate()
|
||||
channels = wf.getnchannels()
|
||||
width = wf.getsampwidth()
|
||||
|
||||
ok = (rate == 16000 and channels == 1 and width == 2)
|
||||
if not ok:
|
||||
raise HTTPException(
|
||||
status_code=400,
|
||||
detail={
|
||||
"ok": False,
|
||||
"rate": rate,
|
||||
"channels": channels,
|
||||
"width": width,
|
||||
"message": (
|
||||
f"⚠️ Invalid format (rate={rate}, channels={channels}, width={width}). "
|
||||
"Expected: 16000Hz, mono, 16-bit PCM."
|
||||
)
|
||||
}
|
||||
)
|
||||
except wave.Error as e:
|
||||
raise HTTPException(status_code=401, detail=f"Invalid WAV file: {e}")
|
||||
return True
|
||||
|
||||
def assess_pronunciation(
|
||||
audio_path: str,
|
||||
reference_text: str,
|
||||
lang: Literal["fr-FR", "ja-JP"] = "fr-FR",
|
||||
grading_system: Literal["HundredMark", "FivePoint"] = "FivePoint",
|
||||
granularity: Literal["Phoneme", "Word", "FullText"] = "Phoneme",
|
||||
enable_miscue: bool = True,
|
||||
) -> Dict[str, Any]:
|
||||
"""
|
||||
使用 Azure Speech SDK 对音频文件进行发音测评。（增强错误输出版）
|
||||
:param audio_path: 音频文件路径（必须是 PCM16/Mono/WAV）
|
||||
:param reference_text: 期望朗读的文本
|
||||
:param lang: 语种代码，例如 'fr-FR'（法语）、'ja-JP'（日语）、'en-US'（英语）
|
||||
:param grading_system: 评分体系 ('HundredMark' / 'FivePoint')
|
||||
:param granularity: 评分粒度 ('Phoneme' / 'Word' / 'FullText')
|
||||
:param enable_miscue: 是否检测漏读/多读（True 推荐）
|
||||
:return: 包含整体分、准确度、流畅度、完整度及识别文本的字典
|
||||
"""
|
||||
# === 1. 加载 Azure Speech 配置 ===
|
||||
subsciption_key = settings.AZURE_SUBSCRIPTION_KEY
|
||||
region = "eastasia"
|
||||
print(">>> Azure Key Loaded:", settings.AZURE_SUBSCRIPTION_KEY[:8], "...")
|
||||
print(">>> Azure Region:", "eastasia")
|
||||
|
||||
if not subsciption_key or not region:
|
||||
raise RuntimeError("缺少 Azure Speech 环境变量 AZURE_SPEECH_KEY / AZURE_SPEECH_REGION")
|
||||
|
||||
speech_config = speechsdk.SpeechConfig(subscription=subsciption_key, region=region)
|
||||
speech_config.speech_recognition_language = lang
|
||||
|
||||
# === 2. 加载音频文件 ===
|
||||
audio_config = speechsdk.audio.AudioConfig(filename=audio_path)
|
||||
recognizer = speechsdk.SpeechRecognizer(speech_config=speech_config, audio_config=audio_config)
|
||||
|
||||
print(reference_text)
|
||||
|
||||
# === 3. 构建发音测评配置 ===
|
||||
pron_assestment = speechsdk.PronunciationAssessmentConfig(
|
||||
reference_text=reference_text,
|
||||
grading_system=getattr(speechsdk.PronunciationAssessmentGradingSystem, grading_system),
|
||||
granularity=getattr(speechsdk.PronunciationAssessmentGranularity, granularity),
|
||||
enable_miscue=enable_miscue
|
||||
)
|
||||
pron_assestment.apply_to(recognizer)
|
||||
|
||||
# === 4. 执行识别与打分 ===
|
||||
result = recognizer.recognize_once()
|
||||
|
||||
if result.reason != speechsdk.ResultReason.RecognizedSpeech:
|
||||
return __parse_azure_error(result)
|
||||
|
||||
pa_result = result.properties.get(speechsdk.PropertyId.SpeechServiceResponse_JsonResult)
|
||||
data = json.loads(pa_result)
|
||||
pa_data = data["NBest"][0]["PronunciationAssessment"]
|
||||
|
||||
return {
|
||||
"ok": True,
|
||||
"recognized_text": data.get("DisplayText"),
|
||||
"overall_score": pa_data.get("PronScore"),
|
||||
"accuracy": pa_data.get("AccuracyScore"),
|
||||
"fluency": pa_data.get("FluencyScore"),
|
||||
"completeness": pa_data.get("CompletenessScore")
|
||||
}
|
||||
|
||||
def __parse_azure_error(result: Any) -> Dict[str, Any]:
|
||||
"""
|
||||
从 Azure Speech 识别结果中提取详细错误信息。
|
||||
用于处理 ResultReason != RecognizedSpeech 的情况。
|
||||
:param result: SpeechRecognizer 的识别结果对象
|
||||
:return: 包含 ok=False 与详细错误字段的 dict
|
||||
"""
|
||||
err_data = {
|
||||
"ok": False,
|
||||
"error": str(result.reason),
|
||||
"details": getattr(result, "error_details", None)
|
||||
}
|
||||
|
||||
# ① 无法识别语音（NoMatch）
|
||||
if result.reason == speechsdk.ResultReason.NoMatch:
|
||||
err_data["no_match_details"] = str(getattr(result, "no_match_details", None))
|
||||
print("[Azure] ⚠️ NoMatch: Speech could not be recognized.")
|
||||
print(f"[Azure] Details: {err_data['no_match_details']}")
|
||||
|
||||
# ② 请求被取消（Canceled）
|
||||
elif result.reason == speechsdk.ResultReason.Canceled:
|
||||
cancellation_details = getattr(result, "cancellation_details", None)
|
||||
if cancellation_details:
|
||||
err_data["cancel_reason"] = str(getattr(cancellation_details, "reason", None))
|
||||
err_data["cancel_error_details"] = getattr(cancellation_details, "error_details", None)
|
||||
err_data["cancel_error_code"] = getattr(cancellation_details, "error_code", None)
|
||||
|
||||
print("[Azure] ❌ Canceled by Speech Service")
|
||||
print(f"[Azure] Reason: {err_data['cancel_reason']}")
|
||||
print(f"[Azure] Error details: {err_data['cancel_error_details']}")
|
||||
print(f"[Azure] Error code: {err_data['cancel_error_code']}")
|
||||
else:
|
||||
print("[Azure] ❌ Canceled but no details provided.")
|
||||
|
||||
# ③ 其他未知类型
|
||||
else:
|
||||
print(f"[Azure] ⚠️ Unexpected recognition result: {result.reason}")
|
||||
print(f"[Azure] Error details: {err_data['details']}")
|
||||
|
||||
return err_data
|
||||
|
||||
def convert_to_pcm16_mono_wav(input_path: str, output_path: str):
|
||||
"""
|
||||
将任意音频格式转换为 Azure Speech API 要求的标准 WAV 文件:
|
||||
- 采样率 16 kHz
|
||||
- 单声道
|
||||
- 16 bit PCM
|
||||
"""
|
||||
from pydub import AudioSegment
|
||||
|
||||
try:
|
||||
audio = AudioSegment.from_file(input_path)
|
||||
duration_ms = len(audio)
|
||||
|
||||
# 重新采样
|
||||
audio = audio.set_frame_rate(16000).set_channels(1).set_sample_width(2)
|
||||
audio.export(output_path, format="wav")
|
||||
|
||||
return {
|
||||
"ok": True,
|
||||
"path": output_path,
|
||||
"message": f"Converted successfully ({duration_ms / 1000:.2f}s)"
|
||||
}
|
||||
|
||||
except Exception as e:
|
||||
return {
|
||||
"ok": False,
|
||||
"path": None,
|
||||
"message": f"Audio conversion failed: {str(e)}"
|
||||
}
|
||||
|
||||
def convert_audio_to_memory(file_obj):
|
||||
"""
|
||||
完全在内存中转化（更快）
|
||||
:param file_obj:
|
||||
:return: 转换后的 BinaryStream
|
||||
"""
|
||||
audio = AudioSegment.from_file(file_obj)
|
||||
audio = audio.set_frame_rate(16000).set_channels(1).set_sample_width(2)
|
||||
buf = BytesIO()
|
||||
audio.export(buf, format="wav")
|
||||
buf.seek(0)
|
||||
return buf
|
||||
|
||||
async def save_pron_result(
|
||||
redis: Redis,
|
||||
user_id: int,
|
||||
sentence_id: int,
|
||||
text: str,
|
||||
scores: Dict[str, float],
|
||||
expire: int = 3600
|
||||
) -> None:
|
||||
"""
|
||||
将测评结果保存到 Redis。
|
||||
结构：test_result:{user_id} -> {"sentences": [ {...}, {...} ]}
|
||||
"""
|
||||
key = f"test_result:{user_id}"
|
||||
existing = await redis.get(key)
|
||||
if existing:
|
||||
data = json.loads(existing)
|
||||
else:
|
||||
data = {"sentences": []}
|
||||
|
||||
# 防止重复写入同一条 sentence_id
|
||||
if not any(item["id"] == sentence_id for item in data["sentences"]):
|
||||
entry = {
|
||||
"id": sentence_id,
|
||||
"text": text,
|
||||
"overall": scores.get("overall_score"),
|
||||
"accuracy": scores.get("accuracy"),
|
||||
"fluency": scores.get("fluency"),
|
||||
"completeness": scores.get("completeness")
|
||||
}
|
||||
data["sentences"].append(entry)
|
||||
await redis.set(key, json.dumps(data), ex=expire)
|
||||
|
||||
async def get_pron_result(
|
||||
redis: Redis,
|
||||
user_id: int,
|
||||
delete_after: bool = False
|
||||
) -> Dict[str, Any]:
|
||||
"""
|
||||
从 Redis 获取用户的所有句子测评结果，
|
||||
返回每句分数 + 总分 + 平均分 + 等级评定。
|
||||
"""
|
||||
key = f"test_result:{user_id}"
|
||||
data = await redis.get(key)
|
||||
|
||||
if not data:
|
||||
return {"ok": False, "error": "No result found"}
|
||||
|
||||
result_data = json.loads(data)
|
||||
sentences: List[Dict[str, Any]] = result_data.get("sentences", [])
|
||||
|
||||
if not sentences:
|
||||
return {"ok": False, "error": "Empty result list"}
|
||||
|
||||
fields = ["overall", "accuracy", "fluency", "completeness"]
|
||||
|
||||
# 计算总分与平均分
|
||||
totals = {f: 0.0 for f in fields}
|
||||
counts = {f: 0 for f in fields}
|
||||
for s in sentences:
|
||||
for f in fields:
|
||||
if s.get(f) is not None:
|
||||
totals[f] += s[f]
|
||||
counts[f] += 1
|
||||
|
||||
averages = {
|
||||
f: round(totals[f] / counts[f], 2) if counts[f] else 0.0
|
||||
for f in fields
|
||||
}
|
||||
|
||||
# 等级映射函数
|
||||
def grade(score: float) -> str:
|
||||
if score >= 4.5:
|
||||
return "优秀 🏆"
|
||||
elif score >= 3.5:
|
||||
return "良好 👍"
|
||||
elif score >= 2.5:
|
||||
return "一般 🙂"
|
||||
elif score > 0:
|
||||
return "需改进 ⚠️"
|
||||
return "无数据"
|
||||
|
||||
# 各项等级 + 总体等级
|
||||
grade_map = {f: grade(averages[f]) for f in fields}
|
||||
grade_map["overall_level"] = grade(averages["overall"])
|
||||
|
||||
if delete_after:
|
||||
await redis.delete(key)
|
||||
|
||||
return {
|
||||
"ok": True,
|
||||
"count": len(sentences),
|
||||
"totals": {f: round(totals[f], 2) for f in fields},
|
||||
"average": averages,
|
||||
"grades": grade_map,
|
||||
"sentences": sentences
|
||||
}
|
||||
|
||||
async def record_test_result(
|
||||
user: User,
|
||||
result: Dict[str, Any],
|
||||
lang: Literal["fr", "jp"]
|
||||
) -> Dict[str, Any]:
|
||||
"""
|
||||
将一次完整测评结果写入数据库。
|
||||
|
||||
:param user: 当前用户对象
|
||||
:param result: 从 get_pron_result() 返回的结果字典
|
||||
:param lang: 测试语种 ('fr' 或 'jp')
|
||||
:return: 数据库存储结果摘要
|
||||
"""
|
||||
if not result.get("ok"):
|
||||
return {"ok": False, "error": "Invalid test result"}
|
||||
|
||||
avg = result.get("average", {})
|
||||
grades = result.get("grades", {})
|
||||
count = result.get("count", 0)
|
||||
sentences = result.get("sentences", [])
|
||||
|
||||
# 构建可存储的数据
|
||||
record = await UserTestRecord.create(
|
||||
user=user, # 外键绑定用户对象
|
||||
username=user.name,
|
||||
language=lang,
|
||||
total_sentences=count,
|
||||
average_score=avg.get("overall", 0.0),
|
||||
accuracy_score=avg.get("accuracy", 0.0),
|
||||
fluency_score=avg.get("fluency", 0.0),
|
||||
completeness_score=avg.get("completeness", 0.0),
|
||||
level=grades.get("overall_level", "无"),
|
||||
raw_result=json.dumps(result, ensure_ascii=False),
|
||||
)
|
||||
|
||||
return {
|
||||
"ok": True,
|
||||
"id": record.id,
|
||||
"user": user.name,
|
||||
"language": lang,
|
||||
"average_score": avg.get("overall"),
|
||||
"level": grades.get("overall_level"),
|
||||
"count": count,
|
||||
"timestamp": record.created_at.isoformat()
|
||||
}
|
||||
|
|
@ -2,12 +2,14 @@ from typing import Literal, List
|
|||
|
||||
from fastapi import APIRouter, Depends, HTTPException, Request
|
||||
|
||||
from app.api.search_dict import service
|
||||
from app.api.search_dict.search_schemas import SearchRequest, WordSearchResponse, SearchItemFr, SearchItemJp, \
|
||||
ProverbSearchRequest
|
||||
from app.api.search_dict.service import suggest_autocomplete
|
||||
from app.api.word_comment.word_comment_schemas import CommentSet
|
||||
from app.models import DefinitionJp, CommentFr, CommentJp
|
||||
from app.models.fr import DefinitionFr
|
||||
from app.schemas.search_schemas import SearchRequest, SearchResponse, SearchItemFr, SearchItemJp
|
||||
from app.utils.all_kana import all_in_kana
|
||||
from app.utils.autocomplete import suggest_autocomplete
|
||||
from app.utils.security import get_current_user
|
||||
from app.utils.textnorm import normalize_text
|
||||
|
||||
|
|
@ -54,7 +56,7 @@ async def __get_comments(
|
|||
return commentlist
|
||||
|
||||
|
||||
@dict_search.post("/search", response_model=SearchResponse)
|
||||
@dict_search.post("/search/word", response_model=WordSearchResponse)
|
||||
async def search(request: Request, body: SearchRequest, user=Depends(get_current_user)):
|
||||
"""
|
||||
精确搜索
|
||||
|
|
@ -96,7 +98,7 @@ async def search(request: Request, body: SearchRequest, user=Depends(get_current
|
|||
eng_explanation=wc.eng_explanation,
|
||||
)
|
||||
)
|
||||
return SearchResponse(
|
||||
return WordSearchResponse(
|
||||
query=query,
|
||||
pos=pos_contents,
|
||||
contents=contents,
|
||||
|
|
@ -126,26 +128,44 @@ async def search(request: Request, body: SearchRequest, user=Depends(get_current
|
|||
example=wc.example,
|
||||
)
|
||||
)
|
||||
return SearchResponse(
|
||||
return WordSearchResponse(
|
||||
query=query,
|
||||
pos=pos_contents,
|
||||
contents=contents,
|
||||
)
|
||||
|
||||
|
||||
@dict_search.post("/search/proverb")
|
||||
async def proverb(request: Request, proverb_id: int, user=Depends(get_current_user)):
|
||||
"""
|
||||
用于法语谚语搜索
|
||||
:param request:
|
||||
:param body: 要求用户输入的内容必须为法语
|
||||
:param user:
|
||||
:return:
|
||||
"""
|
||||
content = await service.accurate_proverb(proverb_id=proverb_id)
|
||||
return content
|
||||
|
||||
|
||||
# TODO 相关度排序（转换为模糊匹配）
|
||||
# TODO 输入搜索框时反馈内容
|
||||
|
||||
@dict_search.post("/search/list")
|
||||
async def search_list(query_word: SearchRequest, user=Depends(get_current_user)):
|
||||
@dict_search.post("/search/word/list")
|
||||
async def search_word_list(query_word: SearchRequest, user=Depends(get_current_user)):
|
||||
"""
|
||||
检索时的提示接口
|
||||
:param query_word: 用户输入的内容
|
||||
:param user:
|
||||
:return: 待选列表
|
||||
"""
|
||||
print(query_word.query, query_word.language, query_word.sort, query_word.order)
|
||||
# print(query_word.query, query_word.language, query_word.sort, query_word.order)
|
||||
word_contents = await suggest_autocomplete(query=query_word)
|
||||
return {"list": word_contents}
|
||||
|
||||
#TODO 用户搜索历史
|
||||
|
||||
@dict_search.post("/search/proverb/list")
|
||||
async def search_proverb_list(query_word: ProverbSearchRequest, user=Depends(get_current_user)):
|
||||
lang: Literal['fr', 'zh'] = 'zh' if service.contains_chinese(query_word.query) else 'fr'
|
||||
suggest_proverbs = await service.suggest_proverb(query=query_word, lang=lang)
|
||||
return {"list": suggest_proverbs}
|
||||
|
|
@ -2,7 +2,6 @@ from typing import Literal, List, Union, Optional
|
|||
|
||||
from pydantic import BaseModel
|
||||
|
||||
from app.models import PosType
|
||||
from app.schemas.admin_schemas import PosEnumFr
|
||||
|
||||
|
||||
|
|
@ -12,6 +11,10 @@ class SearchRequest(BaseModel):
|
|||
sort: Literal['relevance', 'date'] = 'date'
|
||||
order: Literal['asc', 'des'] = 'des'
|
||||
|
||||
class ProverbSearchRequest(BaseModel):
|
||||
query: str
|
||||
language: Literal['fr', 'jp'] = "fr"
|
||||
|
||||
|
||||
class SearchItemJp(BaseModel):
|
||||
chi_exp: str
|
||||
|
|
@ -25,7 +28,12 @@ class SearchItemFr(BaseModel):
|
|||
example: Optional[str]
|
||||
|
||||
|
||||
class SearchResponse(BaseModel):
|
||||
class WordSearchResponse(BaseModel):
|
||||
query: str
|
||||
pos: list
|
||||
contents: Union[List[SearchItemFr], List[SearchItemJp]]
|
||||
|
||||
|
||||
class ProverbSearchResponse(BaseModel):
|
||||
proverb_text: str
|
||||
chi_exp: str
|
||||
|
|
@ -0,0 +1,208 @@
|
|||
import asyncio
|
||||
import re
|
||||
from typing import List, Tuple, Dict, Literal
|
||||
|
||||
from fastapi import HTTPException
|
||||
from tortoise import Tortoise
|
||||
from tortoise.expressions import Q
|
||||
|
||||
from app.api.search_dict.search_schemas import SearchRequest, ProverbSearchResponse, ProverbSearchRequest
|
||||
from app.models import WordlistFr, WordlistJp
|
||||
from app.models.fr import ProverbFr
|
||||
from app.utils.all_kana import all_in_kana
|
||||
from app.utils.textnorm import normalize_text
|
||||
from settings import TORTOISE_ORM
|
||||
|
||||
|
||||
def contains_chinese(text: str) -> bool:
|
||||
"""判断字符串中是否包含至少一个中文字符"""
|
||||
return bool(re.search(r'[\u4e00-\u9fff]', text))
|
||||
|
||||
|
||||
async def accurate_proverb(proverb_id: int) -> ProverbSearchResponse:
|
||||
proverb = await ProverbFr.get_or_none(id=proverb_id)
|
||||
if not proverb:
|
||||
raise HTTPException(status_code=404, detail="Proverb not found")
|
||||
return ProverbSearchResponse(
|
||||
proverb_text=proverb.proverb,
|
||||
chi_exp=proverb.chi_exp,
|
||||
)
|
||||
|
||||
|
||||
async def suggest_proverb(query: ProverbSearchRequest, lang: Literal['fr', 'zh']) -> List[Dict[str, str]]:
|
||||
"""
|
||||
对法语谚语表进行搜索建议。
|
||||
参数:
|
||||
query.query: 搜索关键词
|
||||
lang: 'fr' 或 'zh'
|
||||
逻辑:
|
||||
1. 若 lang='fr'，按谚语字段 (proverb) 搜索；
|
||||
2. 若 lang='zh'，按中文释义字段 (chi_exp) 搜索；
|
||||
3. 优先以输入开头的匹配；
|
||||
4. 其次为包含输入但不以其开头的匹配（按 freq 排序）。
|
||||
:return: [{'id': 1, 'proverb': 'xxx'}, ...]
|
||||
"""
|
||||
keyword = query.query.strip()
|
||||
results: List[Dict[str, str]] = []
|
||||
|
||||
if not keyword:
|
||||
return results
|
||||
|
||||
# ✅ 根据语言决定搜索字段
|
||||
if lang == "zh":
|
||||
startswith_field = "chi_exp__istartswith"
|
||||
contains_field = "chi_exp__icontains"
|
||||
else: # 默认法语
|
||||
startswith_field = "proverb__istartswith"
|
||||
contains_field = "proverb__icontains"
|
||||
|
||||
# ✅ 1. 开头匹配
|
||||
start_matches = await (
|
||||
ProverbFr.filter(**{startswith_field: keyword})
|
||||
.order_by("-freq")
|
||||
.limit(10)
|
||||
.values("id", "proverb", "chi_exp")
|
||||
)
|
||||
|
||||
# ✅ 2. 包含匹配（但不是开头）
|
||||
contain_matches = await (
|
||||
ProverbFr.filter(
|
||||
Q(**{contains_field: keyword}) & ~Q(**{startswith_field: keyword})
|
||||
)
|
||||
.order_by("-freq")
|
||||
.limit(10)
|
||||
.values("id", "proverb", "chi_exp")
|
||||
)
|
||||
|
||||
# ✅ 合并结果（去重并保持顺序）
|
||||
seen_ids = set()
|
||||
for row in start_matches + contain_matches:
|
||||
if row["id"] not in seen_ids:
|
||||
seen_ids.add(row["id"])
|
||||
results.append({
|
||||
"id": row["id"],
|
||||
"proverb": row["proverb"],
|
||||
"chi_exp": row["chi_exp"]
|
||||
})
|
||||
|
||||
return results
|
||||
|
||||
|
||||
async def suggest_autocomplete(query: SearchRequest, limit: int = 10):
|
||||
"""
|
||||
|
||||
:param query: 当前用户输入的内容
|
||||
:param limit: 返回列表限制长度
|
||||
:return: 联想的单词列表（非完整信息，单纯单词）
|
||||
"""
|
||||
if query.language == 'fr':
|
||||
query_word = normalize_text(query.query)
|
||||
exact = await (
|
||||
WordlistFr
|
||||
.get_or_none(search_text=query.query)
|
||||
.values("text", "freq")
|
||||
)
|
||||
if exact:
|
||||
exact_word = [(exact.get("text"), exact.get("freq"))]
|
||||
else:
|
||||
exact_word = []
|
||||
|
||||
qs_prefix = (
|
||||
WordlistFr
|
||||
.filter(Q(search_text__startswith=query_word) | Q(text__startswith=query.query))
|
||||
.exclude(search_text=query.query)
|
||||
.only("text", "freq")
|
||||
)
|
||||
prefix_objs = await qs_prefix[:limit]
|
||||
prefix: List[Tuple[str, int]] = [(o.text, o.freq) for o in prefix_objs]
|
||||
|
||||
need = max(0, limit - len(prefix))
|
||||
contains: List[Tuple[str, int]] = []
|
||||
|
||||
if need > 0:
|
||||
qs_contain = (
|
||||
WordlistFr
|
||||
.filter(Q(search_text__icontains=query_word) | Q(text__icontains=query.query))
|
||||
.exclude(Q(search_text__startswith=query_word) | Q(text__startswith=query.query) | Q(text=query.query))
|
||||
.only("text", "freq")
|
||||
.only("text", "freq")
|
||||
)
|
||||
contains_objs = await qs_contain[: need * 2]
|
||||
contains = [(o.text, o.freq) for o in contains_objs]
|
||||
|
||||
seen_text, out = set(), []
|
||||
for text, freq in list(exact_word) + list(prefix) + list(contains):
|
||||
key = text
|
||||
if key not in seen_text:
|
||||
seen_text.add(key)
|
||||
out.append((text, freq))
|
||||
if len(out) >= limit:
|
||||
break
|
||||
out = sorted(out, key=lambda w: (-w[2], len(w[0]), w[0]))
|
||||
return [text for text, _ in out]
|
||||
|
||||
else:
|
||||
query_word = all_in_kana(query.query)
|
||||
exact = await (
|
||||
WordlistJp
|
||||
.get_or_none(
|
||||
text=query.query
|
||||
)
|
||||
.only("text", "hiragana", "freq")
|
||||
)
|
||||
if exact:
|
||||
exact_word = [(exact.text, exact.hiragana, exact.freq)]
|
||||
else:
|
||||
exact_word = []
|
||||
|
||||
qs_prefix = (
|
||||
WordlistJp
|
||||
.filter(Q(hiragana__startswith=query_word) | Q(text__startswith=query.query))
|
||||
.exclude(text=query.query)
|
||||
.only("text", "hiragana", "freq")
|
||||
)
|
||||
prefix_objs = await qs_prefix[:limit]
|
||||
prefix: List[Tuple[str, str, int]] = [(o.text, o.hiragana, o.freq) for o in prefix_objs]
|
||||
|
||||
need = max(0, limit - len(prefix))
|
||||
contains: List[Tuple[str, str, int]] = []
|
||||
|
||||
if need > 0:
|
||||
qs_contain = await (
|
||||
WordlistJp
|
||||
.filter(Q(hiragana__icontains=query_word) | Q(text__icontains=query.query))
|
||||
.exclude(Q(hiragana__startswith=query_word) | Q(text__startswith=query.query) | Q(text=query.query))
|
||||
.only("text", "hiragana", "freq")
|
||||
)
|
||||
contains_objs = qs_contain[:need * 2]
|
||||
contains: List[Tuple[str, str, int]] = [(o.text, o.hiragana, o.freq) for o in contains_objs]
|
||||
|
||||
seen_text, out = set(), []
|
||||
for text, hiragana, freq in list(exact_word) + list(prefix) + list(contains):
|
||||
key = (text, hiragana)
|
||||
if key not in seen_text:
|
||||
seen_text.add(key)
|
||||
out.append((text, hiragana, freq))
|
||||
if len(out) >= limit:
|
||||
break
|
||||
out = sorted(out, key=lambda w: (-w[2], len(w[0]), w[0]))
|
||||
return [(text, hiragana) for text, hiragana, _ in out]
|
||||
|
||||
|
||||
async def __test():
|
||||
query_word: str = '棋逢'
|
||||
return await (
|
||||
suggest_proverb(
|
||||
query=ProverbSearchRequest(query=query_word),
|
||||
lang='zh'
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
async def __main():
|
||||
await Tortoise.init(config=TORTOISE_ORM)
|
||||
print(await __test())
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
asyncio.run(__main())
|
||||
|
|
@ -1,10 +1,9 @@
|
|||
import json
|
||||
import random
|
||||
from typing import Tuple, Dict
|
||||
|
||||
import redis.asyncio as redis_asyncio
|
||||
import httpx
|
||||
import random
|
||||
import json
|
||||
|
||||
import redis.asyncio as redis_asyncio
|
||||
from fastapi import APIRouter, Depends, HTTPException
|
||||
|
||||
from app.models import User
|
||||
|
|
@ -115,7 +114,7 @@ async def rate_limiter(
|
|||
raise HTTPException(status_code=429, detail=f"Too many requests")
|
||||
|
||||
|
||||
@translator_router.post('/translate', response_model=TransResponse)
|
||||
@translator_router.post('/translate', response_model=TransResponse, dependencies=[Depends(rate_limiter)])
|
||||
async def translate(
|
||||
translate_request: TransRequest,
|
||||
user=Depends(get_current_user)
|
||||
|
|
|
|||
|
|
@ -1,26 +0,0 @@
|
|||
import os
|
||||
|
||||
from fastapi import APIRouter
|
||||
|
||||
"""
|
||||
# 背景
|
||||
你是一个人工智能助手，名字叫EduChat,是一个由华东师范大学开发的教育领域大语言模型。
|
||||
# 对话主题:作文指导
|
||||
## 作文指导主题的要求：
|
||||
EduChat你需要扮演一位经验丰富的语文老师，现在需要帮助一位学生审阅作文并给出修改建议。请按照以下步骤进行：
|
||||
整体评价：先对作文的整体质量进行简要评价，指出主要优点和需要改进的方向。
|
||||
亮点分析：具体指出作文中的亮点（如结构、描写、情感表达等方面的优点）。
|
||||
具体修改建议：针对作文中的不足，从以下几个方面提出具体修改建议，并给出修改后的示例：
|
||||
语言表达：是否生动、准确？有无冗余或重复？可以如何优化？
|
||||
细节描写：是否足够具体？能否加入更多感官描写（视觉、听觉、嗅觉、触觉等）使画面更立体？
|
||||
情感表达：情感是否自然？能否更深入或升华？
|
||||
结构布局：段落衔接是否自然？开头结尾是否呼应？ （注意：每个建议点都要结合原文具体句子进行分析，并给出修改后的句子或段落作为示例）
|
||||
写作技巧提示：提供2-3条实用的写作技巧（如动态描写公式、感官交织法等），帮助学生举一反三。
|
||||
修改效果总结：简要说明按照建议修改后，作文会有哪些方面的提升（如文学性、情感层次、场景沉浸感等）。
|
||||
请用亲切、鼓励的语气进行点评，保持专业性同时让学生易于接受。
|
||||
"""
|
||||
|
||||
article_router = APIRouter()
|
||||
|
||||
ECNU_API_KEY = os.getenv("ECNU_TEACH_AI_KEY")
|
||||
|
||||
|
|
@ -64,6 +64,5 @@ def main(receiver: str, code: int = 123456):
|
|||
|
||||
if __name__ == '__main__':
|
||||
xza = "3480039769@qq.com"
|
||||
bb = "1530799205@qq.com"
|
||||
me = "GodricTan@gmail.com"
|
||||
main(xza, code=123833)
|
||||
|
|
|
|||
|
|
@ -1,5 +1,5 @@
|
|||
from . import signals
|
||||
from .base import User
|
||||
from .comments import CommentFr, CommentJp
|
||||
from .fr import WordlistFr, DefinitionFr, AttachmentFr
|
||||
from .jp import WordlistJp, DefinitionJp, AttachmentJp, PosType
|
||||
from .fr import WordlistFr, DefinitionFr, AttachmentFr, PronunciationTestFr
|
||||
from .jp import WordlistJp, DefinitionJp, AttachmentJp, PosType, PronunciationTestJp
|
||||
|
|
|
|||
|
|
@ -49,3 +49,22 @@ class Language(Model):
|
|||
id = fields.IntField(pk=True)
|
||||
name = fields.CharField(max_length=30, unique=True) # e.g. "Japanese"
|
||||
code = fields.CharField(max_length=10, unique=True) # e.g. "ja", "fr", "zh"
|
||||
|
||||
class UserTestRecord(Model):
|
||||
id = fields.IntField(pk=True)
|
||||
user = fields.ForeignKeyField("models.User", related_name="test_records")
|
||||
username = fields.CharField(max_length=20)
|
||||
|
||||
language = fields.CharField(max_length=10)
|
||||
total_sentences = fields.IntField()
|
||||
average_score = fields.FloatField()
|
||||
accuracy_score = fields.FloatField()
|
||||
fluency_score = fields.FloatField()
|
||||
completeness_score = fields.FloatField()
|
||||
level = fields.CharField(max_length=20)
|
||||
|
||||
raw_result = fields.JSONField() # 或 TextField 存 JSON 字符串
|
||||
created_at = fields.DatetimeField(auto_now_add=True)
|
||||
|
||||
class Meta:
|
||||
table = "user_test_record"
|
||||
|
|
|
|||
|
|
@ -13,6 +13,7 @@ class WordlistFr(Model):
|
|||
attachments: fields.ReverseRelation["AttachmentFr"]
|
||||
freq = fields.IntField(default=0) # 词频排序用
|
||||
search_text = fields.CharField(max_length=255, index=True) # 检索字段
|
||||
proverb = fields.ManyToManyField("models.ProverbFr", related_name="wordlists")
|
||||
|
||||
# attachment = fields.ForeignKeyField("models.Attachment", related_name="wordlists", on_delete=fields.CASCADE)
|
||||
# source = fields.CharField(max_length=20, description="<UNK>", null=True)
|
||||
|
|
@ -41,3 +42,21 @@ class DefinitionFr(Model):
|
|||
example_varification = fields.BooleanField(default=False, description="例句是否审核")
|
||||
class Meta:
|
||||
table = "definitions_fr"
|
||||
|
||||
class ProverbFr(Model):
|
||||
id = fields.IntField(pk=True)
|
||||
proverb = fields.TextField(description="法语谚语及常用表达")
|
||||
chi_exp = fields.TextField(description="中文释义")
|
||||
freq = fields.IntField(default=0)
|
||||
created_at = fields.DatetimeField(auto_now_add=True)
|
||||
updated_at = fields.DatetimeField(auto_now=True)
|
||||
|
||||
class Meta:
|
||||
table = "proverb_fr"
|
||||
|
||||
class PronunciationTestFr(Model):
|
||||
id = fields.IntField(pk=True)
|
||||
text = fields.TextField(description="朗读文段")
|
||||
|
||||
class Meta:
|
||||
table = "pronunciationtest_fr"
|
||||
|
|
@ -1,13 +1,13 @@
|
|||
from __future__ import annotations
|
||||
|
||||
from enum import Enum
|
||||
from app.schemas.admin_schemas import PosEnumJp
|
||||
from typing import Tuple, TypeVar
|
||||
|
||||
import pandas as pd
|
||||
from tortoise.exceptions import DoesNotExist, MultipleObjectsReturned
|
||||
from tortoise.models import Model
|
||||
from tortoise import fields
|
||||
from typing import Tuple, TYPE_CHECKING, TypeVar, Type, Optional
|
||||
from tortoise.exceptions import DoesNotExist
|
||||
from tortoise.models import Model
|
||||
|
||||
from app.schemas.admin_schemas import PosEnumJp
|
||||
|
||||
sheet_name_jp = "日汉释义"
|
||||
|
||||
|
|
@ -80,3 +80,10 @@ class PosType(Model):
|
|||
|
||||
class Meta:
|
||||
table = "pos_type"
|
||||
|
||||
class PronunciationTestJp(Model):
|
||||
id = fields.IntField(pk=True)
|
||||
text = fields.TextField(description="朗读文段")
|
||||
|
||||
class Meta:
|
||||
table = "pronunciationtest_jp"
|
||||
|
|
|
|||
|
|
@ -0,0 +1,13 @@
|
|||
import os
|
||||
|
||||
import ffprobe8_binaries # 或 ffprobe_binaries_only
|
||||
from imageio_ffmpeg import get_ffmpeg_exe
|
||||
from pydub import AudioSegment
|
||||
|
||||
ffprobe_path = os.path.join(os.path.dirname(ffprobe8_binaries.__file__), "bin", "ffprobe")
|
||||
|
||||
AudioSegment.converter = get_ffmpeg_exe()
|
||||
AudioSegment.ffprobe = ffprobe_path # 👈 指定 ffprobe 路径
|
||||
|
||||
print(f"[INIT] ffmpeg: {AudioSegment.converter}")
|
||||
print(f"[INIT] ffprobe: {AudioSegment.ffprobe}")
|
||||
|
|
@ -4,8 +4,8 @@ from typing import List, Literal, Tuple
|
|||
from tortoise import Tortoise
|
||||
from tortoise.expressions import Q
|
||||
|
||||
from app.api.search_dict.search_schemas import SearchRequest
|
||||
from app.models import WordlistFr, WordlistJp
|
||||
from app.schemas.search_schemas import SearchRequest
|
||||
from app.utils.all_kana import all_in_kana
|
||||
from app.utils.textnorm import normalize_text
|
||||
from settings import TORTOISE_ORM
|
||||
|
|
|
|||
9
main.py
9
main.py
|
|
@ -5,13 +5,14 @@ from fastapi import FastAPI
|
|||
from fastapi.middleware.cors import CORSMiddleware
|
||||
from tortoise.contrib.fastapi import register_tortoise
|
||||
|
||||
import app.models.signals
|
||||
import app.utils.audio_init
|
||||
from app.api.admin.router import admin_router
|
||||
from app.api.ai_assist.routes import ai_router
|
||||
from app.api.article_director.routes import article_router
|
||||
from app.api.make_comments.routes import comment_router
|
||||
from app.api.pronounciation_test.routes import pron_test_router
|
||||
from app.api.redis_test import redis_test_router
|
||||
from app.api.search import dict_search
|
||||
from app.api.search_dict.routes import dict_search
|
||||
from app.api.translator import translator_router
|
||||
from app.api.user.routes import users_router
|
||||
from app.api.word_comment.routes import word_comment_router
|
||||
|
|
@ -62,7 +63,9 @@ app.include_router(comment_router, tags=["Comment API"])
|
|||
|
||||
app.include_router(word_comment_router, tags=["Word Comment API"], prefix="/comment/word")
|
||||
|
||||
app.include_router(pron_test_router, tags=["Pron Test API"], prefix="/test")
|
||||
app.include_router(pron_test_router, tags=["Pron Test API"], prefix="/test/pron")
|
||||
|
||||
app.include_router(article_router, tags=["Article API"])
|
||||
|
||||
if __name__ == "__main__":
|
||||
uvicorn.run("main:app", host="127.0.0.1", port=8000, reload=True)
|
||||
|
|
|
|||
|
|
@ -66,3 +66,7 @@ dependencies = [
|
|||
tortoise_orm = "settings.TORTOISE_ORM"
|
||||
location = "./migrations"
|
||||
src_folder = "./."
|
||||
|
||||
[tool.uv.sources]
|
||||
[tool.uv.sources.default]
|
||||
url = "https://pypi.tuna.tsinghua.edu.cn/simple"
|
||||
|
|
|
|||
|
|
@ -1,3 +1,5 @@
|
|||
pydub
|
||||
imageio-ffmpeg
|
||||
aerich==0.9.1
|
||||
aiosqlite==0.21.0
|
||||
annotated-types==0.7.0
|
||||
|
|
@ -5,6 +7,8 @@ anyio==4.10.0
|
|||
async-timeout==5.0.1
|
||||
asyncclick==8.2.2.2
|
||||
asyncmy==0.2.10
|
||||
azure-cognitiveservices-speech==1.46.0
|
||||
azure-core==1.36.0
|
||||
bcrypt==4.3.0
|
||||
certifi==2025.8.3
|
||||
cffi==1.17.1
|
||||
|
|
@ -46,6 +50,7 @@ six==1.17.0
|
|||
sniffio==1.3.1
|
||||
starlette==0.47.2
|
||||
tortoise-orm==0.25.1
|
||||
types-pytz==2025.2.0.20250809
|
||||
typing-inspection==0.4.1
|
||||
typing_extensions==4.14.1
|
||||
tzdata==2025.2
|
||||
|
|
|
|||
|
|
@ -0,0 +1,41 @@
|
|||
import asyncio
|
||||
from pathlib import Path
|
||||
|
||||
import pandas as pd
|
||||
from tortoise import Tortoise
|
||||
|
||||
from app.models.fr import ProverbFr
|
||||
from settings import TORTOISE_ORM
|
||||
|
||||
__xlsx_name = "../DictTable_20251029.xlsx"
|
||||
__table_name = "法语谚语常用表达"
|
||||
|
||||
|
||||
class FrProverb:
|
||||
def __init__(self, __xlsx_name, __table_name):
|
||||
self.__xlsx_name = __xlsx_name
|
||||
self.__table_name = __table_name
|
||||
|
||||
async def get_proverb(self) -> None:
|
||||
df = pd.read_excel(Path(self.__xlsx_name), sheet_name=self.__table_name)
|
||||
df.columns = [col.strip() for col in df.columns]
|
||||
|
||||
for row in df.itertuples():
|
||||
proverb = str(row.法语谚语常用表达).strip()
|
||||
chi_exp = str(row.中文释义).strip()
|
||||
|
||||
cls_proverb, created = await ProverbFr.get_or_create(proverb=proverb, chi_exp=chi_exp)
|
||||
if not created:
|
||||
print(f"{proverb} 已存在！位于第{row.index}行")
|
||||
|
||||
async def build_connection(self):
|
||||
pass
|
||||
|
||||
|
||||
async def main():
|
||||
await Tortoise.init(config=TORTOISE_ORM)
|
||||
proverb = FrProverb(__xlsx_name, __table_name)
|
||||
await proverb.get_proverb()
|
||||
|
||||
if __name__ == '__main__':
|
||||
asyncio.run(main())
|
||||
|
|
@ -1,6 +1,5 @@
|
|||
import asyncio
|
||||
from pathlib import Path
|
||||
from tkinter.scrolledtext import example
|
||||
|
||||
import pandas as pd
|
||||
from tortoise import Tortoise, connections
|
||||
|
|
|
|||
25
settings.py
25
settings.py
|
|
@ -1,5 +1,11 @@
|
|||
from pathlib import Path
|
||||
|
||||
from pydantic.v1 import BaseSettings
|
||||
|
||||
# 计算项目根目录：假设 settings.py 位于 dict_server/settings.py
|
||||
BASE_DIR = Path(__file__).resolve().parent
|
||||
ROOT_DIR = BASE_DIR # 如果 settings.py 就在根目录，否则改成 BASE_DIR.parent
|
||||
|
||||
TORTOISE_ORM = {
|
||||
'connections': {
|
||||
"default": "mysql://local_admin:enterprise@127.0.0.1:3306/dict",
|
||||
|
|
@ -51,8 +57,25 @@ class Settings(BaseSettings):
|
|||
BAIDU_APPKEY: str
|
||||
REDIS_URL: str
|
||||
|
||||
AES_SECRET_KEY: str
|
||||
|
||||
SMTP_HOST: str
|
||||
SMTP_PORT: int
|
||||
SMTP_USER: str
|
||||
SMTP_PASS: str
|
||||
SMTP_SENDER_NAME: str
|
||||
|
||||
RESET_SECRET_KEY: str
|
||||
|
||||
AI_ASSIST_KEY: str
|
||||
|
||||
ECNU_TEACH_AI_KEY: str
|
||||
|
||||
AZURE_SUBSCRIPTION_KEY: str
|
||||
|
||||
class Config:
|
||||
env_file = '.env'
|
||||
env_file = ROOT_DIR / '.env'
|
||||
case_sensitive = False
|
||||
|
||||
|
||||
settings = Settings()
|
||||
|
|
|
|||
Loading…
Reference in New Issue