503 lines
15 KiB
Markdown
503 lines
15 KiB
Markdown
## 外部接口
|
||
|
||
### 1. 分割试卷API:
|
||
|
||
请求:
|
||
|
||
```bash
|
||
curl --location --request POST 'https://ekt1.suzhou.edu.cn/autocut/detect' \
|
||
--header 'Accept: */*' \
|
||
--header 'Host: ekt1.suzhou.edu.cn' \
|
||
--header 'Connection: keep-alive' \
|
||
--header 'Content-Type: multipart/form-data; boundary=--------------------------687058231513459413708701' \
|
||
--form 'image=@"/Users/robertmaxwell/code/zstp/AutoCut/pic.png"' \
|
||
--form 'enable_error_detection="false"'
|
||
```
|
||
|
||
返回:
|
||
|
||
```json
|
||
{
|
||
"data": {
|
||
"crop_results": {
|
||
"cropped_count": 14,
|
||
"cropped_images": [
|
||
{
|
||
"bbox": [
|
||
970,
|
||
69,
|
||
1909,
|
||
246
|
||
],
|
||
"filename": "question_001.jpg",
|
||
"question_id": 1,
|
||
"vertical_id": 1
|
||
},
|
||
{
|
||
"bbox": [
|
||
41,
|
||
107,
|
||
970,
|
||
253
|
||
],
|
||
"filename": "question_002.jpg",
|
||
"question_id": 2,
|
||
"vertical_id": 2
|
||
},
|
||
{
|
||
"bbox": [
|
||
970,
|
||
245,
|
||
1909,
|
||
479
|
||
],
|
||
"filename": "question_003.jpg",
|
||
"question_id": 3,
|
||
"vertical_id": 3
|
||
},
|
||
{
|
||
"bbox": [
|
||
41,
|
||
254,
|
||
970,
|
||
394
|
||
],
|
||
"filename": "question_004.jpg",
|
||
"question_id": 4,
|
||
"vertical_id": 4
|
||
},
|
||
{
|
||
"bbox": [
|
||
41,
|
||
395,
|
||
970,
|
||
506
|
||
],
|
||
"filename": "question_005.jpg",
|
||
"question_id": 5,
|
||
"vertical_id": 5
|
||
},
|
||
{
|
||
"bbox": [
|
||
970,
|
||
480,
|
||
1909,
|
||
661
|
||
],
|
||
"filename": "question_006.jpg",
|
||
"question_id": 6,
|
||
"vertical_id": 6
|
||
},
|
||
{
|
||
"bbox": [
|
||
41,
|
||
507,
|
||
970,
|
||
655
|
||
],
|
||
"filename": "question_007.jpg",
|
||
"question_id": 7,
|
||
"vertical_id": 7
|
||
},
|
||
{
|
||
"bbox": [
|
||
41,
|
||
656,
|
||
970,
|
||
820
|
||
],
|
||
"filename": "question_008.jpg",
|
||
"question_id": 8,
|
||
"vertical_id": 8
|
||
},
|
||
{
|
||
"bbox": [
|
||
970,
|
||
662,
|
||
1909,
|
||
904
|
||
],
|
||
"filename": "question_009.jpg",
|
||
"question_id": 9,
|
||
"vertical_id": 9
|
||
},
|
||
{
|
||
"bbox": [
|
||
41,
|
||
821,
|
||
970,
|
||
968
|
||
],
|
||
"filename": "question_010.jpg",
|
||
"question_id": 10,
|
||
"vertical_id": 10
|
||
},
|
||
{
|
||
"bbox": [
|
||
970,
|
||
905,
|
||
1909,
|
||
1183
|
||
],
|
||
"filename": "question_011.jpg",
|
||
"question_id": 11,
|
||
"vertical_id": 11
|
||
},
|
||
{
|
||
"bbox": [
|
||
41,
|
||
969,
|
||
970,
|
||
1111
|
||
],
|
||
"filename": "question_012.jpg",
|
||
"question_id": 12,
|
||
"vertical_id": 12
|
||
},
|
||
{
|
||
"bbox": [
|
||
41,
|
||
1112,
|
||
970,
|
||
1364
|
||
],
|
||
"filename": "question_013.jpg",
|
||
"question_id": 13,
|
||
"vertical_id": 13
|
||
},
|
||
{
|
||
"bbox": [
|
||
970,
|
||
1184,
|
||
1909,
|
||
1364
|
||
],
|
||
"filename": "question_014.jpg",
|
||
"question_id": 14,
|
||
"vertical_id": 14
|
||
}
|
||
],
|
||
"mode": "all_questions",
|
||
"zip_url": "/download/20251125_082929_f1975f5c"
|
||
},
|
||
"error_count": 8,
|
||
"error_details": [
|
||
{
|
||
"error_boxes": [
|
||
{
|
||
"bbox": [
|
||
1772,
|
||
678,
|
||
44,
|
||
44
|
||
],
|
||
"class_name": "cuo",
|
||
"confidence": 0.7618222236633301
|
||
}
|
||
],
|
||
"match_method": "中心点包含",
|
||
"question_box": {
|
||
"bbox": [
|
||
970,
|
||
662,
|
||
939,
|
||
242
|
||
]
|
||
}
|
||
},
|
||
{
|
||
"error_boxes": [
|
||
{
|
||
"bbox": [
|
||
1209,
|
||
163,
|
||
31,
|
||
32
|
||
],
|
||
"class_name": "cuo",
|
||
"confidence": 0.7555574178695679
|
||
}
|
||
],
|
||
"match_method": "中心点包含",
|
||
"question_box": {
|
||
"bbox": [
|
||
970,
|
||
69,
|
||
939,
|
||
177
|
||
]
|
||
}
|
||
},
|
||
{
|
||
"error_boxes": [
|
||
{
|
||
"bbox": [
|
||
1225,
|
||
533,
|
||
46,
|
||
44
|
||
],
|
||
"class_name": "cuo",
|
||
"confidence": 0.7304249405860901
|
||
}
|
||
],
|
||
"match_method": "中心点包含",
|
||
"question_box": {
|
||
"bbox": [
|
||
970,
|
||
480,
|
||
939,
|
||
181
|
||
]
|
||
}
|
||
},
|
||
{
|
||
"error_boxes": [
|
||
{
|
||
"bbox": [
|
||
797,
|
||
827,
|
||
40,
|
||
42
|
||
],
|
||
"class_name": "cuo",
|
||
"confidence": 0.7287607192993164
|
||
}
|
||
],
|
||
"match_method": "中心点包含",
|
||
"question_box": {
|
||
"bbox": [
|
||
41,
|
||
821,
|
||
929,
|
||
147
|
||
]
|
||
}
|
||
},
|
||
{
|
||
"error_boxes": [
|
||
{
|
||
"bbox": [
|
||
1299,
|
||
363,
|
||
33,
|
||
33
|
||
],
|
||
"class_name": "cuo",
|
||
"confidence": 0.7262661457061768
|
||
}
|
||
],
|
||
"match_method": "中心点包含",
|
||
"question_box": {
|
||
"bbox": [
|
||
970,
|
||
245,
|
||
939,
|
||
234
|
||
]
|
||
}
|
||
},
|
||
{
|
||
"error_boxes": [
|
||
{
|
||
"bbox": [
|
||
1226,
|
||
1028,
|
||
36,
|
||
36
|
||
],
|
||
"class_name": "cuo",
|
||
"confidence": 0.7171130776405334
|
||
}
|
||
],
|
||
"match_method": "中心点包含",
|
||
"question_box": {
|
||
"bbox": [
|
||
970,
|
||
905,
|
||
939,
|
||
278
|
||
]
|
||
}
|
||
},
|
||
{
|
||
"error_boxes": [
|
||
{
|
||
"bbox": [
|
||
814,
|
||
1116,
|
||
36,
|
||
37
|
||
],
|
||
"class_name": "cuo",
|
||
"confidence": 0.7121084332466125
|
||
}
|
||
],
|
||
"match_method": "中心点包含",
|
||
"question_box": {
|
||
"bbox": [
|
||
41,
|
||
1112,
|
||
929,
|
||
252
|
||
]
|
||
}
|
||
},
|
||
{
|
||
"error_boxes": [
|
||
{
|
||
"bbox": [
|
||
793,
|
||
127,
|
||
39,
|
||
30
|
||
],
|
||
"class_name": "xie",
|
||
"confidence": 0.5237091183662415
|
||
}
|
||
],
|
||
"match_method": "中心点包含",
|
||
"question_box": {
|
||
"bbox": [
|
||
41,
|
||
107,
|
||
929,
|
||
146
|
||
]
|
||
}
|
||
}
|
||
],
|
||
"error_symbol_count": 8,
|
||
"question_count": 14,
|
||
"session_id": "20251125_082929_f1975f5c",
|
||
"visualization_images": {
|
||
"error_symbols": "/visualizations/20251125_082929_f1975f5c/error_symbols.jpg",
|
||
"matched_errors": "/visualizations/20251125_082929_f1975f5c/matched_errors.jpg",
|
||
"questions": "/visualizations/20251125_082929_f1975f5c/questions.jpg"
|
||
}
|
||
},
|
||
"success": true
|
||
}
|
||
```
|
||
|
||
被切割后的图片拼接方式:
|
||
|
||
`https://ekt1.suzhou.edu.cn/autocut/files/<session_id>/<filename>`
|
||
|
||
如上图,则被切割了14个图片:
|
||
|
||
```
|
||
|
||
https://ekt1.suzhou.edu.cn/autocut/files/20251125_082929_f1975f5c/question_001.jpg
|
||
https://ekt1.suzhou.edu.cn/autocut/files/20251125_082929_f1975f5c/question_002.jpg
|
||
https://ekt1.suzhou.edu.cn/autocut/files/20251125_082929_f1975f5c/question_003.jpg
|
||
…………
|
||
…………
|
||
https://ekt1.suzhou.edu.cn/autocut/files/20251125_082929_f1975f5c/question_014.jpg
|
||
```
|
||
|
||
question_id 也需要记录。其他字段忽略.
|
||
|
||
返回错误包括:只支持图片格式:
|
||
{
|
||
"message": "不支持的文件格式",
|
||
"success": false
|
||
}
|
||
|
||
字段缺失或错误:
|
||
{
|
||
"message": "没有上传图片文件",
|
||
"success": false
|
||
}
|
||
|
||
### 2. ocr api：
|
||
|
||
输入：上面的切割试卷api所生成的url；
|
||
输出：markdown形式的带有数学公式的题目文本内容。
|
||
大模型api地址：https://ekt1.suzhou.edu.cn/llm/v1/chat/completions
|
||
openai兼容格式
|
||
模型名称：qwen-vl，是一个VL模型。
|
||
上传需要先把url解析成base64。模型不能直接识别url。
|
||
设定温度为0；
|
||
设定提示词是：
|
||
```
|
||
你是一名数学文字识别专家。你的任务是，将图片中的文字，严格转换为markdown格式的文字；如图中有数学公式，你必须使用$或者$$进行包裹，生成markdown可解析的latex格式的数学公式；你无需输出其他任何解释性文字，只需严格输出图片中的文字。
|
||
```
|
||
输入示例：
|
||
|
||
`POST: https://ekt1.suzhou.edu.cn/llm/v1/chat/completions`
|
||
`HEADER: Authonrization: Bearer sk-PqY5sI1od5Hbl6hVsniLDp7G0DOOBlUzE8E22p8HdVSTM5Rs`
|
||
|
||
|
||
```json
|
||
{
|
||
"model": "qwen-vl",
|
||
"messages": [
|
||
{
|
||
"role": "system",
|
||
"content": "你是一名数学图片文字识别专家。"
|
||
},
|
||
{
|
||
"role": "user",
|
||
"content": [
|
||
{
|
||
"type": "text",
|
||
"text": "你的任务是，将图片中的文字，严格转换为markdown格式的文字；如图中有数学公式，你必须使用$或者$$进行包裹，生成markdown可解析的latex格式的数学公式；你无需输出其他任何解释性文字，只需严格输出图片中的文字。"
|
||
},
|
||
{
|
||
"type": "image_url",
|
||
"image_url": {
|
||
"url": "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAABfYAAAGCCAIAAAAt4wW+AAAACXBIWXMAAAsTAAALEwEAm==CC"
|
||
}
|
||
}
|
||
]
|
||
}
|
||
],
|
||
"temperature": 0
|
||
}
|
||
```
|
||
|
||
|
||
输出格式：
|
||
```json
|
||
{
|
||
"id": "chatcmpl-08cc11398014472fa86bd8cbdb2ded8b",
|
||
"object": "chat.completion",
|
||
"created": 1764146440,
|
||
"model": "qwen-vl",
|
||
"choices": [
|
||
{
|
||
"index": 0,
|
||
"message": {
|
||
"role": "assistant",
|
||
"content": "12. 已知集合 $A=\\{1, 2, 4, 8\\}$, $B=\\{z|z=xy, x \\in A, y \\in A\\}$, 则下列四个结论正确的是 ( $\\not A B$ )\n\nA. $A \\nsubseteq B$\n\nB. $\\{1, 4, 16, 64\\} \\subseteq B$\n\nC. 集合 $B$ 中元素的个数为 8\n\nD. 集合 $B$ 的真子集的个数为 127",
|
||
"refusal": null,
|
||
"annotations": null,
|
||
"audio": null,
|
||
"function_call": null,
|
||
"tool_calls": [],
|
||
"reasoning_content": null
|
||
},
|
||
"logprobs": null,
|
||
"finish_reason": "stop",
|
||
"stop_reason": null
|
||
}
|
||
],
|
||
"service_tier": null,
|
||
"system_fingerprint": null,
|
||
"usage": {
|
||
"prompt_tokens": 841,
|
||
"total_tokens": 972,
|
||
"completion_tokens": 131,
|
||
"prompt_tokens_details": null
|
||
},
|
||
"prompt_logprobs": null,
|
||
"kv_transfer_params": null
|
||
}
|
||
|
||
```
|
||
|
||
"content": 这里面的文字就是ocr的结果。
|
||
|
||
### 3. 题目分析api：
|