File size: 3,197 Bytes
60fde3b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
import base64
import mimetypes
import os
import re
from PIL import Image

def encode_image_to_base64(image_path):
    """读取图片并转换为 data:image/jpeg;base64,xxx 格式"""
    if not os.path.exists(image_path):
        raise FileNotFoundError(f"Image not found: {image_path}")
        
    mime_type, _ = mimetypes.guess_type(image_path)
    if not mime_type:
        mime_type = "image/jpeg" # 默认 fallback
        
    with open(image_path, "rb") as image_file:
        encoded_string = base64.b64encode(image_file.read()).decode('utf-8')
        
    return f"data:{mime_type};base64,{encoded_string}"

def parse_multimodal_text(text: str) -> list:
    """
    【API 模式】解析文本,图片转 Base64。
    """
    pattern = r"<image_start>\[(.*?)\]<image_end>"
    content_list = []
    last_end = 0
    
    for match in re.finditer(pattern, text):
        start, end = match.span()
        image_path = match.group(1)
        
        if start > last_end:
            text_part = text[last_end:start]
            if text_part:
                content_list.append({"type": "text", "text": text_part})
        
        try:
            base64_url = encode_image_to_base64(image_path)
            content_list.append({
                "type": "image_url", 
                "image_url": {"url": base64_url}
            })
        except Exception as e:
            print(f"Error encoding image {image_path}: {e}")
            content_list.append({"type": "text", "text": f"[Image Missing: {image_path}]"})

        last_end = end
        
    if last_end < len(text):
        content_list.append({"type": "text", "text": text[last_end:]})
        
    return content_list

def parse_multimodal_text_for_vllm(text: str) -> list:
    """
    【本地 vLLM 模式】解析文本,图片直接加载为 PIL 对象。
    避免 Base64 编解码的开销。
    """
    pattern = r"<image_start>\[(.*?)\]<image_end>"
    content_list = []
    last_end = 0
    
    for match in re.finditer(pattern, text):
        start, end = match.span()
        image_path = match.group(1)
        
        # 1. 文本部分
        if start > last_end:
            text_part = text[last_end:start]
            if text_part:
                content_list.append({"type": "text", "text": text_part})
        
        # 2. 图片部分 (直接加载 PIL)
        try:
            if os.path.exists(image_path):
                image = Image.open(image_path).convert("RGB")
                # 使用自定义类型 "pil_image" 区别于 API 的 "image_url"
                content_list.append({
                    "type": "pil_image", 
                    "image": image
                })
            else:
                print(f"Image not found: {image_path}")
                content_list.append({"type": "text", "text": f"[Image Missing: {image_path}]"})
        except Exception as e:
            print(f"Error loading image {image_path}: {e}")
            content_list.append({"type": "text", "text": f"[Image Error]"})

        last_end = end
        
    if last_end < len(text):
        content_list.append({"type": "text", "text": text[last_end:]})
        
    return content_list