feat:优化NanoImage

This commit is contained in:
2025-12-30 13:46:56 +08:00
parent b44d1589d1
commit 58f17e4497
2 changed files with 99 additions and 39 deletions

Binary file not shown.

After

Width:  |  Height:  |  Size: 593 KiB

View File

@@ -10,6 +10,7 @@ import tomllib
import httpx
import uuid
import base64
import re
from pathlib import Path
from datetime import datetime
from typing import List, Optional
@@ -93,58 +94,117 @@ class NanoImage(PluginBase):
async with client.stream("POST", url, json=payload, headers=headers) as response:
logger.debug(f"收到响应状态码: {response.status_code}")
if response.status_code == 200:
# 处理流式响应
content_type = (response.headers.get("content-type") or "").lower()
is_sse = "text/event-stream" in content_type or "event-stream" in content_type
# 处理流式响应SSE
image_url = None
image_base64 = None
full_content = ""
async for line in response.aiter_lines():
if line.startswith("data: "):
data_str = line[6:]
if data_str == "[DONE]":
break
try:
import json
data = json.loads(data_str)
if "choices" in data and data["choices"]:
delta = data["choices"][0].get("delta", {})
# 方式1: 从 delta.images 中提取(新格式)
images = delta.get("images", [])
if images and len(images) > 0:
img_data = images[0].get("image_url", {}).get("url", "")
if img_data:
if img_data.startswith("data:image"):
# base64 格式
image_base64 = img_data
logger.info(f"从 delta.images 提取到 base64 图片")
elif img_data.startswith("http"):
image_url = img_data
logger.info(f"从 delta.images 提取到图片URL: {image_url}")
# 方式2: 从 content 中提取(旧格式)
content = delta.get("content", "")
if content:
full_content += content
if "http" in content:
import re
urls = re.findall(r'https?://[^\s\)\]"\']+', content)
if urls:
image_url = urls[0].rstrip("'\"")
logger.info(f"从 content 提取到图片URL: {image_url}")
except Exception as e:
logger.warning(f"解析响应数据失败: {e}")
if is_sse:
async for line in response.aiter_lines():
if not line:
continue
if line.startswith("data:"):
data_str = line[5:].lstrip()
if data_str == "[DONE]":
break
try:
import json
data = json.loads(data_str)
if "choices" in data and data["choices"]:
delta = data["choices"][0].get("delta", {})
# 方式1: 从 delta.images 中提取(新格式)
images = delta.get("images", [])
if images and len(images) > 0:
img_data = images[0].get("image_url", {}).get("url", "")
if img_data:
if img_data.startswith("data:image"):
# base64 格式
image_base64 = img_data
logger.info("从 delta.images 提取到 base64 图片")
elif img_data.startswith("http"):
image_url = img_data
logger.info(f"从 delta.images 提取到图片URL: {image_url}")
# 方式2: 从 content 中提取(旧格式)
content = delta.get("content", "")
if content:
full_content += content
if "http" in content:
urls = re.findall(r'https?://[^\s\)\]"\']+', content)
if urls:
image_url = urls[0].rstrip("'\"")
logger.info(f"从 content 提取到图片URL: {image_url}")
except Exception as e:
logger.warning(f"解析响应数据失败: {e}")
continue
else:
# 非流式application/json某些网关即使传了 stream=true 也会返回完整 JSON
raw = await response.aread()
try:
import json
data = json.loads(raw.decode("utf-8", errors="ignore"))
except Exception as e:
logger.error(f"解析 JSON 响应失败: {type(e).__name__}: {e}")
data = None
if isinstance(data, dict):
# 1) 标准 images endpoint 兼容:{"data":[{"url":...}|{"b64_json":...}]}
items = data.get("data")
if isinstance(items, list) and items:
first = items[0] if isinstance(items[0], dict) else {}
if isinstance(first, dict):
b64_json = first.get("b64_json")
if b64_json:
image_base64 = b64_json
logger.info("从 data[0].b64_json 提取到 base64 图片")
else:
u = first.get("url") or ""
if isinstance(u, str) and u:
image_url = u
logger.info(f"从 data[0].url 提取到图片URL: {image_url}")
# 2) chat.completion 兼容choices[0].message.images[0].image_url.url
if not image_url and not image_base64:
try:
choices = data.get("choices") or []
if choices:
msg = (choices[0].get("message") or {}) if isinstance(choices[0], dict) else {}
images = msg.get("images") or []
if isinstance(images, list) and images:
img0 = images[0] if isinstance(images[0], dict) else {}
if isinstance(img0, dict):
img_data = (
(img0.get("image_url") or {}).get("url")
if isinstance(img0.get("image_url"), dict)
else img0.get("url")
)
if isinstance(img_data, str) and img_data:
if img_data.startswith("data:image"):
image_base64 = img_data
logger.info("从 message.images 提取到 base64 图片")
elif img_data.startswith("http"):
image_url = img_data
logger.info(f"从 message.images 提取到图片URL: {image_url}")
except Exception:
pass
# 如果没有从流中提取到URL尝试从完整内容中提取
if not image_url and not image_base64 and full_content:
import re
urls = re.findall(r'https?://[^\s\)\]"\']+', full_content)
if urls:
image_url = urls[0].rstrip("'\"")
logger.info(f"从完整内容提取到图片URL: {image_url}")
if not image_url and not image_base64:
logger.error(f"未能提取到图片,完整响应: {full_content[:500]}")
# 避免把 base64 打到日志里:只输出裁剪后的概要
if full_content:
logger.error(f"未能提取到图片,完整响应(截断): {full_content[:500]}")
else:
# 非SSE时 full_content 可能为空,补充输出 content-type 便于定位
logger.error(f"未能提取到图片content-type={content_type or 'unknown'}")
# 处理 base64 图片
if image_base64: