weichenandCursor
cfde38de31
Why: Uploaded images never reached the model, so a text-only chat model could be asked to read a picture it cannot see. Evidence: go test ./internal/... passed locally before this commit. Image-link rejection is covered by vision package tests. Impact: Current-turn attachments and fetched image links become image input. History keeps a text placeholder. Image turns swap to the configured vision model when the chat model is not in the capable list, and fallbacks are filtered the same way. Refs: Agent-System-Dev feature/four-model-vision Co-authored-by: Cursor <cursoragent@cursor.com>
19 lines
768 B
Go
19 lines
768 B
Go
// Package vision 负责「这一轮要不要把图发给模型」：能力清单、图片字节预算、
|
||
// 以及用户粘贴的图片链接的受控下载。模型名和端点仍由 runtime 的路由解析。
|
||
package vision
|
||
|
||
// 默认可看图清单与兜底模型。网关路由变化时改设置 models.vision.capable，
|
||
// 不必改代码。空设置回落到这里。
|
||
const (
|
||
DefaultVisionModel = "qwen3.8-flash"
|
||
DefaultVisionCapable = "deepseek-v4.1-flash,qwen3.8-flash,ZHIPU/GLM-5.3-FlashX,ZHIPU/GLM-5.3-Flash"
|
||
)
|
||
|
||
// 单图与单轮上限。按网关实测的小图请求体留出余量：base64 约膨胀 4/3，
|
||
// 8MB 原图约 11MB 请求体，单轮合计 20MB 原图。
|
||
const (
|
||
MaxImageBytes = 8 << 20
|
||
MaxImages = 8
|
||
MaxTotalBytes = 20 << 20
|
||
)
|