chiyi/image2api
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

加回 nano-banana-2 与 nano-banana-pro 并存

Browse Source
This commit is contained in:
chiyi
2026-07-10 12:02:14 +08:00
parent 2afb4e7857
commit 8d792e66e0
3 changed files with 49 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files
backend/internal/provider/runway/image.go
+24 -13
View File
@@ -13,14 +13,15 @@ import (
"github.com/google/uuid"
)
// GenerateImage runs the Runway "Nano Banana Pro" (workflow_gemini_image /
// gemini-3-pro-image-preview) text/image-to-image pipeline: upload each
// reference image (DATASET + DATASET_PREVIEW → dataset) to obtain its
// {assetId, url}, create a gemini image task and poll it to completion, then
// download the rendered PNG. teamID is the workspace id; if empty it's derived
// from the token. imageSize is the "1K"/"2K"/"4K" tier (Pro honors 2K/4K).
// refs may be empty (pure text-to-image).
func (c *Client) GenerateImage(ctx context.Context, token, teamID, prompt, imageSize string, refs [][]byte) ([]byte, map[string]any, error) {
// GenerateImage runs a Runway gemini image text/image-to-image pipeline:
// upload each reference image (DATASET + DATASET_PREVIEW → dataset) to obtain
// its {assetId, url}, create a gemini image task and poll it to completion,
// then download the rendered PNG. modelID selects the variant: "nano-banana-2"
// → gemini_3_1_flash_image / gemini-3.1-flash-image-preview (with aspect_ratio),
// anything else → workflow_gemini_image / gemini-3-pro-image-preview. imageSize
// is the "1K"/"2K"/"4K" tier. teamID is the workspace id; if empty it's derived
// from the token. refs may be empty (pure text-to-image).
func (c *Client) GenerateImage(ctx context.Context, token, teamID, modelID, prompt, aspectRatio, imageSize string, refs [][]byte) ([]byte, map[string]any, error) {
token = strings.TrimSpace(strings.TrimPrefix(token, "Bearer "))
if token == "" {
return nil, nil, ErrAuth
@@ -63,7 +64,7 @@ func (c *Client) GenerateImage(ctx context.Context, token, teamID, prompt, image
})
}
taskID, err := c.createImageTask(ctx, submitClient, token, teamID, prompt, imageSize, refImages)
taskID, err := c.createImageTask(ctx, submitClient, token, teamID, modelID, prompt, aspectRatio, imageSize, refImages)
if err != nil {
return nil, nil, err
}
@@ -108,9 +109,10 @@ func (c *Client) uploadReference(ctx context.Context, client tlsclient.HttpClien
return assetID, refURL, nil
}
// createImageTask creates a workflow_gemini_image (Nano Banana Pro) task and
// returns its id.
func (c *Client) createImageTask(ctx context.Context, client tlsclient.HttpClient, token, teamID, prompt, imageSize string, refImages []map[string]any) (string, error) {
// createImageTask creates a gemini image task (workflow_gemini_image for Pro,
// gemini_3_1_flash_image for Nano Banana 2) and returns its id.
func (c *Client) createImageTask(ctx context.Context, client tlsclient.HttpClient, token, teamID, modelID, prompt, aspectRatio, imageSize string, refImages []map[string]any) (string, error) {
taskType := "workflow_gemini_image"
opts := map[string]any{
"name": "Nano Banana Pro - " + prompt,
"text_prompt": prompt,
@@ -120,11 +122,20 @@ func (c *Client) createImageTask(ctx context.Context, client tlsclient.HttpClien
"exploreMode": false,
"creationSource": "tool-mode",
}
if modelID == "nano-banana-2" {
taskType = "gemini_3_1_flash_image"
opts["name"] = "Nano Banana 2 - " + prompt
opts["model"] = "gemini-3.1-flash-image-preview"
if strings.TrimSpace(aspectRatio) == "" {
aspectRatio = "16:9"
}
opts["aspect_ratio"] = aspectRatio
}
if len(refImages) > 0 {
opts["reference_images"] = refImages
}
res, err := c.submitTask(ctx, client, token, teamID, map[string]any{
"taskType": "workflow_gemini_image",
"taskType": taskType,
"options": opts,
"asTeamId": jsonNumberOrString(teamID),
"sessionId": uuid.NewString(),