auto-save 2026-05-13 21:07 (~4)

api/main.py

@@ -220,12 +220,30 @@ def public_api_base() -> str:
     return (LLM_BASE_URL or "https://api.openai.com/v1").rstrip("/")
 
 
+def video_uses_poe() -> bool:
+    if VIDEO_API_BASE_URL:
+        return VIDEO_API_BASE_URL.rstrip("/") == POE_API_BASE_URL.rstrip("/")
+    return bool(POE_API_KEY)
+
+
+def video_uses_ark() -> bool:
+    return "ark.cn-beijing.volces.com" in video_api_base()
+
+
 def video_api_base() -> str:
-    return (VIDEO_API_BASE_URL or LLM_BASE_URL or "https://api.openai.com/v1").rstrip("/")
+    if VIDEO_API_BASE_URL:
+        return VIDEO_API_BASE_URL.rstrip("/")
+    if POE_API_KEY:
+        return POE_API_BASE_URL.rstrip("/")
+    return (LLM_BASE_URL or "https://api.openai.com/v1").rstrip("/")
 
 
 def video_api_key() -> str:
-    return VIDEO_API_KEY or LLM_API_KEY
+    if VIDEO_API_KEY:
+        return VIDEO_API_KEY
+    if video_uses_poe():
+        return POE_API_KEY
+    return LLM_API_KEY
 
 
 def video_path(template: str, **values: str) -> str:
@@ -235,7 +253,7 @@ def video_path(template: str, **values: str) -> str:
 
 def ensure_video_api_configured() -> None:
     if not video_api_key():
-        raise HTTPException(503, "VIDEO_API_KEY 或 LLM_API_KEY 未配置，无法调用生视频 API")
+        raise HTTPException(503, "POE_API_KEY、VIDEO_API_KEY 或 LLM_API_KEY 未配置，无法调用生视频 API")
 
 
 def storyboard_ref_path(job_id: str, ref: dict | None) -> Path | None:
@@ -805,7 +823,8 @@ def health() -> dict:
             "rewrite": REWRITE_MODEL,
             "video": VIDEO_MODEL,
             "video_aliases": VIDEO_MODEL_ALIASES,
-            "video_base_url": video_api_base() if VIDEO_API_BASE_URL else "",
+            "video_provider": "poe" if video_uses_poe() else ("ark" if video_uses_ark() else "custom"),
+            "video_base_url": video_api_base(),
             "video_configured": bool(video_api_key()),
             "video_create_paths": VIDEO_CREATE_PATHS,
         },
@@ -1629,6 +1648,10 @@ class GenerateStoryboardVideoReq(BaseModel):
 
 
 def video_seconds(duration: float) -> str:
+    if video_uses_ark():
+        if duration <= 0:
+            return "5"
+        return str(max(4, min(15, round(duration))))
     if duration <= 6:
         return "4"
     if duration <= 10:
@@ -1683,6 +1706,12 @@ def video_url_from_response(data: dict) -> str:
             v = output.get(key)
             if isinstance(v, str) and v:
                 return v
+    content = data.get("content")
+    if isinstance(content, dict):
+        for key in ("video_url", "url", "download_url", "file_url"):
+            v = content.get(key)
+            if isinstance(v, str) and v:
+                return v
     return ""
 
 
@@ -1696,6 +1725,64 @@ def download_generated_video(client, base: str, headers: dict, provider_id: str,
     out_mp4.write_bytes(r.content)
 
 
+def size_to_video_ratio(size: str) -> str:
+    try:
+        w, h = [int(x) for x in size.lower().replace(" ", "").split("x", 1)]
+    except Exception:
+        return "9:16"
+    if w <= 0 or h <= 0:
+        return "9:16"
+    ratio = w / h
+    known = {
+        "16:9": 16 / 9,
+        "9:16": 9 / 16,
+        "1:1": 1,
+        "4:3": 4 / 3,
+        "3:4": 3 / 4,
+        "21:9": 21 / 9,
+    }
+    return min(known, key=lambda key: abs(known[key] - ratio))
+
+
+def ark_reference_data_url(ref_img: Path) -> str:
+    mime = "image/png" if ref_img.suffix.lower() == ".png" else "image/jpeg"
+    return f"data:{mime};base64,{base64.b64encode(ref_img.read_bytes()).decode('ascii')}"
+
+
+def submit_video_create(client, url: str, headers: dict, ref_img: Path, payload: dict):
+    if video_uses_ark():
+        data = {
+            "model": payload["model"],
+            "content": [
+                {"type": "text", "text": payload["prompt"]},
+                {
+                    "type": "image_url",
+                    "image_url": {"url": ark_reference_data_url(ref_img)},
+                    "role": "first_frame",
+                },
+            ],
+            "ratio": size_to_video_ratio(str(payload.get("size", ""))),
+            "duration": int(float(str(payload.get(VIDEO_DURATION_FIELD, 5)))),
+            "watermark": False,
+            "resolution": "720p",
+        }
+        return client.post(url, headers={**headers, "Content-Type": "application/json"}, json=data)
+
+    if video_uses_poe():
+        data = dict(payload)
+        data[VIDEO_DURATION_FIELD] = int(float(str(data.get(VIDEO_DURATION_FIELD, 4))))
+        data["input_image"] = base64.b64encode(ref_img.read_bytes()).decode("ascii")
+        return client.post(url, headers=headers, json=data)
+
+    with ref_img.open("rb") as fh:
+        return client.post(
+            url,
+            headers=headers,
+            data=payload,
+            files={"input_reference": ("reference.jpg", fh, "image/jpeg")},
+        )
+
+
 def render_storyboard_video(job_id: str, local_id: str, provider_id: str, ref_path: Path, prompt: str, model: str, seconds: str, size: str) -> None:
     import httpx
 
@@ -1714,13 +1801,7 @@ def render_storyboard_video(job_id: str, local_id: str, provider_id: str, ref_pa
             create = None
             create_errors: list[str] = []
             for create_path in VIDEO_CREATE_PATHS:
-                with ref_img.open("rb") as fh:
-                    resp = client.post(
-                        f"{base}{video_path(create_path)}",
-                        headers=headers,
-                        data=payload,
-                        files={"input_reference": ("reference.jpg", fh, "image/jpeg")},
-                    )
+                resp = submit_video_create(client, f"{base}{video_path(create_path)}", headers, ref_img, payload)
                 if resp.status_code < 400:
                     create = resp
                     break