From dcffb5269ab2fc6c01e8b3eb687332a4e126541a Mon Sep 17 00:00:00 2001
From: jnMetaCode <1394485448@qq.com>
Date: Tue, 17 Mar 2026 17:53:03 +0800
Subject: [PATCH] fix: only pass dimensions when explicitly configured in
 embedding config (#6432)

* fix: only pass dimensions param when explicitly configured

Models like bge-m3 don't support the dimensions parameter in the
embedding API, causing HTTP 400 errors. Previously dimensions was
always sent with a default value of 1024, even when the user never
configured it. Now dimensions is only included in the request when
embedding_dimensions is explicitly set in provider config.

Closes #6421

Signed-off-by: JiangNan <1394485448@qq.com>

* fix: handle invalid dimensions config and align get_dim return

- Add try-except around int() conversion in _embedding_kwargs to
  gracefully handle invalid embedding_dimensions config values
- Update get_dim() to return 0 when embedding_dimensions is not
  explicitly configured, so callers know dimensions weren't specified
  and can handle it accordingly
- Both methods now share consistent logic for reading the config

Signed-off-by: JiangNan <1394485448@qq.com>

* fix: improve logging for invalid embedding_dimensions configuration

---------

Signed-off-by: JiangNan <1394485448@qq.com>
Co-authored-by: Soulter <905617992@qq.com>
---
 .../sources/openai_embedding_source.py        | 27 ++++++++++++++++---
 1 file changed, 24 insertions(+), 3 deletions(-)

diff --git a/astrbot/core/provider/sources/openai_embedding_source.py b/astrbot/core/provider/sources/openai_embedding_source.py
index 8bf92ef4d..04397b182 100644
--- a/astrbot/core/provider/sources/openai_embedding_source.py
+++ b/astrbot/core/provider/sources/openai_embedding_source.py
@@ -40,25 +40,46 @@ class OpenAIEmbeddingProvider(EmbeddingProvider):
 
     async def get_embedding(self, text: str) -> list[float]:
         """获取文本的嵌入"""
+        kwargs = self._embedding_kwargs()
         embedding = await self.client.embeddings.create(
             input=text,
             model=self.model,
-            dimensions=self.get_dim(),
+            **kwargs,
         )
         return embedding.data[0].embedding
 
     async def get_embeddings(self, text: list[str]) -> list[list[float]]:
         """批量获取文本的嵌入"""
+        kwargs = self._embedding_kwargs()
         embeddings = await self.client.embeddings.create(
             input=text,
             model=self.model,
-            dimensions=self.get_dim(),
+            **kwargs,
         )
         return [item.embedding for item in embeddings.data]
 
+    def _embedding_kwargs(self) -> dict:
+        """构建嵌入请求的可选参数"""
+        kwargs = {}
+        if "embedding_dimensions" in self.provider_config:
+            try:
+                kwargs["dimensions"] = int(self.provider_config["embedding_dimensions"])
+            except (ValueError, TypeError):
+                logger.warning(
+                    f"embedding_dimensions in embedding configs is not a valid integer: '{self.provider_config['embedding_dimensions']}', ignored."
+                )
+        return kwargs
+
     def get_dim(self) -> int:
         """获取向量的维度"""
-        return int(self.provider_config.get("embedding_dimensions", 1024))
+        if "embedding_dimensions" in self.provider_config:
+            try:
+                return int(self.provider_config["embedding_dimensions"])
+            except (ValueError, TypeError):
+                logger.warning(
+                    f"embedding_dimensions in embedding configs is not a valid integer: '{self.provider_config['embedding_dimensions']}', ignored."
+                )
+        return 0
 
     async def terminate(self):
         if self.client: