fix bug for embedding caused by PR 1188 (#1194)

This commit is contained in:
amstrongzyf
2025-08-18 20:15:18 +08:00
committed by GitHub
parent 3260c0c3ff
commit c150d33603
+5 -2
View File
@@ -481,9 +481,12 @@ class APIBackend(ABC):
or "\\'messages\\' must contain the word \\'json\\' in some form" in e.message
):
kwargs["add_json_in_prompt"] = True
too_long_error_message = "maximum context length" in e.message or "input must have less than" in e.message # type: ignore[attr-defined]
if hasattr(e, "message") and embedding and too_long_error_message:
too_long_error_message = hasattr(e, "message") and (
"maximum context length" in e.message or "input must have less than" in e.message
)
if embedding and too_long_error_message:
if not embedding_truncated:
# Handle embedding text too long error - truncate once and retry
model_name = LLM_SETTINGS.embedding_model