fix: generate summary error when tokens=4097 (langgenius#488)

HuberyHuV1 · Jun 29, 2023 · 7a897b5 · 7a897b5
1 parent 745f146
commit 7a897b5
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 2 deletions.
diff --git a/api/core/generator/llm_generator.py b/api/core/generator/llm_generator.py
@@ -45,7 +45,7 @@ def generate_conversation_summary(cls, tenant_id: str, messages):
         prompt = CONVERSATION_SUMMARY_PROMPT
         prompt_with_empty_context = prompt.format(context='')
         prompt_tokens = TokenCalculator.get_num_tokens(model, prompt_with_empty_context)
-        rest_tokens = llm_constant.max_context_token_length[model] - prompt_tokens - max_tokens
+        rest_tokens = llm_constant.max_context_token_length[model] - prompt_tokens - max_tokens - 1
 
         context = ''
         for message in messages:
@@ -56,6 +56,9 @@ def generate_conversation_summary(cls, tenant_id: str, messages):
             if rest_tokens - TokenCalculator.get_num_tokens(model, context + message_qa_text) > 0:
                 context += message_qa_text
 
+        if not context:
+            return '[message too long, no summary]'
+
         prompt = prompt.format(context=context)
 
         llm: StreamableOpenAI = LLMBuilder.to_llm(

diff --git a/api/tasks/generate_conversation_summary_task.py b/api/tasks/generate_conversation_summary_task.py
@@ -28,7 +28,7 @@ def generate_conversation_summary_task(conversation_id: str):
     try:
         # get conversation messages count
         history_message_count = conversation.message_count
-        if history_message_count >= 5:
+        if history_message_count >= 5 and not conversation.summary:
             app_model = conversation.app
             if not app_model:
                 return