ValueError
builtins.ValueError
Stack trace
Traceback (most recent call last):
File "app.py", line 42, in <module>
rerank_results = reranker.rerank(query, documents)
File "/usr/local/lib/python3.9/site-packages/langchain/rerankers/base.py", line 88, in rerank
raise ValueError("Query string for reranking cannot be empty.")
ValueError: Query string for reranking cannot be empty. Why it happens
Reranking models require a non-empty query string to compare and score documents. If the query string is empty or None, the reranker cannot compute relevance scores, causing it to raise a ValueError. This often happens when upstream code fails to pass or generate a valid query.
Detection
Add input validation before calling rerank to assert the query string is not empty or None, and log the query value to catch missing or empty inputs early.
Causes & fixes
The query string variable passed to the reranker is an empty string.
Ensure the query string is populated with a meaningful non-empty string before calling the reranker.
The query string is None due to a missing or failed extraction from user input or previous pipeline step.
Add checks to confirm the query extraction step returns a valid string and handle None cases with fallback or error messages.
The reranker is called before the query is generated or assigned in asynchronous or multi-step workflows.
Reorder code or add synchronization to guarantee the query string is ready and valid before reranking.
Code: broken vs fixed
from langchain.rerankers import OpenAIReranker
reranker = OpenAIReranker()
query = "" # Empty query string
documents = ["doc1", "doc2"]
rerank_results = reranker.rerank(query, documents) # This line raises ValueError
print(rerank_results)import os
from langchain.rerankers import OpenAIReranker
# Ensure OPENAI_API_KEY is set in environment
# os.environ["OPENAI_API_KEY"] = os.getenv("OPENAI_API_KEY") # Usually set externally
reranker = OpenAIReranker()
query = "What is AI?" # Non-empty query string
if not query:
raise ValueError("Query string for reranking cannot be empty.")
documents = ["doc1", "doc2"]
rerank_results = reranker.rerank(query, documents) # Fixed: query is non-empty
print(rerank_results)Workaround
Wrap the rerank call in try/except ValueError, and if caught, assign a default query string or skip reranking to avoid crashing.
Prevention
Implement strict input validation and pipeline checks to guarantee the query string is always non-empty before reranking, and use type hints or assertions to catch issues early.