Add common issues to question and answer chain node (#2686)

This commit is contained in:
Justin Ellingwood
2024-12-05 18:17:32 +00:00
committed by GitHub
parent de9b9fd90b
commit b558951b71
5 changed files with 50 additions and 3 deletions

View File

@@ -1,6 +1,6 @@
---
#https://www.notion.so/n8n/Frontmatter-432c2b8dff1f43d4b1c8d20075510fe4
title: AI Agent node Common Issues
title: AI Agent node common issues
description: Documentation for common issues and questions in the AI Agent node in n8n, a workflow automation platform. Includes details of the issue and suggested solutions.
contentType: integration
priority: critical

View File

@@ -83,4 +83,4 @@ You might see this error in one of two scenarios:
1. When you've set the **Prompt** to **Define below** and haven't entered anything in the **Text** field.
* To resolve, enter a valid prompt in the **Text** field.
2. When you've set the **Prompt** to **Take from previous node automatically** and the incoming data has no field called `chatInput`.
* The node expects the `chatInput` field. If your previous node doesn't have this field, add an [Edit Fields (Set)](/builtin/core-nodes/n8n-nodes-base.set/) node to edit an incoming field name to `chatInput`.
* The node expects the `chatInput` field. If your previous node doesn't have this field, add an [Edit Fields (Set)](/integrations/builtin/core-nodes/n8n-nodes-base.set/) node to edit an incoming field name to `chatInput`.

View File

@@ -0,0 +1,40 @@
---
#https://www.notion.so/n8n/Frontmatter-432c2b8dff1f43d4b1c8d20075510fe4
title: Question and Answer Chain node common issues
description: Documentation for common issues and questions in the Question and Answer Chain node in n8n, a workflow automation platform. Includes details of the issue and suggested solutions.
contentType: integration
priority: high
---
# Question and Answer Chain node common issues
Here are some common errors and issues with the [Question and Answer Chain node](/integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainretrievalqa/) and steps to resolve or troubleshoot them.
## No prompt specified error
This error displays when the **Prompt** is empty or invalid.
You might see this in one of two scenarios:
1. When you've set the **Prompt** to **Define below** and have an expression in your **Text** that isn't generating a value.
* To resolve, enter a valid prompt in the **Text** field.
* Make sure any expressions reference valid fields and that they resolve to valid input rather than null.
2. When you've set the **Prompt** to **Take from previous node automatically** and the incoming data has null values.
* To resolve, make sure your input contains a `chatInput` field. Add an [Edit Fields (Set)](/integrations/builtin/core-nodes/n8n-nodes-base.set/) node to edit an incoming field name to `chatInput`.
* Remove any null values from the `chatInput` field of the input node.
<!-- vale from-write-good.Passive = NO -->
## A Retriever sub-node must be connected error
<!-- vale from-write-good.Passive = YES -->
This error displays when n8n tries to execute the node without having a Retriever connected.
To resolve this, click the + Retriever button at the bottom of your screen when the node is open, or click the Retriever + connector when the node isn't open. n8n will then open a selection of possible Retrievers to pick from.
## Can't produce longer responses
If you need to generate longer responses than the Question and Answer Chain node produces by default, you can try one or more of the following techniques:
* **Connect a more verbose model**: Some AI models produce more terse results than others. Swapping your model for one with a larger context window and more verbose output can increase the word length of your responses.
* **Increase the maximum number of tokens**: Many model nodes (for example the [OpenAI Chat Model](/integrations/builtin/cluster-nodes/sub-nodes/n8n-nodes-langchain.lmchatopenai/#maximum-number-of-tokens)) include a **Maximum Number of Tokens** option. You can set this to increase the maximum number of tokens the model can use to produce a response.
* **Build larger responses in stages**: For more detailed answers, you may want to construct replies in stages using a variety of AI nodes. You can use AI split up a single question into multiple prompts and create responses for each. You can then compose a final reply by combining the responses again. Though the details are different, you can find a good example of the general idea in this [template for writing a WordPress post with AI](https://n8n.io/workflows/2187-write-a-wordpress-post-with-ai-starting-from-a-few-keywords/).

View File

@@ -28,4 +28,9 @@ The question you want to ask.
Refer to [LangChain's documentation on retrieval chains](https://js.langchain.com/docs/tutorials/rag/){:target=_blank .external-link} for examples of how LangChain can use a vector store as a retriever.
--8<-- "_snippets/integrations/builtin/cluster-nodes/langchain-overview-link.md"
## Common issues
For common errors or issues and suggested resolution steps, refer to [Common Issues](/integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainretrievalqa/common-issues/).
--8<-- "_glossary/ai-glossary.md"

View File

@@ -855,7 +855,9 @@ nav:
- Tools Agent: integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.agent/tools-agent.md
- Common issues: integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.agent/common-issues.md
- Basic LLM Chain: integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainllm.md
- Question and Answer Chain: integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainretrievalqa.md
- Question and Answer Chain:
- Question and Answer Chain: integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainretrievalqa/index.md
- Common issues: integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainretrievalqa/common-issues.md
- Summarization Chain: integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainsummarization.md
- Information Extractor: integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.information-extractor.md
- Text Classifier: integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.text-classifier.md