BC

how do I call all points in a qdrant vector_store

11 comments

@Logan M getting this error when running

getting this error when running LLMQuestionGenerator:

Plain Text

  SubQuestion expected dict not str (type=type_error)

from this code:

Plain Text

query= input("Query: ")

question_gen = LLMQuestionGenerator.from_defaults(llm=Settings.llm)

tool_choices = [
    ToolMetadata(
        name="vyos",
        description="Provides information about VyOS and its contents",
    ),
    ToolMetadata(
        name="ubuntu",
        description="Provides information about Ubuntu and its contents",
    ),
]

# Generate sub-questions
query_bundle = QueryBundle(query_str=query)
choices = question_gen.generate(tool_choices, query_bundle)
for choice in choices:
    print(choice)

4 comments

BBC

@Logan M I'm trying to use filtering_

I'm trying to use filtering_list = MetadataFilters(filters=[
MetadataFilter(key='file_name', operator=FilterOperator.EQ, value=metadata),
MetadataFilter(key='file_name', operator=FilterOperator.EQ, value=metadata),
condition="or")

the "value = metadata" is just the name of the metadata I gave a group of documents in the database to be called to later from the filter. The issue is when my filter faces two different topics and tries to retrieve information it stops working and outputs the wrong response. Is there a way to refine the filter in a way to guarantee it pulls the two different metadatas?

2 comments

BBC

@Logan M in the SimpleDirectoryReader()

in the SimpleDirectoryReader() llama-index-readers-file isn't recognizing PandasExcelReader even though the documentation says it was added:
from llama_index.readers.file import (
DocxReader,
EpubReader,
HWPReader,
ImageReader,
IPYNBReader,
MarkdownReader,
MboxReader,
PandasCSVReader,
PandasExcelReader,
PDFReader,
PptxReader,
VideoAudioReader,
) # pants: no-infer-dep

3 comments

BBC

For SimpleComposableMemory: I'm trying

For SimpleComposableMemory: I'm trying to use the clear method on VectorMemory with different vector stores but it is not a function for any external vector stores and it is not a function for any of the internal vector stores provided by llama-index. This is the function causing issues in the script:

def reset(self) -> None:
"""Reset chat history."""
self.vector_index.vector_store.clear()

Has anyone gotten SimpleComposableMemory to work with a vector store of any kind. Currently working with ChromaDB

40 comments

BBC

I'm having issues when adding documents

I'm having issues when adding documents to an already built vectorstore so when the documents are added it looks like this:
{'page_label': '2213', 'file_name': 'pfsense'}
{'page_label': '2214', 'file_name': 'pfsense'}
{'page_label': '2215', 'file_name': 'pfsense'}
{'page_label': '2216', 'file_name': 'pfsense'}
{'page_label': '2217', 'file_name': 'pfsense'}
{'page_label': '2218', 'file_name': 'pfsense'}
{'page_label': '2219', 'file_name': 'pfsense'}
{'page_label': '2220', 'file_name': 'pfsense'}
{'page_label': '2221', 'file_name': 'pfsense'}
{'page_label': '2222', 'file_name': 'pfsense'}
{'page_label': '2223', 'file_name': 'pfsense'}
I want to get rid of the page_label because what its supposed to look like is this:
{'file_name': 'pfsense'}

this is the code:

12 comments

BBC

@Logan M I'm having issues with the

@Logan M I'm having issues with the response from my agent where function calls are being outputted with the response and getting weird responses asking for my turn to answer. How do I refine the response so it doesn't include this mess?

using CoAAgentWorker() class and getting an odd output as the response from the agent:

Plain Text

Here's my rewritten answer:

Available functions: [FUNC ?(input: string) = y1]

Question: how to find an ip address in vyos?

Abstract plan of reasoning: To find an IP address in VyOS, first access the Vyos terminal by executing Rewrite The client MAY choose to explicitly provide the identifier through the ‘client identifier’ option. If the client supplies a ‘client identifier’, the client MUST use the same ‘client identifier’ in all subsequent messages, and the server MUST use that identifier to identify the client.

Example: set interfaces bonding bond0 vif 10 dhcp-options client-id 'foo-bar' set interfaces bonding <interface> vif <vlan-id> dhcp-options host-name <hostname>

Instead of sending the real system hostname to the DHCP server, overwrite the host-name with this given-value.]. Then, run the command [FUNC "show interfaces eth1" = y3] to display information about eth1. The output will include the IP address associated with eth1.

Your Turn:
Please execute the function calls and fill in the placeholders (y1, "Rewrite The client MAY choose to explicitly provide the identifier through the ‘client identifier’ option. If the client supplies a ‘client identifier’, the client MUST use the same ‘client identifier’ in all subsequent messages, and the server MUST use that identifier to identify the client.

Example: set interfaces bonding bond0 vif 10 dhcp-options client-id 'foo-bar' set interfaces bonding <interface> vif <vlan-id> dhcp-options host-name <hostname>

Instead of sending the real system hostname to the DHCP server, overwrite the host-name with this given-value." and y3) to get the final response.

2 comments

BBC

@Logan M is it possible to change the

@Logan M is it possible to change the functioncalling agent's api so it doesn't require openai I want to use ollama but im only familiar with implementing its base url call to "localhost:XXXX" amd having it connect. Is there a way to implement that in this scenario?