VO.10

is the Pandas Excel Loader no longer supported ? and is there an undergoing update on llama hub ?

19 comments

With v0.10.x
Llama-hub is considered deprecated.

All the loaders/tools have been moved into llama-indec repo.

Also are you talking about pandas CSV reader? Didn't find anything related to excel reader

gghxsted.

i found it in this question and was wondering how to load complexe Excel files, if you have some insights about which loaders to use, or any recommendations, please don't hesitate
https://github.com/run-llama/llama_index/issues/9204

gghxsted.

is there also examples on how to use specific loaders, for example a Docx file reader or something similar, I used to get the examples from llama hub idk where to find them now

gghxsted.

Attachment

WWhiteFang_Jr

DocxReader requires a path object for file argument. Since you are passing a string that is why youa re getting the error.

Attachment

WWhiteFang_Jr

For this, Logan has already mentioned the solution. You can try this:https://github.com/run-llama/llama_index/issues/9204#issuecomment-1832046633

gghxsted.

can you tell me how to specify the reader that SimpleDirectoryReader uses by default, i'm trying to change the reader of the pdf to use llama parse and have docx read by unstructured, It's not mentioned what the default readers are, thanks

WWhiteFang_Jr

Plain Text

from llama_parse import LlamaParse
from llama_index.core import SimpleDirectoryReader

parser = LlamaParse(
    api_key="llx-...",  # can also be set in your env as LLAMA_CLOUD_API_KEY
    result_type="markdown",  # "markdown" and "text" are available
    verbose=True
)

file_extractor = {".pdf": parser}
documents = SimpleDirectoryReader("./data", file_extractor=file_extractor).load_data()

Same you can replace for docx

gghxsted.

Plain Text

files = Path("./DB_Useful/tests")
file_extractor = {".pdf": parser, ".docx":UnstructuredReader,".jpg":UnstructuredReader,".csv": UnstructuredReader}
documents = SimpleDirectoryReader(files, file_extractor=file_extractor, recursive=True).load_data()

gghxsted.

I'm using this code but I get this error since Unstructered take a file argument

gghxsted.

Failed to load file \tests\csv\file.csv with error: UnstructuredReader.load_data() missing 1 required positional argument: 'file'. Skipping...

gghxsted.

https://docs.llamaindex.ai/en/stable/api_reference/readers/file/#llama_index.readers.file.UnstructuredReader

gghxsted.

Do you have any infor about the default reader used by the SimpleDirectoryReader ? in the docs it's mentioned that it supports also Docx and I just want to knpow if using the default one is better than using Unstructured

WWhiteFang_Jr

default library for docs is docx2txt
Also you'll have to pass in like this:
file_extractor = {".pdf": parser,".docx":UnstructuredReader()}

gghxsted.

thanks, but it gives me another error this time

gghxsted.

Attachment

gghxsted.

even tho i have them installed both, unstructured and unstructured[docx]

WWhiteFang_Jr

try restarting the session

gghxsted.

Yep that did it, thank you soooo muuchhhh ;))))

Add a reply

Find answers from the community

VO.10