LlamaIndex
Log in

Find answers from the community

Updated 2 months ago

I think there is a bug in

sstrike
·
I think there is a bug in SimpleDirectoryReader or the Documentation needs to be updated.
I am unable to parse/index .ppt but the .pptx's are just fine.

Anyone can help/confirm if what needs to be done.
Thanks,
L
s
p
24 comments
LLogan M
Is there an error? Its just using the PptxReader
Attachment
image.png
LLogan M
Its using from pptx import Presentation to parse, maybe that package is having trouble reading your file
LLogan M
I might recommend trying llama-parse if you haven't already
sstrike
It is using pptx reader and just parses .pptx and not .ppt.
sstrike
Inside pptx module
Attachment
image.png
sstrike
@Logan M llama-parse is limited to pdfs as of now; at least the web version. That I interacted with.
Any Suggestions?
sstrike
PptxReader parses/processes only .pptx file type not .ppt
LLogan M
llama-parse was recently updated to handle 50+ file types
LLogan M
the .ppt extension is mapped to the PptxReader -- I guess this is an error?
sstrike
yes
sstrike
Is this update in API or web_application?
LLogan M
I think just the API -- pip install -U llama-parse
LLogan M
https://github.com/run-llama/llama_parse/blob/main/examples/other_files/demo_ppt_basic.ipynb
sstrike
awesome; will check it out - thanks
sstrike
Hey; tried the API here's the error:
Attachment
screenshot_2024-03-20_135732.png
LLogan M
awe man, I thought it was more robust than that. Maybe it wasn't fully rolled out...
LLogan M
@pld just curious, we have plans to expand that list right?
sstrike
is there any branch that has that piece that handles .ppt?
ppld
yes, we should extend it.
ppld
https://github.com/run-llama/llama_parse/pull/110 👀
sstrike
@Logan M Are there any plans to bring this extended support to llama-index readers?
LLogan M
No plans currently, but I welcome a PR to add any new reader
sstrike
And how is llama-parse handling the extended format(s)?
is there any code I can look into?
LLogan M
Currently its being provided as a SASS service, with options for on-prem VPC deployments in the future. (So no source code for now)
Add a reply
Sign up and join the conversation on Discord