search

snexus / llm-search

Querying local documents, powered by LLM
MIT License
479 stars 60 forks source link

support for cpp and hpp source code search? Generic text search? #90

Closed bradphelan closed 7 months ago

bradphelan commented 7 months ago

I'd like to try and get this to search my source code.

snexus commented 7 months ago

Hi,

Currently, there is no special mode for source files, but it can be parsed as a regular text file. I believe it won't be optimal, but you are welcome to try.

bradphelan commented 7 months ago

When I add hpp and cpp files to the config file it reports an error message saying that such files are not supported.

On Wed, 14 Feb 2024, 15:42 Denis Lapchev, @.***> wrote:

Hi,

Currently, there is no special mode for source files, but it can be parsed as a regular text file. I believe it won't be optimal, but you are welcome to try.

— Reply to this email directly, view it on GitHub https://github.com/snexus/llm-search/issues/90#issuecomment-1943943312, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAEJ4S73VR2PPNCOGGZDM3YTTENPAVCNFSM6AAAAABDIDCESGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNBTHE2DGMZRGI . You are receiving this because you authored the thread.Message ID: @.***>

snexus commented 7 months ago

You are right, my bad, it has a restriction on available formats at the moment (the only native textual format supported is markdown). Here are all supported formats for the current version - https://github.com/snexus/llm-search/blob/5243360d90254385541f72137f41e27c4e7461b3/src/llmsearch/config.py#L40

It should be easy to extend, as unstructured.io supports source files nowadays, see here - https://unstructured-io.github.io/unstructured/core/partition.html

Will add it as an enhancement...

snexus commented 7 months ago

Support added in 0.6.2