Find answers from the community

Updated 2 years ago

Good morning people everything fine

At a glance
Good morning people, everything fine?

I have some questions related to how to work with unstructured data.

I'm hitting the Github API, getting data from several endpoints and saving it to a Vector database.

Then I'm connected via langchain, to talk to this data.

But the result is pretty bad, anyone doing something similar?

Searching I saw that I will need to transform this data using some entity extraction techniques, training pairs and narrative format.

But I'm in doubt if I'm on the right path, if someone who has knowledge about can help.
L
W
2 comments
are you trying to answer questions about code?
Yes

exemplo json:
Plain Text
        "commit": {
            "author": {
                "name": "Name author",
                "email": "test@test.com",
            },
            "committer": {
                "name": "GitHub",
                "email": "noreply@github.com",
            },
            "message": "Merge pull request #10000 from ezDevs/feat/remove-trial-modal\n\nfeat: remove career acessory and fix modal to schedule a mentory",
        },
        "author": {
            "login": "login author",
            "id": id,
        },
        "committer": {
            "login": "login",
            "id": Id,
        },

i get the json and save to vectorDB, then join the vector bank with langchain to ask questions

Question: List to last commits
Response: Last commit is sha_id

Question: What other information do you have about the commit
Response: I don't have enough information to answer this

How to improve data to conversation Have a better conversation with unstructured data?
Add a reply
Sign up and join the conversation on Discord