Skip to content

cherrry-ai/cherrry-py

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

cherrry-py

Cherrry Python SDK

CleanShot 2022-11-30 at 21 40 41@2x

App | Docs | Blog

Installation + API keys

pip package

Install with pip. See it on pypi

pip install cherrry

API Keys

From https://cherrry.com/dashboard/api to get your API Keys

Private Key

Private keys start with ch_prv

keep it secret and never use it client-side. It has service role privilages: it can read + write data.

Public Key

Public keys start with ch_pub

They're intended to be use client-side and have read-only privilages.

Initalize

from cherrry import CherrryClient

initialize the client

client = CherrryClient("your_api_key")

Concepts

Table

A table is a collection of documents.

Document

A document is respresented as a JSON object literal with three fields: text, image, and metadata. These fields are also JSON object literals, where the keys can be strings with any contents, and their values are also strings.

text and image are semantically searchable each by their type respectively. Each document must have either a text or image field. It can also have both fields. metadata is used to store additional information and for filtering (feature in progress), it is an optional field.

Basic Functions

Create Table

[success, error] = client.create_table("table_name")

Insert a Doc

Documents must be of the following form

{
    "text": {
        "a name for your text": "your desired text in a string"
    },
    "image": {
        "a name for your image": "a url to your downloadable image"
    },
    "metadata": {
        "key": "value"
    }
}

for example:

[data, error] = client.table("recipes").insert({
    "text": {
        "name": "Octopus Cherry Pie"
    },
    "image": {
        "preview": "https://i.imgur.com/lFC8p0L.jpeg"
    },
    "metadata": {
        "author_name": "Davy Jones",
        "author_email": "octo@pus.com"
    }
})

Search

[data, error] = client
    .table("blogs")
    .search({ "prompt": "sea creature desert", "size": 1, "search_type": "image" });

Get Doc by ID

The ID of documents are returned in the responses of /search or /doc

[data, error] = client.table("blogs").doc("1234")

Delete a Doc

[success, error] = client.table("blogs").delete("1234")

Releases

No releases published

Packages

No packages published

Languages