Skip to content

README.md

Latest commit

 

History

History
142 lines (102 loc) · 4.43 KB

README.md

File metadata and controls

142 lines (102 loc) · 4.43 KB

deepset CLI

The deepset CLI is a command-line interface tool that you can use to interact with the deepset SDK and perform various operations, such as uploading files and folders to your deepset workspace.

Installation

To install the deepset CLI, use pip:

pip install deepset-cloud-sdk

Configuration

Before using the deepset CLI, log in and provide your credentials. You can do this by running the command:

On MacOS and Linux:

deepset-cloud login

On Windows:

python -m deepset_cloud_sdk.cli login

This command prompts you to enter your API key and default workspace name. Once you provide these details, the CLI stores your credentials in the ~/.deepset-cloud/.env file. This file is used as the default configuration for subsequent CLI commands.

Alternatively, to use a different environment file for your configuration, you can create an .env file in the local directory. Additionally, you have the flexibility to provide the credentials directly as command-line arguments or set them programmatically in your code.

Usage

You can use the deepset CLI by running the following command:

On MacOS and Linux:

deepset-cloud <command>

On Windows:

python -m deepset_cloud_sdk.cli <command>

Replace with one of the supported commands. To list all available commands, use the --help flag.

Example Commands

Upload Files and Folders

You don't have to follow any special folder structure. If there are multiple files with the same name in your folder, they're all uploaded by default. You can change this behavior with the --write-mode flag. See the examples below.

This command uploads the file example.txt to your deepset workspace. On MacOS and Linux:

deepset-cloud upload ./examples/data/example.txt

On Windows:

python -m deepset_cloud_sdk.cli upload ./examples/data/example.txt

This command uploads all .txt and .pdf files from the folder located in the examples directory to your deepset workspace. By default only .txt and .pdf files are uploaded. To upload different file types see below.

The paths in the examples are relative to the current working directory.

On MacOS and Linux:

deepset-cloud upload ./examples/data

On Windows:

python -m deepset_cloud_sdk.cli upload ./examples/data

To overwrite existing files in your project, use the --write-mode flag. For example:

On MacOS and Linux:

deepset-cloud upload ./examples/data --write-mode OVERWRITE

On Windows:

python -m deepset_cloud_sdk.cli upload ./examples/data --write-mode OVERWRITE

This syncs your local files with the files in your deepset workspace without having to manually delete the files in your workspace.

Upload different file types

To upload other file types than text, specify the desired file types using the flag --use-type. The command below uploads all file types from the ./example/data directory that are supported by deepset AI Platform.

deepset-cloud upload ./examples/data --use-type .csv --use-type .docx --use-type .html --use-type .json --use-type .md --use-type .txt --use-type .pdf --use-type .pptx --use-type .xlsx --use-type .xml

On Windows:

python -m deepset_cloud_sdk.cli upload ./examples/data --use-type .csv --use-type .docx --use-type .html --use-type .json --use-type .md --use-type .txt --use-type .pdf --use-type .pptx --use-type .xlsx --use-type .xml

Downloading Files from deepset AI Platform

This command downloads all files from a workspace to a local directory. For example:

On MacOS and Linux:

deepset-cloud download --workspace-name <your-workspace-name>

On Windows:

python -m deepset_cloud_sdk.cli download --workspace-name <your-workspace-name>

To filter for specific files, use the same filters as for listing files.

List Files

You can run the list-files operation to search files in your deepset workspace. For example:

On MacOS and Linux:

deepset-cloud list-files

On Windows:

python -m deepset_cloud_sdk.cli list-files

with optional arguments:

--name "<your-file-name>"  # search by file name
--content "content" # search by file content
--odata-filter "key eq 'value'" # search by odata filter

Support

If you encounter issues or have questions, reach out to our team on Discord.

We hope you find the deepset CLI useful in your projects. Happy coding!