Pieces CLI
Get the Pieces CLIIn order to use the Pieces CLI, you must have Pieces OS. We recommend also using the Pieces for Developers desktop app.
Installing
To get started with the Pieces Python CLI, you need to:
-
Ensure Pieces OS is installed and running on your system.
-
Install the Python package:
pip install pieces-cli
conda install pieces-cli
Getting Started​
After installing the CLI, you can access its functionalities through the terminal. The tool is initialized with the command pieces
followed by various subcommands and options.
Some important terminologies​
x
-> The indexcurrent asset
-> The asset that you are currently using can be changed by the open commandcurrent conversation
-> The conversation that you currently using in the ask command
Usage​
Run​
The run command starts the CLI in a loop. While you can use each command without running the CLI in a loop you'll get much faster results and a better experience using run.
Once the CLI is running in a loop you can simply type the command.
For instance: open
Instead of: pieces open
If you have a numbered list or search open you can just type the number and it will open the asset associated.
pieces run
List command​
To list assets applications or models, use the command:
Default of 10 assets​
pieces list
Lists your x most recent assets​
pieces list assets x
Lists all registered applications​
pieces list apps
Lists all accessible AI models​
pieces list models
Open an asset:​
Opens an asset from a list or search. If only "open" is used then it will open your most recent asset. This also creates a link to the asset's code.
pieces open [ITEM_INDEX]
Save, Create, Edit, and Delete Assets​
The save create edit and delete commands currently work on the current asset which is by default set to the most recent one and you can change the current asset to anything using the open command above.
Save the current asset:​
You need to edit the snippet code that was opened via the open command pieces open
and then save the changes using the save command
pieces save
Create a new asset:​
Will take whatever text is copied to your clipboard and create an asset. The asset will automatically be scanned and recognized for its file type.
pieces create
Edit an existing asset:​
This will edit the name and reclassify the current asset.
pieces edit
This is used to edit both the classification and name of an asset.
pieces edit --name "new name"
to edit the name
pieces edit --classification python
to edit the classification
You can you -n or -c to change the name and classification respectively with the edit command.
Delete an asset:​
This will delete an opened asset by using a list or search first. If you do not have an opened asset it will open your most recent asset and ask if you'd like to delete it.
pieces delete
Search and Query​
Perform a fuzzy search:​
pieces search [your query]
Finds strings that approximately match patterns. Normal search.
Perform a Neural Code Search:​
pieces search query --mode ncs
Uses machine learning, deep neural networks, and natural language processing. It can understand the intent of a user's search query and match it with the most relevant results.
Perform a Full-Text Search:​
pieces search query --mode fts
Examines all words in a document to find matches to search criteria.
Change the llm model you are using:​
Change the model in the ask command.
pieces change_model [MODEL_INDEX]
Conversations command​
Conversations​
List all conversations that you have. The green conversation is the one that is currently being used in the ask command
pieces conversations
Conversation​
List the messages in the currently used conversation in the ask command.
pieces conversation
You can create a new conversation that will be used in the ask command.
pieces conversation -n
Rename a conversation
pieces conversation -r My awesome name
Or you can make the model rename it for you
pieces conversation -r
Delete a conversation
pieces conversation -d
You can switch the conversation and list its messages. Check the conversations command to get the index
pieces conversation x
Ask a question to a model:​
Requires quotes around question
Ask the Pieces Copilot a question. You can add a relevant file, you can also add relevant snippets based on their index shown in the assets list command. In order to change models from the default (GPT3.5), use the change_model
command You can add a relevant file or snippet index by checking the list assets
command.
You can use .
or /
to refer to the current directory
pieces ask "your question" -f /file1 /file2 folder -s 1 2 3
Auto-Commit to Github​
Auto-commit your GitHub code and generate a commit message based on the changes in your staged area.
you can use the -p
or --push
flags to push the code to the repo too.
pieces commit -p
You can also add the -i
or --issues
flags so the commit message will include the issue numbers in the commit if found otherwise it will list all the issues to choose from.
pieces commit -i
You can use the -a
flag to stage all of your changes and then generate the committing messages
Login and logout​
Login to pieces
pieces login
Logout from pieces
pieces logout
Additional Commands​
Retrieve the version of Pieces OS and the CLI:​
pieces version
For a detailed help menu:​
pieces help
clear:​
It only works in the loop to clear the terminal.
clear
Supported Versions​
- Windows 10 or Greater
- Mac (insert later)
- Linux (insert later)
Please keep the CLI updated to the latest version to ensure compatibility with Pieces OS and access to all features. Please take a look at our documentation for details on supported versions.
Contributing​
Installation​
To run this project locally, follow these steps:
-
Fork this project via GitHub.
-
Clone this project:
git clone https://github.com/pieces-app/cli-agent
- Create a Virtual Environment
python3 -m venv venv
- Activate Your virtualenv
source venv/bin/activate for Mac & Linux OS
cd venv\Scripts for Windows OS
activate
- This project uses poetry for managing dependencies and builds. Install poetry with:
pip install poetry
- Then use poetry to install the required dependencies
poetry install
- You build with
poetry build
- Finally any project dependencies should be added to the pyproject.toml file with
poetry add
- Open the Dist folder
cd dist
- Install the WHL file
pip install pieces-cli-{VERSION}-py3-none-any.whl
replace the VERSION with the version you downloaded Note: Ensure you get latest from the releases of the cli-agent
- To view all the CLI Commands
pieces help
these can be local/github/pypi etc.
Updating​
To update the project, run the following command:
pip install pieces-cli --upgrade
Uninstallation​
To uninstall the project, run the following command:
pip uninstall pieces-cli
Don't forget to remove the virtual environment and dist folder