Skip to content

Commit

Permalink
Merge pull request #155 from hummingbot/update-install-docs
Browse files Browse the repository at this point in the history
doc/update installation
  • Loading branch information
cardosofede authored Jul 3, 2024
2 parents 86c440d + 41e7773 commit c2d8516
Show file tree
Hide file tree
Showing 2 changed files with 70 additions and 110 deletions.
154 changes: 63 additions & 91 deletions INSTALLATION.md
Original file line number Diff line number Diff line change
Expand Up @@ -3,10 +3,41 @@ This repository is maintained by Hummingbot Foundation as a companion for users

## Requirements

* 8 GB memory or more (On AWS, this is a `t2.large` instance)
* Linux / Debian / MacOS
Cloud Server or local machine

## Installation
* Minimum of at least 2vCPU and 8 GB memory or more (On AWS, this is a `t2.large` instance)
* Linux / MacOS / Windows*

* For Windows users, make sure to install WSL2 as well as a Linux distro like Ubuntu and run the commands listed below in a Linux terminal and **NOT** in the Windows Command prompt or Powershell.

## Installation


### Method 1 - Deploy Repo

This is the **recommended** install procedure for normal users

1 - Install dependencies:

* [Docker Engine](https://docs.docker.com/engine/install/ubuntu/)

2 - Clone repo and navigate to the created directory
```bash
git clone https://github.com/hummingbot/deploy.git
cd deploy
```

3 - Run the provided bash script
```bash
bash setup.sh
```


Proceed to the **Launch Dashboard** section

### Method 2 - Source

This method is only recommended if you are a developer and want to make changes to the code.

1 - Install dependencies:

Expand All @@ -21,7 +52,7 @@ cd dashboard

3 - Create `conda` environment and install dependencies
```bash
make env_create
make install
```

4 - Activate the isolated 'conda' environment
Expand All @@ -31,123 +62,64 @@ conda activate dashboard

5 - Start the dashboard
```bash
streamlit run main.py
make run
```
## Enable Authentication

1 - Locate the `config.py` file in the /dashboard folder

2 - Open the file using a text editor

3 - Modify the `AUTH_SYSTEM_ENABLED` variable and set this to `True` to enable. By default this is set to `False`

## Configure Credentials for Authentication

### Method 1: Using hasher_generate
Don't forget to run the **Backend-API** and **Broker** separately for this to work

1 - Update the Environment

Ensure that you have the latest version of the project and reinstall the conda environment if needed.

2 - Generate Hashed Password

Open a Python terminal and utilize hasher_generate from Streamlit to generate a hashed password. Example code:
Proceed to the **Launch Dashboard** section

```
import streamlit_authenticator as st_auth
hashed_password = st_auth.Hasher("YOUR_PLAIN_TEXT_PASSWORD").generate()[0]
print(hashed_password)
```
## Launch the Dashboard

3 - Edit the Credentials YAML File
Open a web browser and navigate to <https://localhost:8501> to view the Dashboard.

Locate the credentials YAML file in the root directory.
Edit the file by inserting the generated hashed password from Step 2.
If you are using a cloud server or VPS, replace localhost with the IP of your server. You may need to edit the firewall rules to allow inbound connections to the necessary ports.

4 - Accessing the Dashboard

Run the Hummingbot Dashboard project.
Use the username and the plain text password (used in Step 2) to log in to the dashboard.

5 - Logging Out

To log out, navigate to the dashboard and select 'Logout'.


### Method 2: Using Pre-Authorized Emails for User Registration

1 - Update the Environment
## Updating

As in Method 1, ensure your conda environment and project version are up-to-date.
Before updating, make sure to stop any running instances first

2 - Whitelist Emails
### Deploy Repo

In the credentials YAML file, add the emails of users who will be pre-authorized to register and create their credentials.
To update - make sure you are in the `deploy` folder then run the bash script

```
pre_authorized_emails:
- [email protected]
- [email protected]
```
```
bash setup.sh
```

3 - User Registration
This will pull any latest images and recreate the Docker containers.

Direct users to the Hummingbot Dashboard.
Users should select 'Register' and enter one of the pre-authorized email addresses.
They will choose their own username and password during registration.

4 - First-Time Login

Users log in with their chosen username and password.
The dashboard will hash and store their credentials automatically.
### Source

5 - Logging Out

Similar to Method 1, users can log out by navigating to the dashboard and selecting 'Logout'.
To update the `dashboard` source for latest version, run:
```
cd dashboard
git pull
```

### Additional Notes
Once updated, start up the dashboard again:

- Video instruction link - https://drive.google.com/file/d/17ecd2aWHZCyuK5Etcv8AZwXMw2B-1sJG/view?usp=share_link
- Users created through Method 2 (Pre-Authorized Emails) will have their hashed passwords automatically generated and stored.
- Always ensure the secure handling and storage of credentials.
- Utilize the 'Logout' feature to secure information and restrict access to the dashboard when it is not in use.
```
make run
```

## Updating
To update the `dashboard` environment, run

To update the `dashboard` environment for changes to dependencies defined in `environment.yml`, remove the environment and re-create it:
```
make env_remove
make env_create
```

To updated the `dashboard` source for latest version, run:
```
cd dashboard
git pull
```
This will remove the `conda` environment and recreate it.

## Troubleshooting

For Dashboard issues, please open a ticket on our Dashboard [Github page](https://github.com/hummingbot/dashboard) or post in the `#hummingbot-deploy` channel in [Discord](https://discord.gg/hummingbot)

### Docker permissions

If you get an error like `Permission denied while trying to connect to the Docker daemon socket at unix:///var/run/docker.sock`, run this command to enable Docker permissions:
```
sudo chmod 6666 /var/run/docker.sock
```

### Sym-link data directory

To use the [Strategy Performance page](https://github.com/hummingbot/dashboard/wiki/%F0%9F%9A%80-Strategy-Performance), you need to establish a symbolic link to the `data` directory of your running Hummingbot instance:

The `data` directory differs for Docker versus Source installed Hummingbot:
* Docker installed: `/path/to/hummingbot/hummingbot_files/data`
* Source installed: `/path/to/hummingbot/data`

Create a symlink to your Hummingbot `/data` directory
```bash
# replace `/path/to/hummingbotdata` with the actual path
ln -s /path/to/hummingbotdata data

# if you need to remove the symlink
unlink data
```
26 changes: 7 additions & 19 deletions README.md
Original file line number Diff line number Diff line change
@@ -1,36 +1,24 @@
# Hummingbot Dashboard

Hummingbot Dashboard is an open-source application designed to assist in the creation, backtesting, and optimization of a wide variety of algorithmic trading strategies. Once refined, these strategies can be deployed as [Hummingbot](https://github.com/hummingbot/hummingbot) instances in either paper trading or live trading modes, providing a seamless transition from strategy formulation to actual trading execution.
Hummingbot Dashboard is an open-source application designed to assist in the creation, backtesting, and optimization of a wide variety of algorithmic trading strategies. Once refined, these strategies can be deployed as [Hummingbot](https://github.com/hummingbot/hummingbot) instances in live trading modes, providing a seamless transition from strategy formulation to actual trading execution.

## Features

- **Bot Orchestration**: Deploy and manage multiple instances of Hummingbot
- **Strategy Backtesting and Optimization**: Evaluate the performance of your strategies against historical data and optimize them with Optuna
- **One-Click Deployment**: Seamlessly deploy your strategies as Hummingbot instances for paper or live trading.
- **Performance Analysis Monitoring**: Monitor and analyze the performance of your deployed strategies.
- **Secure Credentials**: Restrict access to whitelisted users
- **Credential Management**: Create and manage separate accounts for API keys

## Tutorial

Get a comprehensive understanding of Hummingbot Dashboard by exploring our introductory video playlist. These videos will guide you through the various features and functionalities:

1. [Introduction to Dashboard](https://www.youtube.com/watch?v=a-kenMqRB00&list=PLDwlNkL_4MMf0Ifvj-HLuQ_Jtf7OV6uzW&index=1)
2. [Setting up the Environment](https://www.youtube.com/watch?v=AbezIhb6iJg&list=PLDwlNkL_4MMf0Ifvj-HLuQ_Jtf7OV6uzW&index=2)
3. [Managing Credentials](https://www.youtube.com/watch?v=VmlD_WQVe4M&list=PLDwlNkL_4MMf0Ifvj-HLuQ_Jtf7OV6uzW&index=3)
4. [Using the Master Bot Profile](https://www.youtube.com/watch?v=MPQTnlDXPno&list=PLDwlNkL_4MMf0Ifvj-HLuQ_Jtf7OV6uzW&index=4)
5. [Deploying Bots and Running Strategies](https://www.youtube.com/watch?v=915E-C2LWdg&list=PLDwlNkL_4MMf0Ifvj-HLuQ_Jtf7OV6uzW&index=5)
6. Analyzing Strategy Performance (coming soon)
7. [Controllers, Backtesting, and Optimization](https://www.youtube.com/watch?v=bAi2ok7_boo&list=PLDwlNkL_4MMf0Ifvj-HLuQ_Jtf7OV6uzW&index=6)
8. [Deploying Best Strategies from Backtests](https://www.youtube.com/watch?v=BJf3ml-9JIQ&list=PLDwlNkL_4MMf0Ifvj-HLuQ_Jtf7OV6uzW&index=7)
9. [Conclusions and Next Steps](https://www.youtube.com/watch?v=ug_SSZb2HYE&list=PLDwlNkL_4MMf0Ifvj-HLuQ_Jtf7OV6uzW&index=8)

## Documentation

For detailed instructions and further information, visit our [documentation page](https://hummingbot.org/dashboard/).

## Installation

Currently, Dashboard should be installed from source via the process below. In the future, we aim to support a simpler,Docker-based installation process as well.
Currently, we recommend users to install Dashboard using the **[Deploy repo](https://github.com/hummingbot/deploy)** instead as this will automatically launch Dashboard along with the needed components in their own Docker containers.

If you are a developer, and want to make changes to the code then we recommend using the Source installation below - please note that you will also need to launch the Backend API and Broker separately (either through source install or through Docker).

1. **Install Dependencies**:
- Docker Engine
Expand All @@ -44,7 +32,7 @@ Currently, Dashboard should be installed from source via the process below. In t

3. **Create Conda Environment and Install Dependencies**:
```bash
make env_create
make install
```

4. **Activate the Isolated 'conda' Environment**:
Expand All @@ -54,7 +42,7 @@ Currently, Dashboard should be installed from source via the process below. In t

5. **Start the Dashboard**:
```bash
streamlit run main.py
make run
```

For more detailed instructions on how to install and update the dashboard, refer to [INSTALLATION.md](INSTALLATION.md).
Expand Down

0 comments on commit c2d8516

Please sign in to comment.