AutoSight is a metagraph-powered system that enables users to earn income while they drive, by contributing to a dataset of road images that AI companies pay to access.
The system is comprised of four components:
-
Mobile App: Users can use the AutoSight Android app on their phone to capture and upload images of the road from their vehicle, earning tokens for each upload.
-
Image Server: Images are uploaded and saved to an image server that hosts them publicly.
-
Metagraph: A custom metagraph works as the blockchain backend for the AutoSight system and is responsible for validating, storing, and publishing image data.
-
Web App: Customers in demand for road images can view them on the AutoSight web app, which features a menu of image previews and interactive map.
See further project details on Devpost.
Follow the instructions below to set up and configure each of the AutoSight components.
Install the required dependencies:
- Python3
- Node JS
- Euclid Development Environment dependencies
- Android Studio with an Android phone
Follow the Euclid Development Environment quickstart guide to clone and setup the development environment, including the steps to configure docker and a github access token.
In a new terminal window, navigate to the cloned euclid-development-environment directory:
cd euclid-development-environmentInstall the autosight project from this repository into the development environment:
scripts/hydra install-template autosight --repo https://github.com/rhammell/autosight --path metagraphBuild the autosight project containers:
scripts/hydra buildStart the metagraph components from a genesis snapshot:
scripts/hydra start-genesisWhen this process is complete, the terminal will display metagraph information, including the URLs for each metagraph node layer.
Test sending custom data to the deployed metagraph using the included send_data_transaction.js script. This script creates dummy image record data, cryptographically signs it, and sends the structured data to the metagraph's data L1 layer to be processed.
From the euclid-development-environment directory, navigate into the testing_scripts directory:
cd source/project/autosight/testing_scriptsInstall the required Node JS packages:
npm iIn a code editor, open the send_data_transaction.js file, and update the following placeholder values, saving the file when complete:
:your_global_l0_node_url: A metagraph node's Global L0 URL, copied from the terminal output. (ex.http://localhost:9000):your_metagraph_l1_data_url: A metagraph node's Data L1 URL, copied from the terminal output. (ex.http://localhost:9400):your_wallet_address: A valid DAG wallet address, obtainable from the Stargazer wallet.
Run the test script:
node send_data_transaction.jsThe script will output the hash value returned by the metagraph after successfully validating and combining the dummy data. Reward tokens are minted for the provided wallet address in the script.
In a browser, navigate to the following custom endpoints to view the metagraph's state:
<:your_metagraph_L0_url>/data-application/images: Returns a list of image records, stored in the metagraph's on chain state. (Ex.http://localhost:9200/data-application/images)<:your_metagraph_L0_url>/data-application/pending-rewards: Returns a list of pending rewards records, stored in the metagraph's calculated state. (Ex.http://localhost:9200/data-application/pending-rewards)<:your_metagraph_L0_url>/data-application/rewards: Returns a list of addresses and their total minted tokens, stored in the metagraph's calculated state. (Ex.http://localhost:9200/data-application/rewards)
After testing is complete, restart the metagraph from its genesis snapshot to remove the dummy data:
scripts/hydra stop
scripts/hydra start-genesisClone the autosight code repository to your local machine:
git clone https://github.com/rhammell/autosight.gitIn a new terminal window, navigate into the cloned autosight directory, then into the image_upload subdirectory:
cd autosight
cd image_uploadInstall the required Python packages:
pip3 install -r requirements.txtLaunch the server:
python3 server.pyThe server runs on the default port (5050) and is now available at http://localhost:5050.
The AutoSight Android app relies on connections to the metagraph nodes and image upload server. However, since the app is intended to be used while in a vehicle, the phone will not be connected to the same local network as the host machine running the metagraph and image server, and must access them through public URLs.
This can be accomplished by either deploying the metagraph and image server to a cloud service provider, or enabling port forwarding on the host machine's local network router.
If using port forwarding, configure your router to forward inbound requests from its public IP address to the appropriate ports on your local machine. For example:
http://<public-ip>:5050→http://localhost:5050(image server)http://<public-ip>:9200→http://localhost:9200(metagraph L0)http://<public-ip>:9400→http://localhost:9400(metagraph Data L1)
Open Android Studio, select File → Open, select the android_app/AutoSight directory from the autosight cloned repository.
In the upper left corner of the Project tool window, click the 'Project' dropdown menu, and select 'Android' from the options.
Open the AppConfig file within the code editor by double clicking it in the file tree: app/kotlin+java/com.example.autosight/AppConfig
Update the following placeholder values in AppConfig, and save the file when complete:
IMAGE_SERVER_ADDRESS: Public URL of image server (ex.http://<public-ip>:5050)METAGRAPH_L0_ADDRESS: Public URL of metagraph L0 layer (ex.http://<public-ip>:9200)METAGRAPH_DATA_L1_ADDRESS: Public URL of metagraph Data L1 layer (ex.http://<public-ip>:9400)
Synch the project by selecting File → Sync Project with Gradle Files. Then, build the project by selecting Build → Make Project.
Connect an Android device to the development computer via USB, then allow USB debugging on the phone when prompted.
Select Run → Select Device, and select the connected device.
Select Run → Run 'app' to install and run the AutoSight app on the connected Android phone. The app will remain installed on the device after it is disconnected.
In a new terminal window, navigate into the cloned autosight directory, then into the web_app/autosight-explorer subdirectory:
cd autosight
cd web_app/autosight-explorerInstall the required Node JS packages:
npm iIn a code editor, open the .env file, and update the following placeholder value, saving the file when complete:
REACT_APP_AUTOSIGHT_METAGRAPH_LO_IMAGES_URL: The/data-application/imagesendpoint of a metagraph node's Metagraph L0 URL (ex.http://localhost:9200/data-application/images)
Start the development server:
npm startThe AutoSight web application viewable by browsing to the default server address http://localhost:3000.