🧑Step 3: Monitoring the Metrics on the Lens AI Server.

Lens AI server is used for monitoring the data and model drift and access the sampled data.

Lens AI server requires data in the following format. Data need to be aggregated from the sensors in the given directory format. The next release includes support to HTTP/MQTT with a DB. Current version is built using Streamlit.

The format required for the server to visualize the model and data drift.

Step 1: Transfer the lensai directory from the sensor to the server using your own customized method , next releases will support synchronization.

lensai_sensor1
│   ├── imagestats/
│   │   ├── BRIGHTNESS_<timestamp>.png
│   │   ├── NOISE_<timestamp>.png
│   │   ├── SHARPNESS_<timestamp>.png
│   │   ├── MEAN_<metric>_<timestamp>.png
│   │   ├── BRIGHTNESS.bin
│   │   ├── NOISE.bin
│   │   ├── SHARPNESS.bin
│   │   ├── MEAN_<metric>.bin
│   ├── modelstats/
│   │   ├── model_<metric>_<timestamp>.bin
│   │   ├── model_<metric>_<timestamp>.png
│   ├── samples/
│   │   ├── model_<metric>_<timestamp>.bin
│   │   ├── model_<metric>_<timestamp>.png
│   │   ├── model_<metric>_<timestamp>.bin
│   │   ├── model_<metric>_<timestamp>.png
lensai_sensor2
.....

Step 2: Create a directory for every sensor using a sensor Id for example sensor 1 , sensor 2.

mkdir data
cd data
mkdir sensor1
mkdir sensor2
....

Step 3: Create a directory under the sensor directory using the following cmd

cd sensor1 
mkdir "$(date +%s)"
cd sensor2
mkdir "$(date +%s)"

Step 4: Copy the lensai directory content in to it.

cp -r lensai_sensor1/* data/sensor1/<timestamp>/
cp -r lensai_sensor2/* data/sensor2/<timestamp>/
cp -r lensai_sensor3/* data/sensor3/<timestamp>/

after the above step the data folder will have the following structure

data
├── sensor1
│   ├── <timestamp>
│   │   ├── imagestats/
│   │   │   ├── BRIGHTNESS_<timestamp>.png
│   │   │   ├── NOISE_<timestamp>.png
│   │   │   ├── SHARPNESS_<timestamp>.png
│   │   │   ├── MEAN_<metric>_<timestamp>.png
│   │   │   ├── BRIGHTNESS.bin
│   │   │   ├── NOISE.bin
│   │   │   ├── SHARPNESS.bin
│   │   │   ├── MEAN_<metric>.bin
│   │   ├── modelstats/
│   │   │   ├── model_<metric>_<timestamp>.bin
│   │   │   ├── model_<metric>_<timestamp>.png
│   │   ├── samples/
│   │   │   ├── model_<metric>_<timestamp>.bin
│   │   │   ├── model_<metric>_<timestamp>.png
│   │   │   ├── model_<metric>_<timestamp>.bin
│   │   │   ├── model_<metric>_<timestamp>.png
├── sensor2
│   ├── <timestamp>
│   │   ├── imagestats/
│   │   │   ├── BRIGHTNESS_<timestamp>.png
│   │   │   ├── NOISE_<timestamp>.png
│   │   │   ├── SHARPNESS_<timestamp>.png
│   │   │   ├── MEAN_<metric>_<timestamp>.png
│   │   │   ├── BRIGHTNESS.bin
│   │   │   ├── NOISE.bin
│   │   │   ├── SHARPNESS.bin
│   │   │   ├── MEAN_<metric>.bin
│   │   ├── modelstats/
│   │   │   ├── model_<metric>_<timestamp>.bin
│   │   │   ├── model_<metric>_<timestamp>.png
│   │   ├── samples/
│   │   │   ├── model_<metric>_<timestamp>.bin
│   │   │   ├── model_<metric>_<timestamp>.png
│   │   │   ├── model_<metric>_<timestamp>.bin
│   │   │   ├── model_<metric>_<timestamp>.png
│── sensor3 

download the LensAI monitoring server from the following repository

git clone https://github.com/lens-ai/lensai_server

Install the requirements

pip install -r requirements.txt

Run the streamlit server by passing the data folder as an argument

streamlit run lensai_sever.py -- /data