visionaire4 0.1.0

Visionaire4 maintenance tools suite

VISIONAIRE4

A suite of tools used for Visionaire4 maintenance.
Supported function subcommand:

• report: Generate Visionaire4 and FRemisN status report.
• export: Export metrics and edge cases data from current machine to files.
• import: Import metrics and edge cases data from the exported files to current machine.

Installation

Install it directly using pip:

$ pip install visionaire4 --upgrade

How to Use

Export

Metrics

Before running export metrics make sure to have prometheus server to be running.
Export monitoring metrics data by running:

$ visionaire4 export metrics

This command will pull metrics data from prometheus server running in container named visionaire4.
You can configure the container name where the prometheus server is running with:

$ visionaire4 export metrics --prom-name <container name>

This will export the file to the default directory of ~/nodeflux/export.
You can configure where the exported file output directory with:

$ visionaire4 export metrics --out-dir <output directory>

Edge Case

Before running export edge case data make sure to have visionaire4 to be running.
Export edge case data by running:

$ visionaire4 export edge

This command will pull edge case data from visionaire4 container running in container named visionaire4.
You can configure the container name where the prometheus server is running with:

$ visionaire4 export edge --v4-name <container name>

Import

Metrics

Import the exported metrics data to current machine by running:

$ visionaire4 import metrics -f <path to exported file>

This will spawn a prometheus and grafana server and you can view the grafana dashboards at http://localhost:3000 with username admin and password admin.

All the required configs to run the prometheus and grafana server will be generated to the default directory of ~/nodeflux/import.
You can configure the config directory with:

$ visionaire4 import metrics -f <path to exported file> --cfg-dir <config directory>

When you finished viewing the metrics data, you can shutdown the monitoring server by running:

$ visionaire4 import metrics --cfg-dir <config directory> --down

Edge Case

Import the exported edge case data to current machine by running:

$ visionaire4 import edge -f <path to exported file>

This will extract edge case data and separate image data from the label into default config directory ~/nodeflux/import.
You can configure the config directory with:

$ visionaire4 import edge -f <path to exported file> --cfg-dir <config directory>

By default, the data will be dumped as v4 original annotation structure. Optionally, output format can be changed to other common dataset format, i.e. coco (currently only support coco), as:

$ visionaire4 import edge -f <path to exported file> --cfg-dir <config directory> --out-fmt coco

Report

To generate status report for Visionaire4 and or FRemisN, run:

$ visionaire4 report

This will try to locate a running Visionaire4 and or FRemisN, check the system status, and generate a summary for it. If it can't find either Visionaire4 or FRemisN container instance it will throw an error, so make sure to have a valid running instance of either one of it.
If somehow you have the valid instance running and the tool can't find it, you can specify the container name using:

$ visionaire4 report --v4-name <visionaire4 container name> --fr-name <fremisn container name>

If you have a custom Visionaire4 address, specify it using:

$ visionaire4 report --v4-address <IP:PORT>

By default the report will save the last 1000 line of logs for each container, you specify it using:

$ visionaire4 report --tail <N>
# or
$ visionaire4 report -n <N>

By default the report will also plot the last 5 minutes (300 seconds) of useful metrics from prometheus server, you can specify the time range using:

$ visionaire4 report --range <SECONDS>
# or
$ visionaire4 report -r <SECONDS>