- Apache Spark
Clone the repository
git clone https://github.com/mami-project/pto-core.git
and install using pip:
pip install -e pto-core/
For running multiple instances of different versions in parallel you should install pto-core in seperate python virtual environments.
The programs are launched with an arbitrary number of paths to json files.
All json file are automatically merged togheter (the latter overwriting values of the former).
It is advised to have a base config
base.json and an environment-specific config file for each environment.
See conf/base.json for an example base configuration file.
Each instance needs to run on it's own databases. A configuration file can be created with the following
--help for more information about options). In addition it prints out the necessary MongoDB
commands for creating the database users and their permissions. Where
NAME is the name you want to give the new environment and
PATH is an absolute path the analyzer modules will be stored. See a lot more options by running it with
Using ptocore-createconfig is not mandatory, see conf/prod.json for an example how it will look like.
ptocore-createconfig <NAME> <PATH>
The observatory core consists of four python programs each of them run separately and don't have direct connections to each other (only through the database). You should make sure that only one instance of each program runs with the same configuration.
ptocore-sensor <base>.json <env.json> ptocore-supervisor <base>.json <env.json> ptocore-validator <base>.json <env.json>
The administrative RESTful API is powered by flask. A standalone server is started with the following lines. The first line is only needed if you use a python virtual environment and want to start the service with a shell script. (Don't forget to change paths according to your needs.)
source venv/bin/activate export FLASK_APP=ptocore.admin export FLASK_DEBUG=1 export PTO_CONFIG_FILES=/path/to/conf/base.json:/path/to/conf/env.json flask run --host=0.0.0.0 --port=33525