gtfierro/shepherding-metadata

Shepherding Metadata

This repository implements the platform described in our BuildSys 2020 submission. The current code is of "research-quality" with more than a few sharp edges, so be wary!

Getting Started

1. Install the Python requirements in requirements.txt

2. Drivers are run as standalone processes, and are configured with a yml file. Use the run_driver.py file to start a driver; this takes a single argument which is a path to a config file:

   $ python run_driver.py cfgs/bsync_carytown.toml
   2020-11-12:14:12:58,430 INFO    [driver.py:23] Setting up Flask routes
   2020-11-12:14:12:58,431 INFO    [driver.py:29] INITIALIZED
   2020-11-12:14:12:58,431 INFO    [driver.py:63] Push to http://localhost:6483
   2020-11-12:14:12:58,432 INFO    [driver.py:98] Starting driver webserver on port http://localhost:8081
    * Serving Flask app "driver" (lazy loading)
   2020-11-12:14:12:58,432 INFO    [bsync_driver.py:66] Loading BuildingSync file data/buildingsync/examples/bsync-carytown.xml
   2020-11-12:14:12:58,434 INFO    [_internal.py:113]  * Running on http://localhost:8081/ (Press CTRL+C to quit)
   2020-11-12:14:12:58,437 INFO    [bsync_driver.py:130] Loaded 4 records
   

   You can now visit http://localhost:8081/ids to see the entities inferred by the driver; visit http://localhost:8081/id/<identifier> to see the record for a particular entity, including any inferred Brick metadata. (Example: http://localhost:8081/id/Site-1

3. To run the integration server, run python server.py; this will create a SQLite store triples.db which stores the inferred Brick model. Read the source code for available API endpoints.