Konstantin Gründger 2020-05-15 18:40:46 +02:00
commit 29e112eaca
1 zmienionych plików z 20 dodań i 13 usunięć

Wyświetl plik

@ -16,43 +16,50 @@ It requires [PostgreSQL](http://www.postgresql.org/), [PostGIS](http://www.postg
```
git clone https://github.com/glidernet/ogn-python.git
cd ogn-python
```
2. Install python requirements
2. Optional: Create and use a virtual environment
```
python3 -m venv my_environment
source my_environment/bin/activate
```
3. Install python requirements
```
pip install -r requirements.txt
```
3. Install [PostgreSQL](http://www.postgresql.org/) with [PostGIS](http://www.postgis.net/) and [TimescaleDB](https://www.timescale.com) Extension.
4. Install [PostgreSQL](http://www.postgresql.org/) with [PostGIS](http://www.postgis.net/) and [TimescaleDB](https://www.timescale.com) Extension.
Create a database (use "ogn" as default, otherwise you have to modify the configuration, see below)
4. Optional: Install redis for asynchronous tasks (like takeoff/landing-detection)
5. Optional: Install redis for asynchronous tasks (like takeoff/landing-detection)
```
apt-get install redis-server
```
5. Optional: Custom configuration file
6. Optional: Custom configuration file
Write a custom configuration file and let the environment variable `OGN_CONFIG_MODULE` point to the configuration file:
```
export OGN_CONFIG_MODULE="config/my_custom_configuration.py"
```
6. Create database
7. Create database
```
./flask database init
```
7. Optional: Prepare tables for TimescaleDB
8. Optional: Prepare tables for TimescaleDB
```
./flask database init_timescaledb
```
8. Optional: Import world border dataset (needed if you want to know the country a receiver belongs to, etc.)
9. Optional: Import world border dataset (needed if you want to know the country a receiver belongs to, etc.)
Get the [World Borders Dataset](http://thematicmapping.org/downloads/world_borders.php) and unpack it.
Then import it into your database (we use "ogn" as database name).
@ -62,14 +69,14 @@ It requires [PostgreSQL](http://www.postgresql.org/), [PostGIS](http://www.postg
psql -d ogn -c "DROP TABLE world_borders_temp;"
```
9. Get world elevation data (needed for AGL calculation)
10. Get world elevation data (needed for AGL calculation)
Sources: There are many sources for DEM data. It is important that the spatial reference system (SRID) is the same as the database which is 4326.
The [GMTED2010 Viewer](https://topotools.cr.usgs.gov/gmted_viewer/viewer.htm) provides data for the world with SRID 4326. Just download the data you need.
For Europe we can get the DEM as GeoTIFF files from the [European Environment Agency](https://land.copernicus.eu/imagery-in-situ/eu-dem/eu-dem-v1.1).
Because the SRID of these files is 3035 and we want 4326 we have to convert them (next step)
10. Optional: Convert the elevation data into correct SRID
11. Optional: Convert the elevation data into correct SRID
We convert elevation from one SRID (here: 3035) to target SRID (4326):
@ -77,25 +84,25 @@ It requires [PostgreSQL](http://www.postgresql.org/), [PostGIS](http://www.postg
gdalwarp -s_srs "EPSG:3035" -t_srs "EPSG:4326" source.tif target.tif
```
11. Import the GeoTIFF into the elevation table:
12. Import the GeoTIFF into the elevation table:
```
raster2pgsql -s 4326 -c -C -I -M -t 100x100 elevation_data.tif public.elevation | psql -d ogn
```
12. Import Airports (needed for takeoff and landing calculation). A cup file is provided under tests:
13. Import Airports (needed for takeoff and landing calculation). A cup file is provided under tests:
```
flask database import_airports tests/SeeYou.cup
```
13. Import DDB (needed for registration signs in the logbook).
14. Import DDB (needed for registration signs in the logbook).
```
flask database import_ddb
```
14. Optional: Use supervisord
15. Optional: Use supervisord
You can use [Supervisor](http://supervisord.org/) to control the complete system. In the directory deployment/supervisor
we have some configuration files to feed the database (ogn-feed), run the celery worker (celeryd), the celery beat
(celerybeatd), the celery monitor (flower), and the python wsgi server (gunicorn). All files assume that