Add pg_dump step, improve Sqitch deployment, improve docs

Author: rustprooflabs

db/deploy/001.sql

@@ -292,8 +292,12 @@ COMMENT ON FUNCTION geofaker.n_points_in_polygon(GEOMETRY, INT) IS 'Creates N po
 
 
 
--- Call the procedure to ensure the required temp table exists, avoids deploy failure
-CALL geofaker.point_in_place_landuse();
+-- Ensure the required temp table exists, avoids deploy failure on next sproc
+CREATE TEMP TABLE IF NOT EXISTS faker_store_location
+(
+	store_id BIGINT, place_osm_id BIGINT, place_osm_type TEXT, place_name TEXT,
+	road_osm_id BIGINT, geom GEOMETRY
+);
 
 
 CREATE PROCEDURE geofaker.points_around_point()

docs/src/customize.md

@@ -5,23 +5,6 @@ This section builds on the [Quick Start](quick-start.md) section.
 > Warning: This project is in early development!  Things will be changing over the first few releases (e.g. before 0.5.0).
 
 
-## External Postgres connections
-
-The quick start uses the built-in Postgres/PostGIS instance. See
-the PgOSM Flex section on [Using external Postgres connections](https://pgosm-flex.com/postgres-external.html) to use your own Postgres instance.
-This approach does load a lot of data to the target database which may not be
-desired.  Consider using `pg_dump` to load only the target data to your
-database of choice.
-
-The Sqitch deployment step should use additional parameters not set in the quick start
-instructions.
-
-```bash
-source ~/.pgosm-faker-local
-cd pgosm-flex-faker/db
-sqitch db:pg://$POSTGRES_USER:$POSTGRES_PASSWORD@$POSTGRES_HOST:$POSTGRES_PORT/$POSTGRES_DB deploy
-```
-
 
 ## Each time is new data
 
@@ -54,3 +37,13 @@ UNION
 SELECT 'vineyard' AS osm_type
 ;
 ```
+
+
+## External Postgres connections
+
+Geo Faker currently does not support running directly into an external database.
+Technically it can be operated that way, but the instructions and helper scripts
+are specific to using the in-Docker database.
+
+Use `--pg-dump` to extract the generated data.
+

docs/src/quick-start.md

@@ -49,14 +49,18 @@ docker exec -it \
 ## Load and Run Faker Objects
 
 After the data completes processing, load the Geo Faker database structures
-in the `geofaker` schema.
-This is done using Sqitch.
+in the `geofaker` schema.  This deploys the functions and procedures needed,
+runs the processing, and runs `pg_dump` saving the faked
+data into `~/pgosm-data/geofaker_stores_customers.sql`.
 
 
 ```bash
 docker exec -it geofaker /bin/bash run_faker.sh
 ```
 
-Connect to the database and call this stored procedure.  The generated data
-is left in a temp table, each run of the stored procedure will produce new,
-random results.
+Load the saved data into a database of your choice.
+
+```bash
+psql -d pgosm_faker -f ~/pgosm-data/geofaker_stores_customers.sql 
+```
+

run_faker.sh

@@ -1,8 +1,12 @@
 #!/bin/bash
 echo 'Deploying geofaker schema via Sqitch...'
 cd /app/faker/db
-sqitch db:pg://$POSTGRES_USER:$POSTGRES_PASSWORD@localhost:5432/pgosm deploy
+sqitch db:pg://$POSTGRES_USER:$POSTGRES_PASSWORD@localhost/pgosm deploy
 
 echo 'Running GeoFaker generation...'
 psql -d postgres://$POSTGRES_USER:$POSTGRES_PASSWORD@localhost/pgosm \
     -f /app/run_faker.sql
+
+echo 'Running pg_dump...'
+pg_dump -d postgres://$POSTGRES_USER:$POSTGRES_PASSWORD@localhost/pgosm \
+    -n geofaker -f /app/output/geofaker_stores_customers.sql