Merged the documentation additions and most recent makefile modification (branch...

[dead/census-tools.git] / makefile

diff --git a/makefile b/makefile
index f3136a6cda47c33afcb4d13b3d9a8cde272ae5ff..5384ded55875ebdf6de7e85eb0a656d0191aba41 100644 (file)
--- a/makefile
+++ b/makefile
@@ -1,9 +1,92 @@
-.PHONY : test
+DB_NAME='census'
+DB_USER='postgres'
+TIGER_DATA_URL='http://www2.census.gov/cgi-bin/shapefiles/multi-file-download?files=24_MARYLAND%2Ftl_2008_24_tabblock00.zip'
+TIGER_SRID='4269'
 
+# Starting with PostGIS 1.4.0, these paths are calculated at install
+# time using the pg_config utility. Rather than try to guess where
+# PostGIS will wind up installed, we can just check the output of
+# pg_config ourselves.
+PG_BINDIR=`pg_config --bindir`
+PG_SHAREDIR=`pg_config --sharedir`
+
+# Necessary to run test/data without prerequisites.
+#
+.PHONY : test data
+
+
+# The default task, since it comes first in the list.
+#
 all: clean test
 
+
 test:
        ./bin/run_tests
 
+
+# Remove byte-compiled python code.
+#
 clean:
        find ./ -name '*.pyc' -print0 | xargs -0 rm -f
+
+
+# Download the shapefiles from Tiger if they don't already exist.
+# For now, we're only dealing with the Census 2000 Maryland Block
+# data, so the filenames are hard-coded. Easy enough to change.
+#
+data:
+       mkdir -p data/census-2000-block/maryland/
+       if [ ! -f data/census-2000-block/maryland/tl_2008_24_tabblock00.shp ]; then \
+               wget -O tiger.zip $(TIGER_DATA_URL); \
+               unzip tiger.zip; \
+               rm tiger.zip; \
+               unzip srv/ftp/geo/tiger/TIGER2008/24_MARYLAND/tl_2008_24_tabblock00.zip \
+                      -d ./data/census-2000-block/maryland/; \
+               rm -rf srv; \
+       fi;
+
+
+# This task does a couple of things. First, it drops and re-creates
+# the DB_NAME database (or schema, whatever). Then, it adds PL/pgSQL
+# support to the database.
+#
+# At that point, we import the two PostGIS files, postgis.sql and
+# spatial_ref_sys.sql. The postgis.sql file contains the geometry
+# functions, while spatial_ref_sys.sql contains a table of SRIDs, and
+# their associated properties. PostGIS requires both.
+#
+# Then, we import the Tiger data using shp2pgsql. The shapefiles
+# should exist, since this task depends on the "data" task, which
+# downloads said shapefiles.
+#
+# Finally, we create the table for the demographic data (obtained from
+# the geographic header records), and populate that table with the output
+# of the sf1blocks2sql script.
+#
+db: data
+       # Ignore the result of dropdb when it fails.
+       dropdb -U $(DB_USER) $(DB_NAME) || true
+       createdb -U $(DB_USER) $(DB_NAME)
+       createlang -U $(DB_USER) plpgsql $(DB_NAME)
+
+       psql -d $(DB_NAME) \
+             -U $(DB_USER) \
+             -f $(PG_SHAREDIR)/contrib/postgis.sql
+
+       psql -d $(DB_NAME) \
+             -U $(DB_USER) \
+             -f $(PG_SHAREDIR)/contrib/spatial_ref_sys.sql
+
+       $(PG_BINDIR)/shp2pgsql -I                                           \
+                 -s $(TIGER_SRID)                                          \
+                 data/census-2000-block/maryland/tl_2008_24_tabblock00.shp \
+                 tiger                                                     \
+                  | psql -U $(DB_USER) -d $(DB_NAME)
+
+       psql -d $(DB_NAME) \
+             -U $(DB_USER) \
+             -f sql/create-sf1_blocks-table.sql
+
+       bin/sf1blocks2sql src/Tests/Fixtures/SummaryFile1/mdgeo.uf1 sf1_blocks \
+                          | psql -U postgres -d $(DB_NAME)
+