X-Git-Url: http://gitweb.michael.orlitzky.com/?p=dead%2Fcensus-tools.git;a=blobdiff_plain;f=makefile;h=2631b6b11a8ed60dde2a492ca0f2db83bb2879b4;hp=2ab9dbae7b627e27f927061b3dda062abcfd43b0;hb=HEAD;hpb=937b2d7b201b540fcfec7ff64d3838835f690b87 diff --git a/makefile b/makefile index 2ab9dba..2631b6b 100644 --- a/makefile +++ b/makefile @@ -2,237 +2,184 @@ DB_NAME=census DB_USER=postgres TIGER_SRID=4269 -# Root folder for the shapefiles. -TIGER_ROOT=http://www2.census.gov/geo/tiger/TIGER2009 - -# State-specific folders. -DC_ROOT=$(TIGER_ROOT)/11_DISTRICT_OF_COLUMBIA -MD_ROOT=$(TIGER_ROOT)/24_MARYLAND -VA_ROOT=$(TIGER_ROOT)/51_VIRGINIA -PA_ROOT=$(TIGER_ROOT)/42_PENNSYLVANIA -NY_ROOT=$(TIGER_ROOT)/36_NEW_YORK - -# URLs for the TIGER/Line block-level shapefiles. -DC_BLOCKS_URL=$(DC_ROOT)/tl_2009_11_tabblock00.zip -MD_BLOCKS_URL=$(MD_ROOT)/tl_2009_24_tabblock00.zip -VA_BLOCKS_URL=$(VA_ROOT)/tl_2009_51_tabblock00.zip -PA_BLOCKS_URL=$(PA_ROOT)/tl_2009_42_tabblock00.zip -NY_BLOCKS_URL=$(NY_ROOT)/tl_2009_36_tabblock00.zip - -# URLs for the DC county all-lines shapefiles. -# D.C. just has one Census county. -DC_LINES_URL=$(DC_ROOT)/11001_District_of_Columbia/tl_2009_11001_edges.zip - -# This is where it gets ugly. -# -# URLs for the MD county all-lines shapefiles. -ALLEGANY_LINES_URL=$(MD_ROOT)/24001_Allegany_County/tl_2009_24001_edges.zip -ANNE_ARUNDEL_LINES_URL=$(MD_ROOT)/24003_Anne_Arundel_County/tl_2009_24003_edges.zip -BALTIMORE_CO_LINES_URL=$(MD_ROOT)/24005_Baltimore_County/tl_2009_24005_edges.zip -BALTIMORE_CI_LINES_URL=$(MD_ROOT)/24510_Baltimore_city/tl_2009_24510_edges.zip -CALVERT_LINES_URL=$(MD_ROOT)/24009_Calvert_County/tl_2009_24009_edges.zip -CAROLINE_LINES_URL=$(MD_ROOT)/24011_Caroline_County/tl_2009_24011_edges.zip -CARROLL_LINES_URL=$(MD_ROOT)/24013_Carroll_County/tl_2009_24013_edges.zip -CECIL_LINES_URL=$(MD_ROOT)/24015_Cecil_County/tl_2009_24015_edges.zip -CHARLES_LINES_URL=$(MD_ROOT)/24017_Charles_County/tl_2009_24017_edges.zip -DORCHESTER_LINES_URL=$(MD_ROOT)/24019_Dorchester_County/tl_2009_24019_edges.zip -FREDERICK_LINES_URL=$(MD_ROOT)/24021_Frederick_County/tl_2009_24021_edges.zip -GARRETT_LINES_URL=$(MD_ROOT)/24023_Garrett_County/tl_2009_24023_edges.zip -HARFORD_LINES_URL=$(MD_ROOT)/24025_Harford_County/tl_2009_24025_edges.zip -HOWARD_LINES_URL=$(MD_ROOT)/24027_Howard_County/tl_2009_24027_edges.zip -KENT_LINES_URL=$(MD_ROOT)/24029_Kent_County/tl_2009_24029_edges.zip -MONTGOMERY_LINES_URL=$(MD_ROOT)/24031_Montgomery_County/tl_2009_24031_edges.zip -PRINCE_GEORGES_LINES_URL=$(MD_ROOT)/24033_Prince_Georges_County/tl_2009_24033_edges.zip -QUEEN_ANNES_LINES_URL=$(MD_ROOT)/24035_Queen_Annes_County/tl_2009_24035_edges.zip -ST_MARYS_LINES_URL=$(MD_ROOT)/24037_St_Marys_County/tl_2009_24037_edges.zip -SOMERSET_LINES_URL=$(MD_ROOT)/24039_Somerset_County/tl_2009_24039_edges.zip -TALBOT_LINES_URL=$(MD_ROOT)/24041_Talbot_County/tl_2009_24041_edges.zip -WASHINGTON_LINES_URL=$(MD_ROOT)/24043_Washington_County/tl_2009_24043_edges.zip -WICOMICO_LINES_URL=$(MD_ROOT)/24045_Wicomico_County/tl_2009_24045_edges.zip -WORCESTER_LINES_URL=$(MD_ROOT)/24047_Worcester_County/tl_2009_24047_edges.zip - - -# 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` + +# Dark magic. We set these makefile variables to be the result of the +# 'shell' function. The shell function, in turn, executes a Python +# script which determines the locations of these files. +SHP2PGSQL := $(shell bin/find_file_paths --root /usr --single shp2pgsql) +POSTGIS_SQL := $(shell bin/find_file_paths --root /usr lwpostgis.sql postgis.sql) +SPATIAL_REF_SYS_SQL := $(shell bin/find_file_paths --root /usr spatial_ref_sys.sql) # Necessary to run test/data without prerequisites. # -.PHONY : test data +.PHONY : test data lib # The default task, since it comes first in the list. # -all: clean test +all: clean lib test test: ./bin/run_tests +# Download or check out any third-party libraries. +lib: + make -C lib/ + + # 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: + bin/download_data + + +# There is a small issue here with the blocks_db and lines_db +# targets. Each of these requires that the database exists, and might +# therefore depend on the newdb target. However, if /each/ of them +# depends on newdb, the database will be dropped twice and the data +# from one of {blocks, lines} would be lost. # -data: tiger_blocks tiger_lines - -tiger_blocks: dc_blocks md_blocks va_blocks pa_blocks ny_blocks - -dc_blocks: - mkdir -p data/census2000/dc/block - if [ ! -f data/census2000/dc/block/tl_2009_11_tabblock00.shp ]; \ - then \ - wget -O dcblocks.zip $(DC_BLOCKS_URL); \ - unzip dcblocks.zip -d ./data/census2000/dc/block; \ - rm dcblocks.zip; \ - fi; - -md_blocks: - mkdir -p data/census2000/maryland/block - if [ ! -f data/census2000/maryland/block/tl_2009_24_tabblock00.shp ]; \ - then \ - wget -O mdblocks.zip $(MD_BLOCKS_URL); \ - unzip mdblocks.zip -d ./data/census2000/maryland/block; \ - rm mdblocks.zip; \ - fi; - -va_blocks: - mkdir -p data/census2000/virginia/block - if [ ! -f data/census2000/virginia/block/tl_2009_51_tabblock00.shp ]; \ - then \ - wget -O vablocks.zip $(VA_BLOCKS_URL); \ - unzip vablocks.zip -d ./data/census2000/virginia/block; \ - rm vablocks.zip; \ - fi; - -pa_blocks: - mkdir -p data/census2000/pennsylvania/block - if [ ! -f data/census2000/pennsylvania/block/tl_2009_42_tabblock00.shp ]; \ - then \ - wget -O pablocks.zip $(PA_BLOCKS_URL); \ - unzip pablocks.zip -d ./data/census2000/pennsylvania/block; \ - rm pablocks.zip; \ - fi; - -ny_blocks: - mkdir -p data/census2000/new_york/block - if [ ! -f data/census2000/new_york/block/tl_2009_36_tabblock00.shp ]; \ - then \ - wget -O nyblocks.zip $(NY_BLOCKS_URL); \ - unzip nyblocks.zip -d ./data/census2000/new_york/block; \ - rm nyblocks.zip; \ - fi; - - -tiger_lines: dc_lines - -dc_lines: - mkdir -p data/census2000/dc/lines - if [ ! -f data/census2000/dc/lines/tl_2009_11001_edges.shp ]; \ - then \ - wget -O dclines.zip $(DC_LINES_URL); \ - unzip dclines.zip -d ./data/census2000/dc/lines; \ - rm dclines.zip; \ - 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. +# We therefore assume that the database already exists when blocks_db +# or lines_db are initiated. +blocks_db: data blocks_table +# All Blocks # -# 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. +# The table already exists, so we can append to it, and we don't have +# to create the GiST index. + for state in data/census2000/*; do \ + $(SHP2PGSQL) \ + -a \ + -s $(TIGER_SRID) \ + -D \ + $$state/blocks/*.shp \ + tiger_blocks \ + | psql -U $(DB_USER) -d $(DB_NAME); \ + done; + +# Summary File 1 # -# Then, we import the Tiger data using shp2pgsql. The shapefiles +# Run all of the geo (uf1) files through the import script. This has +# to happen after the blocks import since we impose a foreign key +# restriction. + for state in data/census2000/*; do \ + bin/sf1blocks2sql $$state/sf1/*.uf1 sf1_blocks \ + | psql -U $(DB_USER) -d $(DB_NAME) \ + > /dev/null; \ + done; + +# Run the query to combine the two blocks tables, and drop the +# constituents. + psql -U $(DB_USER) \ + -d $(DB_NAME) \ + -f sql/combine-block-tables.sql + + +lines_db: data tiger_lines_table +# All Lines +# +# Since the table and index already exist, we can utilize -a, +# and leave -I out. + for state in data/census2000/*; do \ + for shapefile in $$state/lines/*.shp; do \ + echo "Importing $$shapefile."; \ + $(SHP2PGSQL) \ + -a \ + -s $(TIGER_SRID) \ + $$shapefile \ + tiger_lines \ + | bin/filter-transactions \ + | psql -U $(DB_USER) -d $(DB_NAME) \ + > /dev/null; \ + done; \ + done; + + + +# This imports 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. +# After the TIGER import is done, we use the sf1blocks2sql script to +# parse and import the geographic header record information. # -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 - -# D.C. Blocks - - $(PG_BINDIR)/shp2pgsql \ - -I \ - -s $(TIGER_SRID) \ - -D \ - data/census2000/dc/block/tl_2009_11_tabblock00.shp \ - tiger_blocks \ - | psql -U $(DB_USER) -d $(DB_NAME) - -# Maryland Blocks +db: newdb blocks_data lines_data + # Do nothing except fulfill our prerequisites. - $(PG_BINDIR)/shp2pgsql \ - -I \ - -s $(TIGER_SRID) \ - -D \ - data/census2000/maryland/block/tl_2009_24_tabblock00.shp \ - tiger_blocks \ - | psql -U $(DB_USER) -d $(DB_NAME) -# Virginia Blocks - - $(PG_BINDIR)/shp2pgsql -a \ - -I \ - -s $(TIGER_SRID) \ - -D \ - data/census2000/virginia/block/tl_2009_51_tabblock00.shp \ - tiger_blocks \ - | psql -U $(DB_USER) -d $(DB_NAME) - - -# Pennsylvania Blocks - - $(PG_BINDIR)/shp2pgsql -a \ - -I \ - -s $(TIGER_SRID) \ - -D \ - data/census2000/pennsylvania/block/tl_2009_42_tabblock00.shp \ - tiger_blocks \ - | psql -U $(DB_USER) -d $(DB_NAME) - - -# New York Blocks - - $(PG_BINDIR)/shp2pgsql -a \ - -I \ - -s $(TIGER_SRID) \ - -D \ - data/census2000/new_york/block/tl_2009_36_tabblock00.shp \ - tiger_blocks \ - | 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) > /dev/null +# First, we drop and re-create the DB_NAME database (or schema, +# whatever). Then, we add 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. +# +newdb: +# 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 $(POSTGIS_SQL) \ + > /dev/null + + psql -d $(DB_NAME) \ + -U $(DB_USER) \ + -f $(SPATIAL_REF_SYS_SQL) \ + > /dev/null + + +# This just runs the SQL script to create the sf1_blocks table. +sf1_blocks_table: + psql -d $(DB_NAME) \ + -U $(DB_USER) \ + -f sql/create-sf1_blocks-table.sql \ + > /dev/null + + +# Create the tiger_blocks table, and create its GiST index. Having the +# table already exist makes importing via shp2pgsql much easier. +# Any blocks file will work as an argument. +tiger_blocks_table: + $(SHP2PGSQL) \ + -p \ + -I \ + -s $(TIGER_SRID) \ + data/census2000/maryland/blocks/tl_2009_24_tabblock00.shp \ + tiger_blocks \ + | psql -U $(DB_USER) -d $(DB_NAME) \ + > /dev/null + +# Create the "blocks" table, which is the result of joining +# the tiger_blocks and sf1_blocks tables. +blocks_table: tiger_blocks_table sf1_blocks_table + psql -U $(DB_USER) \ + -d $(DB_NAME) \ + -f sql/create-blocks-table.sql + + +# Prepare the tiger_lines table, and create the GiST index on its +# geometry column. Any lines shapefile will do here. +tiger_lines_table: + $(SHP2PGSQL) \ + -p \ + -I \ + -s $(TIGER_SRID) \ + data/census2000/maryland/lines/tl_2009_24510_edges.shp \ + tiger_lines \ + | psql -U $(DB_USER) -d $(DB_NAME) \ + > /dev/null + +# Add a unique index on the "tlid" column. + psql -U $(DB_USER) \ + -d $(DB_NAME) \ + -f sql/create_tlid_unique_index.sql