Add comments

CONTRIBUTING.md

@@ -12,7 +12,7 @@ name must be created.
 ### Version numbers
 
 The version of both the SQL extension and the Python package shall
-follow the [Semantic Versioning 2.0](http://semver.org/) guidelines:
+follow the[Semantic Versioning 2.0](http://semver.org/) guidelines:
 
 * When backwards incompatibility is introduced the major number is incremented
 * When functionally is added (in a backwards-compatible manner) the minor number

Makefile

@@ -1,5 +1,5 @@
-EXT_DIR = src/pg
+EXT_DIR = pg
-PYP_DIR = src/py
+PYP_DIR = python
 
 .PHONY: install
 .PHONY: run_tests

README.md

@@ -4,87 +4,11 @@ CartoDB Spatial Analysis extension for PostgreSQL.
 
 ## Code organization
 
-* *doc* documentation
+* *pg* contains the PostgreSQL extension source code
-* *src* source code
+* *python* Python module
-* - *src/pg* contains the PostgreSQL extension source code
+
-* - *src/py* Python module source code
+FIXME: should it be `./extension` and `./lib/python' ?
-* *release* reselesed versions
 
 ## Requirements
 
-* pip, virtualenv, PostgreSQL
+* pip
-
-# Working Process
-
-## Development
-
-Work in `src/pg/sql`, `src/py/crankshaft`;
-use topic branch.
-
-Update local installation with `sudo make install`
-(this will update the 'dev' version of the extension in 'src/pg/')
-
-Run the tests with `PGUSER=postgres make test`
-
-Update extension in working database with
-
-* `ALTER EXTENSION crankshaft VERSION TO 'current';`
-  `ALTER EXTENSION crankshaft VERSION TO 'dev';`
-
-Note: we keep the current development version install as 'dev' always;
-we update through the 'current' alias to allow changing the extension
-contents but not the version identifier. This will fail if the
-changes involve incompatible function changes such as a different
-return type; in that case the offending function (or the whole extension)
-should be dropped manually before the update.
-
-If the extension has not previously been installed in a database
-we can:
-
-Add tests...
-
-* `CREATE EXTENSION crankshaft WITH VERSION 'dev';`
-
-Test
-
-Commit, push, create PR, wait for CI tests, CR, ...
-
-## Release
-
-To release current development version
-(working directory should be clean in dev branch)
-
-(process to be gradually automated)
-
-For backwards compatible changes (no return value, num of arguments, etc. changes...)
-new version number increasing either patch level (no new functionality)
-or minor level (new functionality) => 'X.Y.Z'.
-Update version in src/pg/crankshaft.control
-Copy release/crankshaft--current.sql to release/crankshaft--X.Y.Z.sql
-Prepare incremental downgrade, upgrade scripts....
-
-Python: ...
-
-Install the new release
-
-`make install-release`
-
-Test the new release
-
-`make test-release`
-
-Push the release
-
-Wait for CI tests
-
-Merge into master
-
-Deploy: install extension and python to production hosts,
-update extension in databases (limited to team users, data observatory, ...)
-
-Release manager role: ...
-
-.sql release scripts
-commit
-tests: staging....
-merge, tag, deploy...

pg/.gitignore

@@ -0,0 +1,3 @@
+regression.diffs
+regression.out
+results/

pg/Makefile

@@ -0,0 +1,30 @@
+# Makefile to generate the extension out of separate sql source files.
+# Once a version is released, it is not meant to be changed. E.g: once version 0.0.1 is out, it SHALL NOT be changed.
+
+EXTENSION    = crankshaft
+EXTVERSION   = $(shell grep default_version $(EXTENSION).control | sed -e "s/default_version[[:space:]]*=[[:space:]]*'\([^']*\)'/\1/")
+
+# The new version to be generated from templates
+NEW_EXTENSION_ARTIFACT = $(EXTENSION)--$(EXTVERSION).sql
+
+# DATA is a special variable used by postgres build infrastructure
+# These are the files to be installed in the server shared dir,
+# for installation from scratch, upgrades and downgrades.
+# @see http://www.postgresql.org/docs/current/static/extend-pgxs.html
+DATA =  $(NEW_EXTENSION_ARTIFACT)
+
+SOURCES_DATA_DIR = sql/$(EXTVERSION)
+SOURCES_DATA = $(wildcard sql/$(EXTVERSION)/*.sql)
+
+# The extension installation artifacts are stored in the base subdirectory
+$(NEW_EXTENSION_ARTIFACT): $(SOURCES_DATA)
+	rm -f $@
+	cat $(SOURCES_DATA_DIR)/*.sql >> $@
+
+REGRESS = $(notdir $(basename $(wildcard test/$(EXTVERSION)/sql/*test.sql)))
+TEST_DIR = test/$(EXTVERSION)
+REGRESS_OPTS = --inputdir='$(TEST_DIR)' --outputdir='$(TEST_DIR)'
+
+PG_CONFIG = pg_config
+PGXS := $(shell $(PG_CONFIG) --pgxs)
+include $(PGXS)

pg/crankshaft--0.0.1.sql

@@ -0,0 +1,136 @@
+-- Internal function.
+-- Set the seeds of the RNGs (Random Number Generators)
+-- used internally.
+CREATE OR REPLACE FUNCTION
+_cdb_random_seeds (seed_value INTEGER) RETURNS VOID
+AS $$
+  from crankshaft import random_seeds
+  random_seeds.set_random_seeds(seed_value)
+$$ LANGUAGE plpythonu;
+-- Moran's I
+CREATE OR REPLACE FUNCTION
+  cdb_moran_local (
+      t TEXT,
+  	  attr TEXT,
+  	  significance float DEFAULT 0.05,
+  	  num_ngbrs INT DEFAULT 5,
+  	  permutations INT DEFAULT 99,
+  	  geom_column TEXT DEFAULT 'the_geom',
+  	  id_col TEXT DEFAULT 'cartodb_id',
+      w_type TEXT DEFAULT 'knn')
+RETURNS TABLE (moran FLOAT, quads TEXT, significance FLOAT, ids INT)
+AS $$
+  from crankshaft.clustering import moran_local
+  # TODO: use named parameters or a dictionary
+  return moran_local(t, attr, significance, num_ngbrs, permutations, geom_column, id_col, w_type)
+$$ LANGUAGE plpythonu;
+
+-- Moran's I Local Rate
+CREATE OR REPLACE FUNCTION
+  cdb_moran_local_rate(t TEXT,
+		 numerator TEXT,
+		 denominator TEXT,
+		 significance FLOAT DEFAULT 0.05,
+		 num_ngbrs INT DEFAULT 5,
+		 permutations INT DEFAULT 99,
+		 geom_column TEXT DEFAULT 'the_geom',
+		 id_col TEXT DEFAULT 'cartodb_id',
+		 w_type TEXT DEFAULT 'knn')
+RETURNS TABLE(moran FLOAT, quads TEXT, significance FLOAT, ids INT, y numeric)
+AS $$
+  from crankshaft.clustering import moran_local_rate
+  # TODO: use named parameters or a dictionary
+  return moran_local_rate(t, numerator, denominator, significance, num_ngbrs, permutations, geom_column, id_col, w_type)
+$$ LANGUAGE plpythonu;
+-- Function by Stuart Lynn for a simple interpolation of a value
+-- from a polygon table over an arbitrary polygon
+-- (weighted by the area proportion overlapped)
+-- Aereal weighting is a very simple form of aereal interpolation.
+--
+-- Parameters:
+--   * geom a Polygon geometry which defines the area where a value will be
+--     estimated as the area-weighted sum of a given table/column
+--   * target_table_name table name of the table that provides the values
+--   * target_column column name of the column that provides the values
+--   * schema_name optional parameter to defina the schema the target table
+--     belongs to, which is necessary if its not in the search_path.
+--     Note that target_table_name should never include the schema in it.
+-- Return value:
+--   Aereal-weighted interpolation of the column values over the geometry
+CREATE OR REPLACE
+FUNCTION cdb_overlap_sum(geom geometry, target_table_name text, target_column text, schema_name text DEFAULT NULL)
+  RETURNS numeric AS
+$$
+DECLARE
+	result numeric;
+  qualified_name text;
+BEGIN
+  IF schema_name IS NULL THEN
+    qualified_name := Format('%I', target_table_name);
+  ELSE
+    qualified_name := Format('%I.%s', schema_name, target_table_name);
+  END IF;
+  EXECUTE Format('
+    SELECT sum(%I*ST_Area(St_Intersection($1, a.the_geom))/ST_Area(a.the_geom))
+    FROM %s AS a
+    WHERE $1 && a.the_geom
+  ', target_column, qualified_name)
+  USING geom
+  INTO result;
+  RETURN result;
+END;
+$$ LANGUAGE plpgsql;
+--
+-- Creates N points randomly distributed arround the polygon
+--
+-- @param g - the geometry to be turned in to points
+--
+-- @param no_points - the number of points to generate
+--
+-- @params max_iter_per_point - the function generates points in the polygon's bounding box
+-- and discards points which don't lie in the polygon. max_iter_per_point specifies how many
+-- misses per point the funciton accepts before giving up.
+--
+-- Returns: Multipoint with the requested points
+CREATE OR REPLACE FUNCTION cdb_dot_density(geom geometry , no_points Integer, max_iter_per_point Integer DEFAULT 1000)
+RETURNS GEOMETRY AS $$
+DECLARE
+  extent GEOMETRY;
+  test_point Geometry;
+  width                NUMERIC;
+  height               NUMERIC;
+  x0                   NUMERIC;
+  y0                   NUMERIC;
+  xp                   NUMERIC;
+  yp                   NUMERIC;
+  no_left              INTEGER;
+  remaining_iterations INTEGER;
+  points               GEOMETRY[];
+  bbox_line            GEOMETRY;
+  intersection_line    GEOMETRY;
+BEGIN
+  extent  := ST_Envelope(geom);
+  width   := ST_XMax(extent) - ST_XMIN(extent);
+  height  := ST_YMax(extent) - ST_YMIN(extent);
+  x0 	  := ST_XMin(extent);
+  y0 	  := ST_YMin(extent);
+  no_left := no_points;
+
+  LOOP
+    if(no_left=0) THEN
+      EXIT;
+    END IF;
+    yp = y0 + height*random();
+    bbox_line  = ST_MakeLine(
+      ST_SetSRID(ST_MakePoint(yp, x0),4326),
+      ST_SetSRID(ST_MakePoint(yp, x0+width),4326)
+    );
+    intersection_line = ST_Intersection(bbox_line,geom);
+  	test_point = ST_LineInterpolatePoint(st_makeline(st_linemerge(intersection_line)),random());
+	  points := points || test_point;
+	  no_left = no_left - 1 ;
+  END LOOP;
+  RETURN ST_Collect(points);
+END;
+$$
+LANGUAGE plpgsql VOLATILE

pg/sql/0.0.1/01_random_seeds.sql

@@ -4,7 +4,6 @@
 CREATE OR REPLACE FUNCTION
 _cdb_random_seeds (seed_value INTEGER) RETURNS VOID
 AS $$
-  plpy.execute('SELECT cdb_crankshaft._cdb_crankshaft_activate_py()')
   from crankshaft import random_seeds
   random_seeds.set_random_seeds(seed_value)
 $$ LANGUAGE plpythonu;

pg/sql/0.0.1/02_moran.sql

@@ -11,7 +11,6 @@ CREATE OR REPLACE FUNCTION
       w_type TEXT DEFAULT 'knn')
 RETURNS TABLE (moran FLOAT, quads TEXT, significance FLOAT, ids INT)
 AS $$
-  plpy.execute('SELECT cdb_crankshaft._cdb_crankshaft_activate_py()')
   from crankshaft.clustering import moran_local
   # TODO: use named parameters or a dictionary
   return moran_local(t, attr, significance, num_ngbrs, permutations, geom_column, id_col, w_type)
@@ -30,7 +29,6 @@ CREATE OR REPLACE FUNCTION
 		 w_type TEXT DEFAULT 'knn')
 RETURNS TABLE(moran FLOAT, quads TEXT, significance FLOAT, ids INT, y numeric)
 AS $$
-  plpy.execute('SELECT cdb_crankshaft._cdb_crankshaft_activate_py()')
   from crankshaft.clustering import moran_local_rate
   # TODO: use named parameters or a dictionary
   return moran_local_rate(t, numerator, denominator, significance, num_ngbrs, permutations, geom_column, id_col, w_type)

pg/sql/0.0.1/04_dot_density.sql

@@ -51,4 +51,4 @@ BEGIN
   RETURN ST_Collect(points);
 END;
 $$
-LANGUAGE plpgsql VOLATILE;
+LANGUAGE plpgsql VOLATILE

pg/sql/0.0.1/population.sql

@@ -0,0 +1,138 @@
+-- Function to obtain an estimate of the population living inside
+-- an area (polygon) from the CartoDB Data Observatory
+CREATE OR REPLACE FUNCTION cdb_population(area geometry)
+RETURNS NUMERIC AS $$
+DECLARE
+  georef_column TEXT;
+  table_id TEXT;
+  tag_value TEXT;
+  table_name TEXT;
+  column_name TEXT;
+  population NUMERIC;
+BEGIN
+
+  -- Note: comments contain pseudo-code that should be implemented
+
+  -- Register metadata tables:
+  -- This would require super-user privileges
+  /*
+  SELECT cdb_add_remote_table('observatory', 'bmd_column_table');
+  SELECT cdb_add_remote_table('observatory', 'bmd_column_2_column');
+  SELECT cdb_add_remote_table('observatory', 'bmd_table');
+  SELECT cdb_add_remote_table('observatory', 'bmd_column_table');
+  SELECT cdb_add_remote_table('observatory', 'bmd_column_tag');
+  SELECT cdb_add_remote_table('observatory', 'bmd_tag');
+  */
+
+  tag_value := 'population';
+
+
+  -- Determine the georef column id to be used: it must have type 'geometry',
+  -- the maximum weight.
+  -- TODO: in general, multiple columns with maximal weight could be found;
+  -- we should use the timespan of the table to disambiguate (choose the
+  -- most recent). Also a rank of geometry columns should be introduced to
+  -- find select the greatest resolution available.
+  /*
+  WITH selected_tables AS (
+    -- Find tables that have population columns and cover the input area
+    SELECT tab.id AS id
+    FROM observatory.bmd_column col,
+         observatory.bmd_column_table coltab,
+         observatory.bmd_table tab,
+         observatory.bmd_tag tag,
+         observatory.bmd_column_tag coltag
+    WHERE coltab.column_id = col.id
+      AND coltab.table_id = tab.id
+      AND coltag.tag_id = tag.id
+      AND coltag.column_id = col.id
+      AND tag.name ILIKE tag_value
+      AND tab.id = table_id
+      AND tab.bounds && area;
+  )
+  SELECT
+    FROM bmd_column col
+    JOIN bmd_table tab ON col.table_id = tab.id
+    WHERE type = 'geometry'
+      AND tab.id IN (selected_tables)
+    ORDER BY weight DESC LIMIT 1;
+  */
+  georef_column := '"us.census.tiger".block_group_2013';
+
+  -- Now we will query the metadata to find which actual tables correspond
+  -- to this datasource and resolution/timespan
+  -- and choose the 'parent' or more general of them.
+  /*
+  SELECT from_table_geoid.id data_table_id
+  FROM observatory.bmd_column_table from_column_table_geoid,
+       observatory.bmd_column_table to_column_table_geoid,
+       observatory.bmd_column_2_column rel,
+       observatory.bmd_column_table to_column_table_geom,
+       observatory.bmd_table from_table_geoid,
+       observatory.bmd_table to_table_geoid,
+       observatory.bmd_table to_table_geom
+  WHERE from_column_table_geoid.column_id = to_column_table_geoid.column_id
+    AND to_column_table_geoid.column_id = rel.from_id
+    AND rel.reltype = 'geom_ref'
+    AND rel.to_id = to_column_table_geom.column_id
+    AND to_column_table_geom.column_id = georef_column
+    AND from_table_geoid.id = from_column_table_geoid.table_id
+    AND to_table_geoid.id = to_column_table_geoid.table_id
+    AND to_table_geom.id = to_column_table_geom.table_id
+    AND from_table_geoid.bounds && area
+  ORDER by from_table_geoid.timespan desc
+  INTO table_id;
+  */
+  table_id := '"us.census.acs".extract_2013_5yr_block_group';
+
+  -- Next will fetch the columns of that table that are tagged as population:
+  -- and get the more general one (not having a parent or denominator)
+  /*
+  WITH column_ids AS (
+    SELECT col.id AS id
+    FROM observatory.bmd_column col,
+         observatory.bmd_column_table coltab,
+         observatory.bmd_table tab,
+         observatory.bmd_tag tag,
+         observatory.bmd_column_tag coltag
+    WHERE coltab.column_id = col.id
+      AND coltab.table_id = tab.id
+      AND coltag.tag_id = tag.id
+      AND coltag.column_id = col.id
+      AND tag.name ILIKE tag_value
+      AND tab.id = table_id;
+  ),
+  excluded_column_ids AS (
+    SELECT from_id AS id
+    FROM observatory.bmd_column_2_column
+    WHERE from_id in (column_ids)
+      AND reltype in ('parent', 'denominator')
+      AND to_id in (column_ids)
+  ),
+  SELECT bmd_table.tablename, bmd_column_table.colname
+  FROM observatory.bmd_column_table,
+       observatory.bmd_table
+  WHERE bmd_column_table.table_id = bmd_table.id
+    AND bmd_column_table.column_id IN (column_ids)
+    AND NOT bmd_column_table.column_id IN (exclude_column_ids)
+  INTO (table_name, column_name);
+  */
+  table_name := 'us_census_acs2013_5yr_block_group';
+  column_name := 'total_pop';
+
+  -- Register the foreign table
+  -- This would require super-user privileges
+  -- SELECT cdb_add_remote_table('observatory', table_name);
+
+  -- Perform the query
+  SELECT cdb_crankshaft.cdb_overlap_sum(
+    area,
+    table_name,
+    column_name,
+    schema_name := 'observatory')
+  INTO population;
+
+  RETURN population;
+END;
+$$
+LANGUAGE plpgsql VOLATILE

pg/test/0.0.1/expected/01_install_test.out

@@ -3,4 +3,4 @@ CREATE EXTENSION plpythonu;
 CREATE EXTENSION postgis;
 CREATE EXTENSION cartodb;
 -- Install the extension
-CREATE EXTENSION crankshaft VERSION 'dev';
+CREATE EXTENSION crankshaft;

pg/test/0.0.1/results/01_install_test.out

@@ -0,0 +1,6 @@
+-- Install dependencies
+CREATE EXTENSION plpythonu;
+CREATE EXTENSION postgis;
+CREATE EXTENSION cartodb;
+-- Install the extension
+CREATE EXTENSION crankshaft;

pg/test/0.0.1/sql/01_install_test.sql

@@ -4,4 +4,4 @@ CREATE EXTENSION postgis;
 CREATE EXTENSION cartodb;
 
 -- Install the extension
-CREATE EXTENSION crankshaft VERSION 'dev';
+CREATE EXTENSION crankshaft;

python/.gitignore

@@ -1,2 +1 @@
 *.pyc
-dev/

python/Makefile

@@ -0,0 +1,11 @@
+# Install the package (needs root privileges)
+install:
+	pip install ./crankshaft --upgrade
+
+# Test from source code
+test:
+	(cd crankshaft && nosetests test/)
+
+# Test currently installed package
+testinstalled:
+	nosetests crankshaft/test/

python/crankshaft/setup.py

@@ -10,7 +10,7 @@ from setuptools import setup, find_packages
 setup(
     name='crankshaft',
 
-    version='0.0.1',
+    version='0.0.01',
 
     description='CartoDB Spatial Analysis Python Library',
 

src/pg/.gitignore

@@ -1,6 +0,0 @@
-regression.diffs
-regression.out
-results/
-crankshaft--dev.sql
-crankshaft--dev--current.sql
-crankshaft--current--dev.sql

src/pg/Makefile

@@ -1,41 +0,0 @@
-# Generation of a new development version 'dev' (with an alias 'current' for
-# updating easily by upgrading to 'current', then 'dev')
-
-# sudo make install -- generate the 'dev' version from current source
-#                      and make it available to PostgreSQL
-# PGUSER=postgres make installcheck -- test the 'dev' extension
-
-EXTENSION    = crankshaft
-
-DATA         = $(EXTENSION)--dev.sql \
-	             $(EXTENSION)--current--dev.sql \
-	             $(EXTENSION)--dev--current.sql
-
-SOURCES_DATA_DIR = sql
-SOURCES_DATA = $(wildcard $(SOURCES_DATA_DIR)/*.sql)
-
-$(DATA): $(SOURCES_DATA)
-	cat $(SOURCES_DATA_DIR)/*.sql > $@
-
-TEST_DIR = test
-REGRESS = $(notdir $(basename $(wildcard $(TEST_DIR)/sql/*test.sql)))
-REGRESS_OPTS = --inputdir='$(TEST_DIR)' --outputdir='$(TEST_DIR)'
-
-PG_CONFIG = pg_config
-PGXS := $(shell $(PG_CONFIG) --pgxs)
-include $(PGXS)
-
-# This seems to be needed at least for PG 9.3.11
-all: $(DATA)
-
-# WIP: goals for releasing the extension...
-
-EXTVERSION   = $(shell grep default_version $(EXTENSION).control | sed -e "s/default_version[[:space:]]*=[[:space:]]*'\([^']*\)'/\1/")
-
-../release/$(EXTENSION).control: $(EXTENSION).control
-	cp $< $@
-
-release: ../release/$(EXTENSION).control
-	cp $(EXTENSION)--dev.sql $(EXTENSION)--$(EXTVERSION).sql
-	# pending: create upgrade/downgrade scripts,
-	#          commit, push, tag....

src/pg/sql/00_header.sql

@@ -1,3 +0,0 @@
---DO NOT MODIFY THIS FILE, IT IS GENERATED AUTOMATICALLY FROM SOURCES
--- Complain if script is sourced in psql, rather than via CREATE EXTENSION
-\echo Use "CREATE EXTENSION crankshaft" to load this file. \quit

src/pg/sql/01_py.sql

@@ -1,18 +0,0 @@
--- Use the crankshaft python module
-CREATE OR REPLACE FUNCTION _cdb_crankshaft_activate_py()
-RETURNS VOID
-AS $$
-    # activate virtualenv
-    # TODO: parameterize with environment variables or something
-    venv_path = '/home/ubuntu/crankshaft/src/py/dev'
-    activate_path = venv_path + '/bin/activate_this.py'
-    exec(open(activate_path).read(),
-         dict(__file__=activate_path))
-
-    # import something from virtualenv
-    # from crankshaft import random_seeds
-
-    # do some stuff
-    # random_seeds.set_random_seeds(123)
-    # plpy.notice('here we are')
-$$ LANGUAGE plpythonu;

src/pg/sql/90_permissions.sql

@@ -1,9 +0,0 @@
--- Make sure by default there are no permissions for publicuser
--- NOTE: this happens at extension creation time, as part of an implicit transaction.
--- REVOKE ALL PRIVILEGES ON SCHEMA cdb_crankshaft FROM PUBLIC, publicuser CASCADE;
-
--- Grant permissions on the schema to publicuser (but just the schema)
-GRANT USAGE ON SCHEMA cdb_crankshaft TO publicuser;
-
--- Revoke execute permissions on all functions in the schema by default
--- REVOKE EXECUTE ON ALL FUNCTIONS IN SCHEMA cdb_crankshaft FROM PUBLIC, publicuser;

src/pg/test/sql/90_permissions.sql

@@ -1,18 +0,0 @@
-SELECT cdb_crankshaft._cdb_random_seeds(1234);
-
--- Use regular user role
-SET ROLE test_regular_user;
-
--- Add to the search path the schema
-SET search_path TO public,cartodb,cdb_crankshaft;
-
--- Exercise public functions
-SELECT ppoints.code, m.quads
-  FROM ppoints
-  JOIN cdb_moran_local('ppoints', 'value') m
-    ON ppoints.cartodb_id = m.ids
-  ORDER BY ppoints.code;
-SELECT round(cdb_overlap_sum(
-  '0106000020E61000000100000001030000000100000004000000FFFFFFFFFF3604C09A0B9ECEC42E444000000000C060FBBF30C7FD70E01D44400000000040AD02C06481F1C8CD034440FFFFFFFFFF3604C09A0B9ECEC42E4440'::geometry,
-  'values', 'value'
-), 2);

src/py/Makefile

@@ -1,9 +0,0 @@
-# Install the package locally for development
-install:
-	virtualenv dev
-	./dev/bin/pip install ./crankshaft --upgrade
-	./dev/bin/pip install nose
-
-# Test develpment install
-testinstalled:
-	./dev/bin/nosetests crankshaft/test/