add full tests

src/pg/.gitignore

@@ -4,3 +4,4 @@ results/
 crankshaft--dev.sql
 crankshaft--dev--current.sql
 crankshaft--current--dev.sql
+crankshaft--*--dev.sql

src/pg/Makefile

@@ -7,22 +7,19 @@ include ../../Makefile.global
 #   requires sudo. In additionof the current development version
 #   named 'dev', an alias 'current' is generating for ease of
 #   update (upgrade to 'current', then to 'dev').
-#   the python module is installed in a virtualenv in envs/dev/
 # * test runs the tests for the currently generated Development
 #   extension.
 
-DATA         = $(EXTENSION)--dev.sql \
-	             $(EXTENSION)--current--dev.sql \
-	             $(EXTENSION)--dev--current.sql
+DATA = \
+    $(EXTENSION)--dev.sql \
+    $(EXTENSION)--current--dev.sql \
+    $(EXTENSION)--dev--current.sql \
+    $(EXTENSION)--$(RELEASE_VERSION)--dev.sql
 
 SOURCES_DATA_DIR = sql
 SOURCES_DATA = $(wildcard $(SOURCES_DATA_DIR)/*.sql)
 
-VIRTUALENV_PATH = $(realpath ../../envs)
-ESC_VIRVIRTUALENV_PATH = $(subst /,\/,$(VIRTUALENV_PATH))
-
-REPLACEMENTS = -e 's/@@VERSION@@/$(EXTVERSION)/g' \
-               -e 's/@@VIRTUALENV_PATH@@/$(ESC_VIRVIRTUALENV_PATH)/g'
+REPLACEMENTS = -e 's/@@VERSION@@/$(EXTVERSION)/g'
 
 $(DATA): $(SOURCES_DATA)
 	$(SED) $(REPLACEMENTS) $(SOURCES_DATA_DIR)/*.sql > $@
@@ -54,7 +51,6 @@ release: ../../release/$(EXTENSION).control $(SOURCES_DATA)
 	$(SED) $(REPLACEMENTS) $(SOURCES_DATA_DIR)/*.sql > ../../release/$(EXTENSION)--$(EXTVERSION).sql
 
 # Install the current relese into the PostgreSQL extensions directory
-# and the Python package in a virtual environment envs/X.Y.Z
 deploy:
 	$(INSTALL_DATA) ../../release/$(EXTENSION).control '$(DESTDIR)$(datadir)/extension/'
 	$(INSTALL_DATA) ../../release/*.sql '$(DESTDIR)$(datadir)/extension/'

src/pg/crankshaft.control

@@ -1,5 +1,5 @@
 comment = 'CartoDB Spatial Analysis extension'
-default_version = '0.0.2'
-requires = 'plpythonu, postgis, cartodb'
+default_version = '0.2.0'
+requires = 'plpythonu, postgis'
 superuser = true
 schema = cdb_crankshaft

src/pg/sql/02_py.sql

@@ -1,23 +0,0 @@
-CREATE OR REPLACE FUNCTION _cdb_crankshaft_virtualenvs_path()
-RETURNS text
-AS $$
-  BEGIN
-    -- RETURN '/opt/virtualenvs/crankshaft';
-    RETURN '@@VIRTUALENV_PATH@@';
-  END;
-$$ language plpgsql IMMUTABLE STRICT;
-
--- Use the crankshaft python module
-CREATE OR REPLACE FUNCTION _cdb_crankshaft_activate_py()
-RETURNS VOID
-AS $$
-    import os
-    # plpy.notice('%',str(os.environ))
-    # activate virtualenv
-    crankshaft_version = plpy.execute('SELECT cdb_crankshaft._cdb_crankshaft_internal_version()')[0]['_cdb_crankshaft_internal_version']
-    base_path = plpy.execute('SELECT cdb_crankshaft._cdb_crankshaft_virtualenvs_path()')[0]['_cdb_crankshaft_virtualenvs_path']
-    default_venv_path = os.path.join(base_path, crankshaft_version)
-    venv_path =  os.environ.get('CRANKSHAFT_VENV', default_venv_path)
-    activate_path = venv_path + '/bin/activate_this.py'
-    exec(open(activate_path).read(), dict(__file__=activate_path))
-$$ LANGUAGE plpythonu;

src/pg/sql/03_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;

src/pg/sql/04_py_agg.sql

@@ -0,0 +1,31 @@
+CREATE OR REPLACE FUNCTION
+    CDB_PyAggS(current_state Numeric[], current_row Numeric[]) 
+    returns NUMERIC[] as $$
+    BEGIN
+        if array_upper(current_state,1) is null  then
+            RAISE NOTICE 'setting state %',array_upper(current_row,1);
+            current_state[1] = array_upper(current_row,1);
+        end if;
+        return array_cat(current_state,current_row) ;
+    END
+    $$ LANGUAGE plpgsql;
+
+-- Create aggregate if it did not exist
+DO $$
+BEGIN
+    IF NOT EXISTS (
+        SELECT *
+        FROM pg_catalog.pg_proc p
+            LEFT JOIN pg_catalog.pg_namespace n ON n.oid = p.pronamespace
+        WHERE n.nspname = 'cdb_crankshaft'
+            AND p.proname = 'cdb_pyagg'
+            AND p.proisagg)
+    THEN
+        CREATE AGGREGATE CDB_PyAgg(NUMERIC[]) (
+            SFUNC = CDB_PyAggS,
+            STYPE = Numeric[],
+            INITCOND = "{}"
+        );
+    END IF;
+END
+$$ LANGUAGE plpgsql;

src/pg/sql/05_segmentation.sql

@@ -0,0 +1,53 @@
+
+CREATE OR REPLACE FUNCTION
+  CDB_CreateAndPredictSegment(
+    target NUMERIC[],
+    features NUMERIC[],
+    target_features NUMERIC[],
+    target_ids NUMERIC[],
+    n_estimators INTEGER DEFAULT 1200,
+    max_depth INTEGER DEFAULT 3,
+    subsample DOUBLE PRECISION DEFAULT 0.5,
+    learning_rate DOUBLE PRECISION DEFAULT 0.01,
+    min_samples_leaf INTEGER DEFAULT 1)
+RETURNS TABLE(cartodb_id NUMERIC, prediction NUMERIC, accuracy NUMERIC)
+AS $$
+    import numpy as np
+    import plpy
+
+    from crankshaft.segmentation import create_and_predict_segment_agg
+    model_params = {'n_estimators': n_estimators,
+                    'max_depth': max_depth,
+                    'subsample': subsample,
+                    'learning_rate': learning_rate,
+                    'min_samples_leaf': min_samples_leaf}
+
+    def unpack2D(data):
+        dimension = data.pop(0)
+        a = np.array(data, dtype=float)
+        return a.reshape(len(a)/dimension, dimension)
+
+    return create_and_predict_segment_agg(np.array(target, dtype=float),
+            unpack2D(features),
+            unpack2D(target_features),
+            target_ids,
+            model_params)
+
+$$ LANGUAGE plpythonu;
+
+CREATE OR REPLACE FUNCTION
+  CDB_CreateAndPredictSegment (
+      query TEXT,
+      variable_name TEXT,
+      target_table TEXT,
+      n_estimators INTEGER DEFAULT 1200,
+      max_depth INTEGER DEFAULT 3,
+      subsample DOUBLE PRECISION DEFAULT 0.5,
+      learning_rate DOUBLE PRECISION DEFAULT 0.01,
+      min_samples_leaf INTEGER DEFAULT 1)
+RETURNS TABLE (cartodb_id TEXT, prediction NUMERIC, accuracy NUMERIC)
+AS $$
+  from crankshaft.segmentation import create_and_predict_segment
+  model_params = {'n_estimators': n_estimators, 'max_depth':max_depth, 'subsample' : subsample, 'learning_rate': learning_rate, 'min_samples_leaf' : min_samples_leaf}
+  return create_and_predict_segment(query,variable_name,target_table, model_params)
+$$ LANGUAGE plpythonu;

src/pg/sql/07_gravity.sql

@@ -0,0 +1,115 @@
+CREATE OR REPLACE FUNCTION CDB_Gravity(
+    IN target_query text,
+    IN weight_column text,
+    IN source_query text,
+    IN pop_column text,
+    IN target bigint,
+    IN radius integer,
+    IN minval numeric DEFAULT -10e307
+    )
+RETURNS TABLE(
+    the_geom geometry,
+    source_id bigint,
+    target_id bigint,
+    dist numeric,
+    h numeric,
+    hpop numeric)  AS $$
+DECLARE
+    t_id bigint[];
+    t_geom geometry[];
+    t_weight numeric[];
+    s_id bigint[];
+    s_geom geometry[];
+    s_pop numeric[];
+BEGIN
+    EXECUTE 'WITH foo as('+target_query+') SELECT array_agg(cartodb_id), array_agg(the_geom), array_agg(' || weight_column || ') FROM foo' INTO t_id, t_geom, t_weight;
+    EXECUTE 'WITH foo as('+source_query+') SELECT array_agg(cartodb_id), array_agg(the_geom), array_agg(' || pop_column || ') FROM foo' INTO s_id, s_geom, s_pop;
+    RETURN QUERY
+    SELECT g.* FROM t, s, CDB_Gravity(t_id, t_geom, t_weight, s_id, s_geom, s_pop, target, radius, minval) g;
+END;
+$$ language plpgsql;
+
+CREATE OR REPLACE FUNCTION CDB_Gravity(
+    IN t_id bigint[],
+    IN t_geom geometry[],
+    IN t_weight numeric[],
+    IN s_id bigint[],
+    IN s_geom geometry[],
+    IN s_pop numeric[],
+    IN target bigint,
+    IN radius integer,
+    IN minval numeric DEFAULT -10e307
+    )
+RETURNS TABLE(
+    the_geom geometry,
+    source_id bigint,
+    target_id bigint,
+    dist numeric,
+    h numeric,
+    hpop numeric)  AS $$
+DECLARE
+    t_type text;
+    s_type text;
+    t_center geometry[];
+    s_center geometry[];
+BEGIN
+    t_type := GeometryType(t_geom[1]);
+    s_type := GeometryType(s_geom[1]);
+    IF t_type = 'POINT' THEN
+        t_center := t_geom;
+    ELSE
+        WITH tmp as (SELECT unnest(t_geom) as g) SELECT array_agg(ST_Centroid(g)) INTO t_center FROM tmp;
+    END IF;
+    IF s_type = 'POINT' THEN
+        s_center := s_geom;
+    ELSE
+        WITH tmp as (SELECT unnest(s_geom) as g) SELECT array_agg(ST_Centroid(g)) INTO s_center FROM tmp;
+    END IF;
+    RETURN QUERY
+        with target0 as(
+            SELECT unnest(t_center) as tc, unnest(t_weight) as tw, unnest(t_id) as td
+        ),
+        source0 as(
+            SELECT unnest(s_center) as sc, unnest(s_id) as sd, unnest (s_geom) as sg, unnest(s_pop) as sp
+        ),
+        prev0 as(
+            SELECT
+                source0.sg,
+                source0.sd as sourc_id,
+                coalesce(source0.sp,0) as sp,
+                target.td as targ_id,
+                coalesce(target.tw,0) as tw,
+                GREATEST(1.0,ST_Distance(geography(target.tc), geography(source0.sc)))::numeric as distance
+            FROM source0
+            CROSS JOIN LATERAL
+                (
+                SELECT
+                    *
+                FROM target0
+                    WHERE tw > minval
+                    AND ST_DWithin(geography(source0.sc), geography(tc), radius)
+                ) AS target
+        ),
+        deno as(
+            SELECT
+                sourc_id,
+                sum(tw/distance) as h_deno
+            FROM
+                prev0
+            GROUP BY sourc_id
+        )
+        SELECT
+            p.sg as the_geom,
+            p.sourc_id as source_id,
+            p.targ_id as target_id,
+            case when p.distance > 1 then p.distance else 0.0 end as dist,
+            100*(p.tw/p.distance)/d.h_deno as h,
+            p.sp*(p.tw/p.distance)/d.h_deno as hpop
+        FROM
+            prev0 p,
+            deno d
+        WHERE
+            p.targ_id = target AND
+            p.sourc_id = d.sourc_id;
+END;
+$$ language plpgsql;

src/pg/sql/08_interpolation.sql

@@ -14,9 +14,12 @@ $$
 DECLARE
     gs geometry[];
     vs numeric[];
+    output numeric;
 BEGIN
     EXECUTE 'WITH a AS('||query||') SELECT array_agg(the_geom), array_agg(attrib) FROM a' INTO gs, vs;
-    RETURN QUERY SELECT CDB_SpatialInterpolation(gs, vs, point, method, p1,p2) FROM a;
+    SELECT CDB_SpatialInterpolation(gs, vs, point, method, p1,p2) INTO output FROM a;
+
+    RETURN output;
 END;
 $$
 language plpgsql IMMUTABLE;
@@ -71,11 +74,11 @@ BEGIN
                 SELECT array_agg(v) INTO vertex FROM a;
 
             -- retrieve the value of each vertex
-        WITH a AS(SELECT unnest(vertex) as geo, unnest(colin) as c)
+        WITH a AS(SELECT unnest(geomin) as geo, unnest(colin) as c)
             SELECT c INTO va FROM a WHERE ST_Equals(geo, vertex[1]);
-        WITH a AS(SELECT unnest(vertex) as geo, unnest(colin) as c)
+        WITH a AS(SELECT unnest(geomin) as geo, unnest(colin) as c)
             SELECT c INTO vb FROM a WHERE ST_Equals(geo, vertex[2]);
-        WITH a AS(SELECT unnest(vertex) as geo, unnest(colin) as c)
+        WITH a AS(SELECT unnest(geomin) as geo, unnest(colin) as c)
                 SELECT c INTO vc FROM a WHERE ST_Equals(geo, vertex[3]);
 
         SELECT ST_area(g), ST_area(ST_MakePolygon(ST_MakeLine(ARRAY[point, vertex[2], vertex[3], point]))), ST_area(ST_MakePolygon(ST_MakeLine(ARRAY[point, vertex[1], vertex[3], point]))), ST_area(ST_MakePolygon(ST_MakeLine(ARRAY[point,vertex[1],vertex[2], point]))) INTO sg, sa, sb, sc;

src/pg/sql/10_moran.sql

@@ -10,7 +10,6 @@ CREATE OR REPLACE FUNCTION
       id_col TEXT DEFAULT 'cartodb_id')
 RETURNS TABLE (moran NUMERIC, significance NUMERIC)
 AS $$
-  plpy.execute('SELECT cdb_crankshaft._cdb_crankshaft_activate_py()')
   from crankshaft.clustering import moran_local
   # TODO: use named parameters or a dictionary
   return moran(subquery, column_name, w_type, num_ngbrs, permutations, geom_col, id_col)
@@ -28,7 +27,6 @@ CREATE OR REPLACE FUNCTION
       id_col TEXT)
 RETURNS TABLE (moran NUMERIC, quads TEXT, significance NUMERIC, rowid INT, vals NUMERIC)
 AS $$
-  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(subquery, column_name, w_type, num_ngbrs, permutations, geom_col, id_col)
@@ -122,7 +120,6 @@ CREATE OR REPLACE FUNCTION
       id_col TEXT DEFAULT 'cartodb_id')
 RETURNS TABLE (moran FLOAT, significance FLOAT)
 AS $$
-  plpy.execute('SELECT cdb_crankshaft._cdb_crankshaft_activate_py()')
   from crankshaft.clustering import moran_local
   # TODO: use named parameters or a dictionary
   return moran_rate(subquery, numerator, denominator, w_type, num_ngbrs, permutations, geom_col, id_col)
@@ -143,7 +140,6 @@ CREATE OR REPLACE FUNCTION
 RETURNS
 TABLE(moran NUMERIC, quads TEXT, significance NUMERIC, rowid INT, vals NUMERIC)
 AS $$
-  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(subquery, numerator, denominator, w_type, num_ngbrs, permutations, geom_col, id_col)

src/pg/sql/11_kmeans.sql

@@ -0,0 +1,63 @@
+CREATE OR REPLACE FUNCTION  CDB_KMeans(query text, no_clusters integer,no_init integer default 20)
+RETURNS table (cartodb_id integer, cluster_no integer) as $$
+    
+    from crankshaft.clustering import kmeans
+    return kmeans(query,no_clusters,no_init)
+
+$$ language plpythonu;
+
+
+CREATE OR REPLACE FUNCTION CDB_WeightedMeanS(state Numeric[],the_geom GEOMETRY(Point, 4326), weight NUMERIC)
+RETURNS Numeric[] AS 
+$$
+DECLARE 
+    newX NUMERIC;
+    newY NUMERIC;
+    newW NUMERIC;
+BEGIN
+    IF weight IS NULL OR the_geom IS NULL THEN 
+        newX = state[1];
+        newY = state[2];
+        newW = state[3];
+    ELSE
+        newX = state[1] + ST_X(the_geom)*weight;
+        newY = state[2] + ST_Y(the_geom)*weight;
+        newW = state[3] + weight;
+    END IF;
+    RETURN Array[newX,newY,newW];
+
+END
+$$ LANGUAGE plpgsql;
+
+CREATE OR REPLACE FUNCTION CDB_WeightedMeanF(state Numeric[])
+RETURNS GEOMETRY AS 
+$$
+BEGIN
+    IF state[3] = 0 THEN 
+        RETURN ST_SetSRID(ST_MakePoint(state[1],state[2]), 4326);
+    ELSE 
+        RETURN ST_SETSRID(ST_MakePoint(state[1]/state[3], state[2]/state[3]),4326);
+    END IF;
+END
+$$ LANGUAGE plpgsql;
+
+-- Create aggregate if it did not exist
+DO $$
+BEGIN
+    IF NOT EXISTS (
+        SELECT *
+        FROM pg_catalog.pg_proc p
+            LEFT JOIN pg_catalog.pg_namespace n ON n.oid = p.pronamespace
+        WHERE n.nspname = 'cdb_crankshaft'
+            AND p.proname = 'cdb_weightedmean'
+            AND p.proisagg)
+    THEN
+        CREATE AGGREGATE CDB_WeightedMean(geometry(Point, 4326), NUMERIC) (
+            SFUNC = CDB_WeightedMeanS,
+            FINALFUNC = CDB_WeightedMeanF,
+            STYPE = Numeric[],
+            INITCOND = "{0.0,0.0,0.0}" 
+        );
+    END IF;
+END
+$$ LANGUAGE plpgsql;

src/pg/sql/11_markov.sql

@@ -0,0 +1,81 @@
+-- Spatial Markov
+
+-- input table format:
+-- id | geom | date_1 | date_2 | date_3
+--  1 | Pt1  | 12.3   | 13.1   | 14.2
+--  2 | Pt2  | 11.0   | 13.2   | 12.5
+--  ...
+-- Sample Function call:
+-- SELECT CDB_SpatialMarkov('SELECT * FROM real_estate',
+--                          Array['date_1', 'date_2', 'date_3'])
+
+CREATE OR REPLACE FUNCTION
+  CDB_SpatialMarkovTrend (
+      subquery TEXT,
+      time_cols TEXT[],
+      num_classes INT DEFAULT 7,
+      w_type TEXT DEFAULT 'knn',
+      num_ngbrs INT DEFAULT 5,
+  	  permutations INT DEFAULT 99,
+  	  geom_col TEXT DEFAULT 'the_geom',
+  	  id_col TEXT DEFAULT 'cartodb_id')
+RETURNS TABLE (trend NUMERIC, trend_up NUMERIC, trend_down NUMERIC, volatility NUMERIC, rowid INT)
+AS $$
+
+  from crankshaft.space_time_dynamics import spatial_markov_trend
+
+  ## TODO: use named parameters or a dictionary
+  return spatial_markov_trend(subquery, time_cols, num_classes, w_type, num_ngbrs, permutations, geom_col, id_col)
+$$ LANGUAGE plpythonu;
+
+-- input table format: identical to above but in a predictable format
+-- Sample function call:
+-- SELECT cdb_spatial_markov('SELECT * FROM real_estate',
+--                           'date_1')
+
+
+-- CREATE OR REPLACE FUNCTION
+--   cdb_spatial_markov (
+--       subquery TEXT,
+--       time_col_min text,
+--       time_col_max text,
+--       date_format text, -- '_YYYY_MM_DD'
+--       num_time_per_bin INT DEFAULT 1,
+--   	  permutations INT DEFAULT 99,
+--   	  geom_column TEXT DEFAULT 'the_geom',
+--   	  id_col TEXT DEFAULT 'cartodb_id',
+--       w_type TEXT DEFAULT 'knn',
+--       num_ngbrs int DEFAULT 5)
+-- RETURNS TABLE (moran FLOAT, quads TEXT, significance FLOAT, ids INT)
+-- AS $$
+--   plpy.execute('SELECT cdb_crankshaft._cdb_crankshaft_activate_py()')
+--   from crankshaft.clustering import moran_local
+--   # TODO: use named parameters or a dictionary
+--   return spatial_markov(subquery, time_cols, permutations, geom_column, id_col, w_type, num_ngbrs)
+-- $$ LANGUAGE plpythonu;
+--
+-- -- input table format:
+-- -- id | geom | date  | measurement
+-- --  1 | Pt1  | 12/3  | 13.2
+-- --  2 | Pt2  | 11/5  | 11.3
+-- --  3 | Pt1  | 11/13 | 12.9
+-- --  4 | Pt3  | 12/19 | 10.1
+-- -- ...
+--
+-- CREATE OR REPLACE FUNCTION
+--   cdb_spatial_markov (
+--       subquery TEXT,
+--       time_col text,
+--       num_time_per_bin INT DEFAULT 1,
+--   	  permutations INT DEFAULT 99,
+--   	  geom_column TEXT DEFAULT 'the_geom',
+--   	  id_col TEXT DEFAULT 'cartodb_id',
+--       w_type TEXT DEFAULT 'knn',
+--       num_ngbrs int DEFAULT 5)
+-- RETURNS TABLE (moran FLOAT, quads TEXT, significance FLOAT, ids INT)
+-- AS $$
+--   plpy.execute('SELECT cdb_crankshaft._cdb_crankshaft_activate_py()')
+--   from crankshaft.clustering import moran_local
+--   # TODO: use named parameters or a dictionary
+--   return spatial_markov(subquery, time_cols, permutations, geom_column, id_col, w_type, num_ngbrs)
+-- $$ LANGUAGE plpythonu;

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

@@ -1,6 +1,18 @@
 -- Install dependencies
 CREATE EXTENSION plpythonu;
-CREATE EXTENSION postgis;
-CREATE EXTENSION cartodb;
+CREATE EXTENSION postgis VERSION '2.2.2';
+-- Create role publicuser if it does not exist
+DO
+$$
+BEGIN
+   IF NOT EXISTS (
+      SELECT *
+      FROM   pg_catalog.pg_user
+      WHERE  usename = 'publicuser') THEN
+
+      CREATE ROLE publicuser LOGIN;
+   END IF;
+END
+$$ LANGUAGE plpgsql;
 -- Install the extension
 CREATE EXTENSION crankshaft VERSION 'dev';

src/pg/test/expected/05_kmeans_test.out

@@ -0,0 +1,10 @@
+\pset format unaligned
+\set ECHO all
+SELECT count(DISTINCT cluster_no) as clusters from cdb_crankshaft.cdb_kmeans('select * from ppoints', 2);
+clusters
+2
+(1 row)
+SELECT count(*) clusters from (select  cdb_crankshaft.CDB_WeightedMean(the_geom, value::NUMERIC), code from ppoints group by code) p;
+clusters
+52
+(1 row)

src/pg/test/expected/05_markov_test.out

@@ -0,0 +1,10 @@
+SET client_min_messages TO WARNING;
+\set ECHO none
+_cdb_random_seeds
+
+(1 row)
+cartodb_id|trend_test|trend_up_test|trend_down_test|volatility_test
+1|t|t|t|t
+2|t|t|t|t
+3|t|t|t|t
+(3 rows)

src/pg/test/expected/06_segmentation_test.out

@@ -0,0 +1,27 @@
+\pset format unaligned
+\set ECHO none
+_cdb_random_seeds
+
+(1 row)
+within_tolerance
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+t
+(20 rows)

src/pg/test/expected/07_gravity_test.out

@@ -0,0 +1,14 @@
+SET client_min_messages TO WARNING;
+\set ECHO none
+                  the_geom                  |            h            |           hpop           |      dist      
+--------------------------------------------+-------------------------+--------------------------+----------------
+ 01010000001361C3D32B650140DD24068195B34440 |  1.51078258369747945249 |  12.08626066957983561994 | 4964.714459152
+ 01010000002497FF907EFB0040713D0AD7A3B04440 | 98.29730954183620807430 | 688.08116679285345652007 |   99.955141922
+ 0101000000A167B3EA733501401D5A643BDFAF4440 | 63.70532894711274639196 | 382.23197368267647835174 | 2488.330566505
+ 010100000062A1D634EF380140BE9F1A2FDDB44440 | 35.35415870080995954879 | 176.77079350404979774397 | 4359.370460594
+ 010100000052B81E85EB510140355EBA490CB24440 | 33.12290506987740864904 | 132.49162027950963459615 | 3703.664449828
+ 0101000000C286A757CA320140736891ED7CAF4440 | 65.45251754279248087849 | 196.35755262837744263547 | 2512.092358644
+ 01010000007DD0B359F5390140C976BE9F1AAF4440 | 62.83927792471345639225 | 125.67855584942691278449 |  2926.25725244
+ 0101000000D237691A140D01407E6FD39FFDB44440 | 53.54905726651871279586 |  53.54905726651871279586 | 3744.515577777
+(8 rows)
+

src/pg/test/expected/08_interpolation_test.out

@@ -1,7 +1,12 @@
 SET client_min_messages TO WARNING;
 \set ECHO none
+<<<<<<< HEAD
  nn  |           nni            |       idw       
 -----+--------------------------+-----------------
  200 | 780.79470198683925288365 | 341.46260750526
+=======
+cdb_spatialinterpolation
+t
+>>>>>>> ebbf68af5700f33737ac18516e2bdc22ea330b24
 (1 row)
 

src/pg/test/sql/01_install_test.sql

@@ -1,7 +1,20 @@
 -- Install dependencies
 CREATE EXTENSION plpythonu;
-CREATE EXTENSION postgis;
-CREATE EXTENSION cartodb;
+CREATE EXTENSION postgis VERSION '2.2.2';
+
+-- Create role publicuser if it does not exist
+DO
+$$
+BEGIN
+   IF NOT EXISTS (
+      SELECT *
+      FROM   pg_catalog.pg_user
+      WHERE  usename = 'publicuser') THEN
+
+      CREATE ROLE publicuser LOGIN;
+   END IF;
+END
+$$ LANGUAGE plpgsql;
 
 -- Install the extension
 CREATE EXTENSION crankshaft VERSION 'dev';

src/pg/test/sql/05_kmeans_test.sql

@@ -0,0 +1,6 @@
+\pset format unaligned
+\set ECHO all
+
+SELECT count(DISTINCT cluster_no) as clusters from cdb_crankshaft.cdb_kmeans('select * from ppoints', 2);
+
+SELECT count(*) clusters from (select  cdb_crankshaft.CDB_WeightedMean(the_geom, value::NUMERIC), code from ppoints group by code) p;

src/pg/test/sql/05_markov_test.sql

@@ -0,0 +1,33 @@
+SET client_min_messages TO WARNING;
+\set ECHO none
+\pset format unaligned
+\i test/fixtures/markov_usjoin_example.sql
+
+-- Areas of Interest functions perform some nondeterministic computations
+-- (to estimate the significance); we will set the seeds for the RNGs
+-- that affect those results to have repeatable results
+SELECT cdb_crankshaft._cdb_random_seeds(1234);
+
+SELECT
+  m1.cartodb_id,
+  CASE WHEN m1.cartodb_id = 1 THEN abs(m2.trend - 0.069767441860465115) / 0.069767441860465115 < 0.05
+       WHEN m1.cartodb_id = 2 THEN abs(m2.trend - 0.15151515151515152) / 0.15151515151515152 < 0.05
+       WHEN m1.cartodb_id = 3 THEN abs(m2.trend - 0.069767441860465115) / 0.069767441860465115 < 0.05
+       ELSE NULL END As trend_test,
+  CASE WHEN m1.cartodb_id = 1 THEN abs(m2.trend_up - 0.065217391304347824) / 0.065217391304347824 < 0.05
+       WHEN m1.cartodb_id = 2 THEN abs(m2.trend_up - 0.13157894736842105) / 0.13157894736842105 < 0.05
+       WHEN m1.cartodb_id = 3 THEN abs(m2.trend_up - 0.065217391304347824) / 0.065217391304347824 < 0.05
+       ELSE NULL END As trend_up_test,
+  CASE WHEN m1.cartodb_id = 1 THEN m2.trend_down = 0.0
+       WHEN m1.cartodb_id = 2 THEN m2.trend_down = 0.0
+       WHEN m1.cartodb_id = 3 THEN m2.trend_down = 0.0
+       ELSE NULL END As trend_down_test,
+  CASE WHEN m1.cartodb_id = 1 THEN abs(m2.volatility - 0.367574633389) / 0.367574633389  < 0.1
+       WHEN m1.cartodb_id = 2 THEN abs(m2.volatility - 0.33807340742635211) / 0.33807340742635211  < 0.1
+       WHEN m1.cartodb_id = 3 THEN abs(m2.volatility - 0.3682585596149513) / 0.3682585596149513  < 0.1
+       ELSE NULL END As volatility_test
+  FROM markov_usjoin_example As m1
+  JOIN cdb_crankshaft.CDB_SpatialMarkovTrend('SELECT * FROM markov_usjoin_example ORDER BY cartodb_id DESC', Array['y1995', 'y1996', 'y1997', 'y1998', 'y1999', 'y2000', 'y2001', 'y2002', 'y2003', 'y2004', 'y2005', 'y2006', 'y2007', 'y2008', 'y2009']::text[], 5::int, 'knn'::text, 5::int, 0::int, 'the_geom'::text, 'cartodb_id'::text) As m2
+    ON m1.cartodb_id = m2.rowid
+ORDER BY m1.cartodb_id
+LIMIT 3;

src/pg/test/sql/06_segmentation_test.sql

@@ -0,0 +1,33 @@
+\pset format unaligned
+\set ECHO none
+\i test/fixtures/ml_values.sql
+SELECT cdb_crankshaft._cdb_random_seeds(1234);
+
+WITH expected AS (
+  SELECT generate_series(1000,1020) AS id, unnest(ARRAY[
+    4.5656517130822492,
+    1.7928053473230694,
+    1.0283378773916563,
+    2.6586517814904593,
+    2.9699056242935944,
+    3.9550646059951347,
+    4.1662572444459745,
+    3.8126334839264162,
+    1.8809821053623488,
+    1.6349065129019873,
+    3.0391288591472954,
+    3.3035970359672553,
+    1.5835471589451968,
+    3.7530378537263638,
+    1.0833589653009252,
+    3.8104965452882897,
+    2.665217959294802,
+    1.5850334252802472,
+    3.679401198805563,
+    3.5332033186588636
+    ]) AS expected LIMIT 20
+), prediction AS (
+  SELECT cartodb_id::integer id, prediction
+  FROM cdb_crankshaft.CDB_CreateAndPredictSegment('SELECT target, x1, x2, x3 FROM ml_values WHERE class = $$train$$','target','SELECT cartodb_id, target, x1, x2, x3 FROM ml_values WHERE class = $$test$$')
+  LIMIT 20
+) SELECT abs(e.expected - p.prediction) <= 1e-9 AS within_tolerance FROM expected e, prediction p WHERE e.id = p.id;

src/pg/test/sql/07_gravity_test.sql

@@ -0,0 +1,24 @@
+SET client_min_messages TO WARNING;
+\set ECHO none
+
+WITH t AS (
+    SELECT
+    ARRAY[1,2,3] AS id,
+    ARRAY[7.0,8.0,3.0] AS w,
+    ARRAY[ST_GeomFromText('POINT(2.1744 41.4036)'),ST_GeomFromText('POINT(2.1228 41.3809)'),ST_GeomFromText('POINT(2.1511 41.3742)')] AS g
+),
+s AS (
+    SELECT
+    ARRAY[10,20,30,40,50,60,70,80] AS id,
+    ARRAY[800, 700, 600, 500, 400, 300, 200, 100] AS p,
+    ARRAY[ST_GeomFromText('POINT(2.1744 41.403)'),ST_GeomFromText('POINT(2.1228 41.380)'),ST_GeomFromText('POINT(2.1511 41.374)'),ST_GeomFromText('POINT(2.1528 41.413)'),ST_GeomFromText('POINT(2.165 41.391)'),ST_GeomFromText('POINT(2.1498 41.371)'),ST_GeomFromText('POINT(2.1533 41.368)'),ST_GeomFromText('POINT(2.131386 41.41399)')] AS g
+)
+SELECT
+    g.the_geom,
+    g.h,
+    g.hpop,
+    g.dist
+FROM
+    t,
+    s,
+    cdb_crankshaft.CDB_Gravity(t.id, t.g, t.w, s.id, s.g, s.p, 2, 100000, 3) g;

src/pg/test/sql/08_interpolation_test.sql

@@ -1,12 +1,21 @@
 SET client_min_messages TO WARNING;
 \set ECHO none
+<<<<<<< HEAD
+=======
+\pset format unaligned
+
+>>>>>>> ebbf68af5700f33737ac18516e2bdc22ea330b24
 WITH a AS (
     SELECT
     ARRAY[800, 700, 600, 500, 400, 300, 200, 100] AS vals,
     ARRAY[ST_GeomFromText('POINT(2.1744 41.403)'),ST_GeomFromText('POINT(2.1228 41.380)'),ST_GeomFromText('POINT(2.1511 41.374)'),ST_GeomFromText('POINT(2.1528 41.413)'),ST_GeomFromText('POINT(2.165 41.391)'),ST_GeomFromText('POINT(2.1498 41.371)'),ST_GeomFromText('POINT(2.1533 41.368)'),ST_GeomFromText('POINT(2.131386 41.41399)')] AS g
 )
+<<<<<<< HEAD
 SELECT
     CDB_SpatialInterpolation(g, vals, ST_GeomFromText('POINT(2.154 41.37)'),0) as NN,
     CDB_SpatialInterpolation(g, vals, ST_GeomFromText('POINT(2.154 41.37)'),1) as NNI,
     CDB_SpatialInterpolation(g, vals, ST_GeomFromText('POINT(2.154 41.37)'),2) as IDW
 FROM a;
+=======
+SELECT (cdb_crankshaft.CDB_SpatialInterpolation(g, vals, ST_GeomFromText('POINT(2.154 41.37)'), 1) - 780.79470198683925288365) / 780.79470198683925288365 < 0.001 As cdb_spatialinterpolation FROM a;
+>>>>>>> ebbf68af5700f33737ac18516e2bdc22ea330b24

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

@@ -4,7 +4,7 @@ SELECT cdb_crankshaft._cdb_random_seeds(1234);
 SET ROLE test_regular_user;
 
 -- Add to the search path the schema
-SET search_path TO public,cartodb,cdb_crankshaft;
+SET search_path TO public,cdb_crankshaft;
 
 -- Exercise public functions
 SELECT ppoints.code, m.quads

src/py/Makefile

@@ -2,21 +2,16 @@ include ../../Makefile.global
 
 # Install the package locally for development
 install:
-	virtualenv --system-site-packages ../../envs/dev
-	# source ../../envs/dev/bin/activate
-	../../envs/dev/bin/pip install -I ./crankshaft
-	../../envs/dev/bin/pip install -I nose
+	pip install --upgrade ./crankshaft
 
 # Test develpment install
 test:
-	../../envs/dev/bin/nosetests crankshaft/test/
+	nosetests crankshaft/test/
 
 release: ../../release/$(EXTENSION).control $(SOURCES_DATA)
 	mkdir -p ../../release/python/$(EXTVERSION)
 	cp -r ./$(PACKAGE) ../../release/python/$(EXTVERSION)/
 	$(SED) -i -r 's/version='"'"'[0-9]+\.[0-9]+\.[0-9]+'"'"'/version='"'"'$(EXTVERSION)'"'"'/g'  ../../release/python/$(EXTVERSION)/$(PACKAGE)/setup.py
 
-deploy:
-	virtualenv --system-site-packages $(VIRTUALENV_PATH)/$(RELEASE_VERSION)
-	$(VIRTUALENV_PATH)/$(RELEASE_VERSION)/bin/pip install -I -U ../../release/python/$(RELEASE_VERSION)/$(PACKAGE)
-	$(VIRTUALENV_PATH)/$(RELEASE_VERSION)/bin/pip install -I nose
+deploy: 
+	pip install $(RUN_OPTIONS) --upgrade ../../release/python/$(RELEASE_VERSION)/$(PACKAGE)

src/py/README.md

@@ -10,7 +10,6 @@ nosetests test/
 
 ## Notes about Python dependencies
 * This extension is targeted at production databases. Therefore certain restrictions must be assumed about the production environment vs other experimental environments.
-* We're using `pip` and `virtualenv` to generate a suitable isolated environment for python code that has  all the dependencies
 * Every dependency should be:
   - Added to the `setup.py` file
   - Installed through it
@@ -30,21 +29,7 @@ PySAL 1.10 or later, so we'll stick to 1.9.1.
 apt-get install -y python-scipy
 ```
 
-We'll use virtual environments to install our packages,
-but configued to use also system modules so that the
-mentioned scipy and numpy are used.
-
-    # Create a virtual environment for python
-    $ virtualenv --system-site-packages dev
-
-    # Activate the virtualenv
-    $ source dev/bin/activate
-
-    # Install all the requirements
-    # expect this to take a while, as it will trigger a few compilations
-    (dev) $ pip install -I ./crankshaft
-
-#### Test the libraries with that virtual env
+#### Test the libraries
 
 ##### Test numpy library dependency:
 

src/py/crankshaft/crankshaft/__init__.py

@@ -1,2 +1,5 @@
-import random_seeds
-import clustering
+"""Import all modules"""
+import crankshaft.random_seeds
+import crankshaft.clustering
+import crankshaft.space_time_dynamics
+import crankshaft.segmentation

src/py/crankshaft/crankshaft/clustering/__init__.py

@@ -1 +1,3 @@
+"""Import all functions from for clustering"""
 from moran import *
+from kmeans import *

src/py/crankshaft/crankshaft/clustering/kmeans.py

@@ -0,0 +1,18 @@
+from sklearn.cluster import KMeans
+import plpy
+
+def kmeans(query, no_clusters, no_init=20):
+    data = plpy.execute('''select array_agg(cartodb_id order by cartodb_id) as ids,
+        array_agg(ST_X(the_geom) order by cartodb_id) xs,
+        array_agg(ST_Y(the_geom) order by cartodb_id) ys from ({query}) a
+        where the_geom is not null
+    '''.format(query=query))
+
+    xs  = data[0]['xs']
+    ys  = data[0]['ys']
+    ids = data[0]['ids']
+
+    km = KMeans(n_clusters= no_clusters, n_init=no_init)
+    labels = km.fit_predict(zip(xs,ys))
+    return zip(ids,labels)
+

src/py/crankshaft/crankshaft/clustering/moran.py

@@ -7,6 +7,7 @@ Moran's I geostatistics (global clustering & outliers presence)
 
 import pysal as ps
 import plpy
+from collections import OrderedDict
 
 # crankshaft module
 import crankshaft.pysal_utils as pu
@@ -21,11 +22,11 @@ def moran(subquery, attr_name,
      core clusters with PySAL.
     Andy Eschbacher
     """
-    qvals = {"id_col": id_col,
-             "attr1": attr_name,
-             "geom_col": geom_col,
-             "subquery": subquery,
-             "num_ngbrs": num_ngbrs}
+    qvals = OrderedDict([("id_col", id_col),
+                         ("attr1", attr_name),
+                         ("geom_col", geom_col),
+                         ("subquery", subquery),
+                         ("num_ngbrs", num_ngbrs)])
 
     query = pu.construct_neighbor_query(w_type, qvals)
 
@@ -65,11 +66,11 @@ def moran_local(subquery, attr,
     # geometries with attributes that are null are ignored
     # resulting in a collection of not as near neighbors
 
-    qvals = {"id_col": id_col,
-             "attr1": attr,
-             "geom_col": geom_col,
-             "subquery": subquery,
-             "num_ngbrs": num_ngbrs}
+    qvals = OrderedDict([("id_col", id_col),
+                         ("attr1", attr),
+                         ("geom_col", geom_col),
+                         ("subquery", subquery),
+                         ("num_ngbrs", num_ngbrs)])
 
     query = pu.construct_neighbor_query(w_type, qvals)
 
@@ -101,12 +102,12 @@ def moran_rate(subquery, numerator, denominator,
     Moran's I Rate (global)
     Andy Eschbacher
     """
-    qvals = {"id_col": id_col,
-             "attr1": numerator,
-             "attr2": denominator,
-             "geom_col": geom_col,
-             "subquery": subquery,
-             "num_ngbrs": num_ngbrs}
+    qvals = OrderedDict([("id_col", id_col),
+                         ("attr1", numerator),
+                         ("attr2", denominator)
+                         ("geom_col", geom_col),
+                         ("subquery", subquery),
+                         ("num_ngbrs", num_ngbrs)])
 
     query = pu.construct_neighbor_query(w_type, qvals)
 
@@ -145,13 +146,14 @@ def moran_local_rate(subquery, numerator, denominator,
     # geometries with values that are null are ignored
     # resulting in a collection of not as near neighbors
 
-    query = pu.construct_neighbor_query(w_type,
-                                     {"id_col": id_col,
-                                      "numerator": numerator,
-                                      "denominator": denominator,
-                                      "geom_col": geom_col,
-                                      "subquery": subquery,
-                                      "num_ngbrs": num_ngbrs})
+    qvals = OrderedDict([("id_col", id_col),
+                         ("numerator", numerator),
+                         ("denominator", denominator),
+                         ("geom_col", geom_col),
+                         ("subquery", subquery),
+                         ("num_ngbrs", num_ngbrs)])
+
+    query = pu.construct_neighbor_query(w_type, qvals)
 
     try:
         result = plpy.execute(query)
@@ -174,7 +176,7 @@ def moran_local_rate(subquery, numerator, denominator,
     lisa = ps.esda.moran.Moran_Local_Rate(numer, denom, weight,
                                           permutations=permutations)
 
-    # find units of significance
+    # find quadrants for each geometry
     quads = quad_position(lisa.q)
 
     return zip(lisa.Is, quads, lisa.p_sim, weight.id_order, lisa.y)
@@ -186,12 +188,12 @@ def moran_local_bv(subquery, attr1, attr2,
     """
     plpy.notice('** Constructing query')
 
-    qvals = {"num_ngbrs": num_ngbrs,
-             "attr1": attr1,
-             "attr2": attr2,
-             "subquery": subquery,
-             "geom_col": geom_col,
-             "id_col": id_col}
+    qvals = OrderedDict([("id_col", id_col),
+                         ("attr1", attr1),
+                         ("attr2", attr2),
+                         ("geom_col", geom_col),
+                         ("subquery", subquery),
+                         ("num_ngbrs", num_ngbrs)])
 
     query = pu.construct_neighbor_query(w_type, qvals)
 

src/py/crankshaft/crankshaft/pysal_utils/__init__.py

@@ -1 +1,2 @@
-from pysal_utils import *
+"""Import all functions for pysal_utils"""
+from crankshaft.pysal_utils.pysal_utils import *

src/py/crankshaft/crankshaft/pysal_utils/pysal_utils.py

@@ -1,5 +1,6 @@
 """
-    Utilities module for generic PySAL functionality, mainly centered on translating queries into numpy arrays or PySAL weights objects
+    Utilities module for generic PySAL functionality, mainly centered on
+      translating queries into numpy arrays or PySAL weights objects
 """
 
 import numpy as np
@@ -20,19 +21,23 @@ def construct_neighbor_query(w_type, query_vals):
 def get_weight(query_res, w_type='knn', num_ngbrs=5):
     """
         Construct PySAL weight from return value of query
-        @param query_res: query results with attributes and neighbors
+        @param query_res dict-like: query results with attributes and neighbors
     """
-    if w_type.lower() == 'knn':
-        row_normed_weights = [1.0 / float(num_ngbrs)] * num_ngbrs
-        weights = {x['id']: row_normed_weights for x in query_res}
-    else:
-        weights = {x['id']: [1.0 / len(x['neighbors'])] * len(x['neighbors'])
-                            if len(x['neighbors']) > 0
-                            else [] for x in query_res}
+    # if w_type.lower() == 'knn':
+    #     row_normed_weights = [1.0 / float(num_ngbrs)] * num_ngbrs
+    #     weights = {x['id']: row_normed_weights for x in query_res}
+    # else:
+    #     weights = {x['id']: [1.0 / len(x['neighbors'])] * len(x['neighbors'])
+    #                         if len(x['neighbors']) > 0
+    #                         else [] for x in query_res}
 
     neighbors = {x['id']: x['neighbors'] for x in query_res}
+    print 'len of neighbors: %d' % len(neighbors)
 
-    return ps.W(neighbors, weights)
+    built_weight = ps.W(neighbors)
+    built_weight.transform = 'r'
+
+    return built_weight
 
 def query_attr_select(params):
     """
@@ -41,32 +46,63 @@ def query_attr_select(params):
                        table name, etc.)
     """
 
-    attrs = [k for k in params
-             if k not in ('id_col', 'geom_col', 'subquery', 'num_ngbrs')]
-
-    template = "i.\"{%(col)s}\"::numeric As attr%(alias_num)s, "
-
     attr_string = ""
+    template = "i.\"%(col)s\"::numeric As attr%(alias_num)s, "
 
-    for idx, val in enumerate(sorted(attrs)):
-        attr_string += template % {"col": val, "alias_num": idx + 1}
+    if 'time_cols' in params:
+        ## if markov analysis
+        attrs = params['time_cols']
+
+        for idx, val in enumerate(attrs):
+            attr_string += template % {"col": val, "alias_num": idx + 1}
+    else:
+        ## if moran's analysis
+        attrs = [k for k in params
+                 if k not in ('id_col', 'geom_col', 'subquery', 'num_ngbrs', 'subquery')]
+
+        for idx, val in enumerate(sorted(attrs)):
+            attr_string += template % {"col": params[val], "alias_num": idx + 1}
 
     return attr_string
 
 def query_attr_where(params):
     """
+      Construct where conditions when building neighbors query
         Create portion of WHERE clauses for weeding out NULL-valued geometries
+        Input: dict of params:
+            {'subquery': ...,
+             'numerator': 'data1',
+             'denominator': 'data2',
+             '': ...}
+        Output: 'idx_replace."data1" IS NOT NULL AND idx_replace."data2" IS NOT NULL'
+        Input:
+        {'subquery': ...,
+         'time_cols': ['time1', 'time2', 'time3'],
+         'etc': ...}
+        Output: 'idx_replace."time1" IS NOT NULL AND idx_replace."time2" IS NOT
+          NULL AND idx_replace."time3" IS NOT NULL'
     """
-    attrs = sorted([k for k in params
-                    if k not in ('id_col', 'geom_col', 'subquery', 'num_ngbrs')])
-
     attr_string = []
+    template = "idx_replace.\"%s\" IS NOT NULL"
 
-    for attr in attrs:
-        attr_string.append("idx_replace.\"{%s}\" IS NOT NULL" % attr)
+    if 'time_cols' in params:
+        ## markov where clauses
+        attrs = params['time_cols']
+        # add values to template
+        for attr in attrs:
+            attr_string.append(template % attr)
+    else:
+        ## moran where clauses
 
-    if len(attrs) == 2:
-        attr_string.append("idx_replace.\"{%s}\" <> 0" % attrs[1])
+        # get keys
+        attrs = sorted([k for k in params
+                        if k not in ('id_col', 'geom_col', 'subquery', 'num_ngbrs', 'subquery')])
+        # add values to template
+        for attr in attrs:
+            attr_string.append(template % params[attr])
+
+        if len(attrs) == 2:
+            attr_string.append("idx_replace.\"%s\" <> 0" % params[attrs[1]])
 
     out = " AND ".join(attr_string)
 

src/py/crankshaft/crankshaft/random_seeds.py

@@ -1,3 +1,4 @@
+"""Random seed generator used for non-deterministic functions in crankshaft"""
 import random
 import numpy
 

src/py/crankshaft/crankshaft/segmentation/__init__.py

@@ -0,0 +1 @@
+from segmentation import * 

src/py/crankshaft/crankshaft/segmentation/segmentation.py

@@ -0,0 +1,176 @@
+"""
+Segmentation creation and prediction
+"""
+
+import sklearn
+import numpy as np
+import plpy
+from sklearn.ensemble import GradientBoostingRegressor
+from sklearn import metrics
+from sklearn.cross_validation import train_test_split
+
+# Lower level functions
+#----------------------
+
+def replace_nan_with_mean(array):
+    """
+        Input:
+            @param array: an array of floats which may have null-valued entries
+        Output:
+            array with nans filled in with the mean of the dataset
+    """
+    # returns an array of rows and column indices
+    indices = np.where(np.isnan(array))
+
+    # iterate through entries which have nan values
+    for row, col in zip(*indices):
+            array[row, col] = np.mean(array[~np.isnan(array[:, col]), col])
+
+    return array
+
+def get_data(variable, feature_columns, query):
+    """
+        Fetch data from the database, clean, and package into
+          numpy arrays
+        Input:
+            @param variable: name of the target variable
+            @param feature_columns: list of column names
+            @param query: subquery that data is pulled from for the packaging
+        Output:
+            prepared data, packaged into NumPy arrays
+    """
+
+    columns = ','.join(['array_agg("{col}") As "{col}"'.format(col=col) for col in feature_columns])
+
+    try:
+        data = plpy.execute('''SELECT array_agg("{variable}") As target, {columns} FROM ({query}) As a'''.format(
+            variable=variable,
+            columns=columns,
+            query=query))
+    except Exception, e:
+        plpy.error('Failed to access data to build segmentation model: %s' % e)
+
+    # extract target data from plpy object
+    target = np.array(data[0]['target'])
+
+    # put n feature data arrays into an n x m array of arrays
+    features = np.column_stack([np.array(data[0][col], dtype=float) for col in feature_columns])
+
+    return replace_nan_with_mean(target), replace_nan_with_mean(features)
+
+# High level interface
+# --------------------
+
+def create_and_predict_segment_agg(target, features, target_features, target_ids, model_parameters):
+    """
+    Version of create_and_predict_segment that works on arrays that come stright form the SQL calling
+    the function.
+
+        Input:
+            @param target: The 1D array of lenth NSamples containing the target variable we want the model to predict
+            @param features: Thw 2D array of size NSamples * NFeatures that form the imput to the model
+            @param target_ids: A 1D array of target_ids that will be used to associate the results of the prediction with the rows which they come from
+            @param model_parameters: A dictionary containing parameters for the model.
+    """
+
+    clean_target = replace_nan_with_mean(target)
+    clean_features = replace_nan_with_mean(features)
+    target_features = replace_nan_with_mean(target_features)
+
+    model, accuracy = train_model(clean_target, clean_features, model_parameters, 0.2)
+    prediction = model.predict(target_features)
+    accuracy_array = [accuracy]*prediction.shape[0]
+    return zip(target_ids, prediction, np.full(prediction.shape, accuracy_array))
+
+
+
+def create_and_predict_segment(query, variable, target_query, model_params):
+    """
+    generate a segment with machine learning
+    Stuart Lynn
+    """
+
+    ## fetch column names
+    try:
+        columns = plpy.execute('SELECT * FROM ({query}) As a LIMIT 1  '.format(query=query))[0].keys()
+    except Exception, e:
+        plpy.error('Failed to build segmentation model: %s' % e)
+
+    ## extract column names to be used in building the segmentation model
+    feature_columns = set(columns) - set([variable, 'cartodb_id', 'the_geom', 'the_geom_webmercator'])
+    ## get data from database
+    target, features = get_data(variable, feature_columns, query)
+
+    model, accuracy = train_model(target, features, model_params, 0.2)
+    cartodb_ids, result = predict_segment(model, feature_columns, target_query)
+    accuracy_array = [accuracy]*result.shape[0]
+    return zip(cartodb_ids, result, accuracy_array)
+
+
+def train_model(target, features, model_params, test_split):
+    """
+        Train the Gradient Boosting model on the provided data and calculate the accuracy of the model
+        Input:
+            @param target: 1D Array of the variable that the model is to be trianed to predict
+            @param features: 2D Array NSamples * NFeatures to use in trining the model
+            @param model_params: A dictionary of model parameters, the full specification can be found on the
+                scikit learn page for [GradientBoostingRegressor](http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingRegressor.html)
+            @parma test_split: The fraction of the data to be withheld for testing the model / calculating the accuray
+    """
+    features_train, features_test, target_train, target_test = train_test_split(features, target, test_size=test_split)
+    model = GradientBoostingRegressor(**model_params)
+    model.fit(features_train, target_train)
+    accuracy = calculate_model_accuracy(model, features, target)
+    return model, accuracy
+
+def calculate_model_accuracy(model, features, target):
+    """
+        Calculate the mean squared error of the model prediction
+        Input:
+            @param model: model trained from input features
+            @param features: features to make a prediction from
+            @param target: target to compare prediction to
+        Output:
+            mean squared error of the model prection compared to the target
+    """
+    prediction = model.predict(features)
+    return metrics.mean_squared_error(prediction, target)
+
+def predict_segment(model, features, target_query):
+    """
+    Use the provided model to predict the values for the new feature set
+        Input:
+            @param model: The pretrained model
+            @features: A list of features to use in the model prediction (list of column names)
+            @target_query: The query to run to obtain the data to predict on and the cartdb_ids associated with it.
+    """
+
+    batch_size = 1000
+    joined_features = ','.join(['"{0}"::numeric'.format(a) for a in features])
+
+    try:
+        cursor = plpy.cursor('SELECT Array[{joined_features}] As features FROM ({target_query}) As a'.format(
+            joined_features=joined_features,
+            target_query=target_query))
+    except Exception, e:
+        plpy.error('Failed to build segmentation model: %s' % e)
+
+    results = []
+
+    while True:
+        rows = cursor.fetch(batch_size)
+        if not rows:
+            break
+        batch = np.row_stack([np.array(row['features'], dtype=float) for row in rows])
+
+        #Need to fix this. Should be global mean. This will cause weird effects
+        batch = replace_nan_with_mean(batch)
+        prediction = model.predict(batch)
+        results.append(prediction)
+
+    try:
+        cartodb_ids = plpy.execute('''SELECT array_agg(cartodb_id ORDER BY cartodb_id) As cartodb_ids FROM ({0}) As a'''.format(target_query))[0]['cartodb_ids']
+    except Exception, e:
+        plpy.error('Failed to build segmentation model: %s' % e)
+
+    return cartodb_ids, np.concatenate(results)

src/py/crankshaft/crankshaft/space_time_dynamics/__init__.py

@@ -0,0 +1,2 @@
+"""Import all functions from clustering libraries."""
+from markov import *

src/py/crankshaft/crankshaft/space_time_dynamics/markov.py

@@ -0,0 +1,189 @@
+"""
+Spatial dynamics measurements using Spatial Markov
+"""
+
+
+import numpy as np
+import pysal as ps
+import plpy
+import crankshaft.pysal_utils as pu
+
+def spatial_markov_trend(subquery, time_cols, num_classes=7,
+                         w_type='knn', num_ngbrs=5, permutations=0,
+                         geom_col='the_geom', id_col='cartodb_id'):
+    """
+        Predict the trends of a unit based on:
+        1. history of its transitions to different classes (e.g., 1st quantile -> 2nd quantile)
+        2. average class of its neighbors
+
+        Inputs:
+        @param subquery string: e.g., SELECT the_geom, cartodb_id,
+          interesting_time_column FROM table_name
+        @param time_cols list of strings: list of strings of column names
+        @param num_classes (optional): number of classes to break distribution
+          of values into. Currently uses quantile bins.
+        @param w_type string (optional): weight type ('knn' or 'queen')
+        @param num_ngbrs int (optional): number of neighbors (if knn type)
+        @param permutations int (optional): number of permutations for test
+          stats
+        @param geom_col string (optional): name of column which contains the
+          geometries
+        @param id_col string (optional): name of column which has the ids of
+          the table
+
+        Outputs:
+        @param trend_up float: probablity that a geom will move to a higher
+          class
+        @param trend_down float: probablity that a geom will move to a lower
+          class
+        @param trend float: (trend_up - trend_down) / trend_static
+        @param volatility float: a measure of the volatility based on
+          probability stddev(prob array)
+    """
+
+    if len(time_cols) < 2:
+        plpy.error('More than one time column needs to be passed')
+
+    qvals = {"id_col": id_col,
+             "time_cols": time_cols,
+             "geom_col": geom_col,
+             "subquery": subquery,
+             "num_ngbrs": num_ngbrs}
+
+    try:
+        query_result = plpy.execute(
+            pu.construct_neighbor_query(w_type, qvals)
+        )
+        if len(query_result) == 0:
+            return zip([None], [None], [None], [None], [None])
+    except plpy.SPIError, err:
+        plpy.debug('Query failed with exception %s: %s' % (err, pu.construct_neighbor_query(w_type, qvals)))
+        plpy.error('Query failed, check the input parameters')
+        return zip([None], [None], [None], [None], [None])
+
+    ## build weight
+    weights = pu.get_weight(query_result, w_type)
+    weights.transform = 'r'
+
+    ## prep time data
+    t_data = get_time_data(query_result, time_cols)
+
+    plpy.debug('shape of t_data %d, %d' % t_data.shape)
+    plpy.debug('number of weight objects: %d, %d' % (weights.sparse).shape)
+    plpy.debug('first num elements: %f' % t_data[0, 0])
+
+    sp_markov_result = ps.Spatial_Markov(t_data,
+                                         weights,
+                                         k=num_classes,
+                                         fixed=False,
+                                         permutations=permutations)
+
+    ## get lag classes
+    lag_classes = ps.Quantiles(
+        ps.lag_spatial(weights, t_data[:, -1]),
+        k=num_classes).yb
+
+    ## look up probablity distribution for each unit according to class and lag class
+    prob_dist = get_prob_dist(sp_markov_result.P,
+                              lag_classes,
+                              sp_markov_result.classes[:, -1])
+
+    ## find the ups and down and overall distribution of each cell
+    trend_up, trend_down, trend, volatility = get_prob_stats(prob_dist,
+                                                             sp_markov_result.classes[:, -1])
+
+    ## output the results
+    return zip(trend, trend_up, trend_down, volatility, weights.id_order)
+
+def get_time_data(markov_data, time_cols):
+    """
+        Extract the time columns and bin appropriately
+    """
+    num_attrs = len(time_cols)
+    return np.array([[x['attr' + str(i)] for x in markov_data]
+                     for i in range(1, num_attrs+1)], dtype=float).transpose()
+
+## not currently used
+def rebin_data(time_data, num_time_per_bin):
+    """
+        Convert an n x l matrix into an (n/m) x l matrix where the values are
+         reduced (averaged) for the intervening states:
+          1 2 3 4    1.5 3.5
+          5 6 7 8 -> 5.5 7.5
+          9 8 7 6    8.5 6.5
+          5 4 3 2    4.5 2.5
+
+          if m = 2, the 4 x 4 matrix is transformed to a 2 x 4 matrix.
+
+        This process effectively resamples the data at a longer time span n
+         units longer than the input data.
+        For cases when there is a remainder (remainder(5/3) = 2), the remaining
+         two columns are binned together as the last time period, while the
+         first three are binned together for the first period.
+
+        Input:
+          @param time_data n x l  ndarray: measurements of an attribute at
+           different time intervals
+          @param num_time_per_bin int: number of columns to average into a new
+           column
+        Output:
+          ceil(n / m) x l ndarray of resampled time series
+    """
+
+    if time_data.shape[1] % num_time_per_bin == 0:
+        ## if fit is perfect, then use it
+        n_max = time_data.shape[1] / num_time_per_bin
+    else:
+        ## fit remainders into an additional column
+        n_max = time_data.shape[1] / num_time_per_bin + 1
+
+    return np.array([time_data[:, num_time_per_bin * i:num_time_per_bin * (i+1)].mean(axis=1)
+                     for i in range(n_max)]).T
+
+def get_prob_dist(transition_matrix, lag_indices, unit_indices):
+    """
+        Given an array of transition matrices, look up the probability
+        associated with the arrangements passed
+
+        Input:
+        @param transition_matrix ndarray[k,k,k]:
+        @param lag_indices ndarray:
+        @param unit_indices ndarray:
+
+        Output:
+        Array of probability distributions
+    """
+
+    return np.array([transition_matrix[(lag_indices[i], unit_indices[i])]
+                     for i in range(len(lag_indices))])
+
+def get_prob_stats(prob_dist, unit_indices):
+    """
+        get the statistics of the probability distributions
+
+        Outputs:
+            @param trend_up ndarray(float): sum of probabilities for upward
+               movement (relative to the unit index of that prob)
+            @param trend_down ndarray(float): sum of probabilities for downward
+               movement (relative to the unit index of that prob)
+            @param trend ndarray(float): difference of upward and downward
+               movements
+    """
+
+    num_elements = len(unit_indices)
+    trend_up = np.empty(num_elements, dtype=float)
+    trend_down = np.empty(num_elements, dtype=float)
+    trend = np.empty(num_elements, dtype=float)
+
+    for i in range(num_elements):
+        trend_up[i] = prob_dist[i, (unit_indices[i]+1):].sum()
+        trend_down[i] = prob_dist[i, :unit_indices[i]].sum()
+        if prob_dist[i, unit_indices[i]] > 0.0:
+            trend[i] = (trend_up[i] - trend_down[i]) / prob_dist[i, unit_indices[i]]
+        else:
+            trend[i] = None
+
+    ## calculate volatility of distribution
+    volatility = prob_dist.std(axis=1)
+
+    return trend_up, trend_down, trend, volatility

src/py/crankshaft/setup.py

@@ -40,9 +40,10 @@ setup(
 
     # The choice of component versions is dictated by what's
     # provisioned in the production servers.
-    install_requires=['pysal==1.9.1'],
+    # IMPORTANT NOTE: please don't change this line. Instead issue a ticket to systems for evaluation.
+    install_requires=['joblib==0.8.3', 'numpy==1.6.1', 'scipy==0.14.0', 'pysal==1.11.2', 'scikit-learn==0.14.1'],
 
-    requires=['pysal', 'numpy' ],
+    requires=['pysal', 'numpy', 'sklearn'],
 
     test_suite='test'
 )

src/py/crankshaft/test/fixtures/kmeans.json

@@ -0,0 +1 @@
+[{"xs": [9.917239463463458, 9.042767302696836, 10.798929825304187, 8.763751051762995, 11.383882954810852, 11.018206993460897, 8.939526075734316, 9.636159342565252, 10.136336896960058, 11.480610059427342, 12.115011910725082, 9.173267848893428, 10.239300931201738, 8.00012512174072, 8.979962292282131, 9.318376124429575, 10.82259513754284, 10.391747171927115, 10.04904588886165, 9.96007160443463, -0.78825626804569, -0.3511819898577426, -1.2796410003764271, -0.3977049391203402, 2.4792311265774667, 1.3670311632092624, 1.2963504112955613, 2.0404844103073025, -1.6439708506073223, 0.39122885445645805, 1.026031821452462, -0.04044477160482201, -0.7442346929085072, -0.34687120826243034, -0.23420359971379054, -0.5919629143336708, -0.202903054395391, -0.1893399644841902, 1.9331834251176807, -0.12321054392851609], "ys": [8.735627063679981, 9.857615954045011, 10.81439096759407, 10.586727233537191, 9.232919976568622, 11.54281262696508, 8.392787912674466, 9.355119689665944, 9.22380703532752, 10.542142541823122, 10.111980619367035, 10.760836265570738, 8.819773453269804, 10.25325722424816, 9.802077905695608, 8.955420161552611, 9.833801181904477, 10.491684241001613, 12.076108669877556, 11.74289693140474, -0.5685725015474191, -0.5715728344759778, -0.20180907868635137, 0.38431336480089595, -0.3402202083684184, -2.4652736827783586, 0.08295159401756182, 0.8503818775816505, 0.6488691600321166, 0.5794762568230527, -0.6770063922144103, -0.6557616416449478, -1.2834289177624947, 0.1096318195532717, -0.38986922166834853, -1.6224497706950238, 0.09429787743230483, 0.4005097316394031, -0.508002811195673, -1.2473463371366507], "ids": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39]}]␍

src/py/crankshaft/test/fixtures/markov.json

@@ -0,0 +1 @@
+[[0.11111111111111112, 0.10000000000000001, 0.0, 0.35213633723318016, 0], [0.03125, 0.030303030303030304, 0.0, 0.3850273981640871, 1], [0.03125, 0.030303030303030304, 0.0, 0.3850273981640871, 2], [0.0, 0.10000000000000001, 0.10000000000000001, 0.30331501776206204, 3], [0.0, 0.065217391304347824, 0.065217391304347824, 0.33605067580764519, 4], [-0.054054054054054057, 0.0, 0.05128205128205128, 0.37488547451276033, 5], [0.1875, 0.23999999999999999, 0.12, 0.23731835158706122, 6], [0.034482758620689655, 0.0625, 0.03125, 0.35388469167230169, 7], [0.030303030303030304, 0.078947368421052627, 0.052631578947368418, 0.33560628561957595, 8], [0.19047619047619049, 0.16, 0.0, 0.32594478059941379, 9], [-0.23529411764705882, 0.0, 0.19047619047619047, 0.31356338348865387, 10], [0.030303030303030304, 0.078947368421052627, 0.052631578947368418, 0.33560628561957595, 11], [-0.22222222222222224, 0.13333333333333333, 0.26666666666666666, 0.22310934040908681, 12], [0.027777777777777783, 0.11111111111111112, 0.088888888888888892, 0.30339641183779581, 13], [0.03125, 0.030303030303030304, 0.0, 0.3850273981640871, 14], [0.052631578947368425, 0.090909090909090912, 0.045454545454545456, 0.33352611505171165, 15], [-0.22222222222222224, 0.13333333333333333, 0.26666666666666666, 0.22310934040908681, 16], [-0.20512820512820512, 0.0, 0.1702127659574468, 0.32172013908826891, 17], [-0.20512820512820512, 0.0, 0.1702127659574468, 0.32172013908826891, 18], [-0.0625, 0.095238095238095233, 0.14285714285714285, 0.28634850244519822, 19], [0.0, 0.10000000000000001, 0.10000000000000001, 0.30331501776206204, 20], [0.078947368421052641, 0.073170731707317083, 0.0, 0.36451788667842738, 21], [0.030303030303030304, 0.078947368421052627, 0.052631578947368418, 0.33560628561957595, 22], [-0.16666666666666663, 0.18181818181818182, 0.27272727272727271, 0.20246415864836445, 23], [-0.22222222222222224, 0.13333333333333333, 0.26666666666666666, 0.22310934040908681, 24], [0.1875, 0.23999999999999999, 0.12, 0.23731835158706122, 25], [-0.20512820512820512, 0.0, 0.1702127659574468, 0.32172013908826891, 26], [-0.043478260869565216, 0.0, 0.041666666666666664, 0.37950991789118999, 27], [0.22222222222222221, 0.18181818181818182, 0.0, 0.31701083225750354, 28], [-0.054054054054054057, 0.0, 0.05128205128205128, 0.37488547451276033, 29], [-0.0625, 0.095238095238095233, 0.14285714285714285, 0.28634850244519822, 30], [0.0, 0.10000000000000001, 0.10000000000000001, 0.30331501776206204, 31], [0.030303030303030304, 0.078947368421052627, 0.052631578947368418, 0.33560628561957595, 32], [-0.0625, 0.095238095238095233, 0.14285714285714285, 0.28634850244519822, 33], [0.034482758620689655, 0.0625, 0.03125, 0.35388469167230169, 34], [0.0, 0.10000000000000001, 0.10000000000000001, 0.30331501776206204, 35], [-0.054054054054054057, 0.0, 0.05128205128205128, 0.37488547451276033, 36], [0.11111111111111112, 0.10000000000000001, 0.0, 0.35213633723318016, 37], [-0.22222222222222224, 0.13333333333333333, 0.26666666666666666, 0.22310934040908681, 38], [-0.0625, 0.095238095238095233, 0.14285714285714285, 0.28634850244519822, 39], [0.034482758620689655, 0.0625, 0.03125, 0.35388469167230169, 40], [0.11111111111111112, 0.10000000000000001, 0.0, 0.35213633723318016, 41], [0.052631578947368425, 0.090909090909090912, 0.045454545454545456, 0.33352611505171165, 42], [0.0, 0.0, 0.0, 0.40000000000000002, 43], [0.0, 0.065217391304347824, 0.065217391304347824, 0.33605067580764519, 44], [0.078947368421052641, 0.073170731707317083, 0.0, 0.36451788667842738, 45], [0.052631578947368425, 0.090909090909090912, 0.045454545454545456, 0.33352611505171165, 46], [-0.20512820512820512, 0.0, 0.1702127659574468, 0.32172013908826891, 47]]

src/py/crankshaft/test/mock_plpy.py

@@ -1,5 +1,16 @@
 import re
 
+class MockCursor:
+    def __init__(self, data):
+        self.cursor_pos = 0
+        self.data = data
+
+    def fetch(self, batch_size):
+        batch = self.data[self.cursor_pos : self.cursor_pos + batch_size]
+        self.cursor_pos += batch_size
+        return batch
+
+
 class MockPlPy:
     def __init__(self):
         self._reset()
@@ -24,9 +35,16 @@ class MockPlPy:
     def notice(self, msg):
         self.notices.append(msg)
 
+    def debug(self, msg):
+        self.notices.append(msg)
+
     def info(self, msg):
         self.infos.append(msg)
 
+    def cursor(self, query):
+        data = self.execute(query)
+        return MockCursor(data)
+
     def execute(self, query): # TODO: additional arguments
        for result in self.results:
           if result[0].match(query):

src/py/crankshaft/test/test_cluster_kmeans.py

@@ -0,0 +1,38 @@
+import unittest
+import numpy as np
+
+
+# from mock_plpy import MockPlPy
+# plpy = MockPlPy()
+#
+# import sys
+# sys.modules['plpy'] = plpy
+from helper import plpy, fixture_file
+import numpy as np
+import crankshaft.clustering as cc
+import crankshaft.pysal_utils as pu
+from crankshaft import random_seeds
+import json
+
+class KMeansTest(unittest.TestCase):
+    """Testing class for Moran's I functions"""
+
+    def setUp(self):
+        plpy._reset()
+        self.cluster_data = json.loads(open(fixture_file('kmeans.json')).read())
+        self.params = {"subquery": "select * from table",
+                       "no_clusters": "10"
+                       }
+
+    def test_kmeans(self):
+        data = self.cluster_data
+        plpy._define_result('select' ,data)
+        clusters = cc.kmeans('subquery', 2)
+        labels  = [a[1] for a in clusters]
+        c1 = [a for a in clusters if a[1]==0]
+        c2 = [a for a in clusters if a[1]==1]
+
+        self.assertEqual(len(np.unique(labels)),2)
+        self.assertEqual(len(c1),20)
+        self.assertEqual(len(c2),20)
+

src/py/crankshaft/test/test_clustering_moran.py

@@ -25,6 +25,11 @@ class MoranTest(unittest.TestCase):
                        "subquery": "SELECT * FROM a_list",
                        "geom_col": "the_geom",
                        "num_ngbrs": 321}
+        self.params_markov = {"id_col": "cartodb_id",
+                              "time_cols": ["_2013_dec", "_2014_jan", "_2014_feb"],
+                              "subquery": "SELECT * FROM a_list",
+                              "geom_col": "the_geom",
+                              "num_ngbrs": 321}
         self.neighbors_data = json.loads(open(fixture_file('neighbors.json')).read())
         self.moran_data = json.loads(open(fixture_file('moran.json')).read())
 

src/py/crankshaft/test/test_pysal_utils.py

@@ -15,22 +15,38 @@ class PysalUtilsTest(unittest.TestCase):
                        "geom_col": "the_geom",
                        "num_ngbrs": 321}
 
+        self.params_array = {"id_col": "cartodb_id",
+                             "time_cols": ["_2013_dec", "_2014_jan", "_2014_feb"],
+                             "subquery": "SELECT * FROM a_list",
+                             "geom_col": "the_geom",
+                             "num_ngbrs": 321}
+
     def test_query_attr_select(self):
         """Test query_attr_select"""
 
-        ans = "i.\"{attr1}\"::numeric As attr1, " \
-              "i.\"{attr2}\"::numeric As attr2, "
+        ans = "i.\"andy\"::numeric As attr1, " \
+              "i.\"jay_z\"::numeric As attr2, "
+
+        ans_array = "i.\"_2013_dec\"::numeric As attr1, " \
+                    "i.\"_2014_jan\"::numeric As attr2, " \
+                    "i.\"_2014_feb\"::numeric As attr3, "
 
         self.assertEqual(pu.query_attr_select(self.params), ans)
+        self.assertEqual(pu.query_attr_select(self.params_array), ans_array)
 
     def test_query_attr_where(self):
         """Test pu.query_attr_where"""
 
-        ans = "idx_replace.\"{attr1}\" IS NOT NULL AND " \
-              "idx_replace.\"{attr2}\" IS NOT NULL AND " \
-              "idx_replace.\"{attr2}\" <> 0"
+        ans = "idx_replace.\"andy\" IS NOT NULL AND " \
+              "idx_replace.\"jay_z\" IS NOT NULL AND " \
+              "idx_replace.\"jay_z\" <> 0"
+
+        ans_array = "idx_replace.\"_2013_dec\" IS NOT NULL AND " \
+                    "idx_replace.\"_2014_jan\" IS NOT NULL AND " \
+                    "idx_replace.\"_2014_feb\" IS NOT NULL"
 
         self.assertEqual(pu.query_attr_where(self.params), ans)
+        self.assertEqual(pu.query_attr_where(self.params_array), ans_array)
 
     def test_knn(self):
         """Test knn neighbors constructor"""
@@ -53,8 +69,27 @@ class PysalUtilsTest(unittest.TestCase):
                     "i.\"jay_z\" IS NOT NULL AND " \
                     "i.\"jay_z\" <> 0 " \
               "ORDER BY i.\"cartodb_id\" ASC;"
+        
+        ans_array = "SELECT i.\"cartodb_id\" As id, " \
+              "i.\"_2013_dec\"::numeric As attr1, " \
+              "i.\"_2014_jan\"::numeric As attr2, " \
+              "i.\"_2014_feb\"::numeric As attr3, " \
+              "(SELECT ARRAY(SELECT j.\"cartodb_id\" " \
+                            "FROM (SELECT * FROM a_list) As j " \
+                            "WHERE i.\"cartodb_id\" <> j.\"cartodb_id\" AND " \
+                                  "j.\"_2013_dec\" IS NOT NULL AND " \
+                                  "j.\"_2014_jan\" IS NOT NULL AND " \
+                                  "j.\"_2014_feb\" IS NOT NULL " \
+                            "ORDER BY j.\"the_geom\" <-> i.\"the_geom\" ASC " \
+                            "LIMIT 321)) As neighbors " \
+              "FROM (SELECT * FROM a_list) As i " \
+              "WHERE i.\"_2013_dec\" IS NOT NULL AND " \
+                    "i.\"_2014_jan\" IS NOT NULL AND " \
+                    "i.\"_2014_feb\" IS NOT NULL "\
+              "ORDER BY i.\"cartodb_id\" ASC;"
 
         self.assertEqual(pu.knn(self.params), ans)
+        self.assertEqual(pu.knn(self.params_array), ans_array)
 
     def test_queen(self):
         """Test queen neighbors constructor"""

src/py/crankshaft/test/test_segmentation.py

@@ -0,0 +1,64 @@
+import unittest
+import numpy as np
+from helper import plpy, fixture_file
+import crankshaft.segmentation as segmentation
+import json
+
+class SegmentationTest(unittest.TestCase):
+    """Testing class for Moran's I functions"""
+
+    def setUp(self):
+        plpy._reset()
+
+    def generate_random_data(self,n_samples,random_state,  row_type=False):
+        x1 = random_state.uniform(size=n_samples)
+        x2 = random_state.uniform(size=n_samples)
+        x3 = random_state.randint(0, 4, size=n_samples)
+
+        y = x1+x2*x2+x3
+        cartodb_id  = range(len(x1))
+
+        if row_type:
+            return [ {'features': vals} for vals in zip(x1,x2,x3)], y
+        else:
+            return  [dict( zip(['x1','x2','x3','target', 'cartodb_id'],[x1,x2,x3,y,cartodb_id]))]
+
+    def test_replace_nan_with_mean(self):
+        test_array = np.array([1.2, np.nan, 3.2, np.nan, np.nan])
+
+    def test_create_and_predict_segment(self):
+        n_samples = 1000
+
+        random_state_train = np.random.RandomState(13)
+        random_state_test = np.random.RandomState(134)
+        training_data = self.generate_random_data(n_samples, random_state_train)
+        test_data, test_y = self.generate_random_data(n_samples, random_state_test, row_type=True)
+
+
+        ids =  [{'cartodb_ids': range(len(test_data))}]
+        rows =  [{'x1': 0,'x2':0,'x3':0,'y':0,'cartodb_id':0}]
+
+        plpy._define_result('select \* from  \(select \* from training\) a  limit 1',rows)
+        plpy._define_result('.*from \(select \* from training\) as a' ,training_data)
+        plpy._define_result('select array_agg\(cartodb\_id order by cartodb\_id\) as cartodb_ids from \(.*\) a',ids)
+        plpy._define_result('.*select \* from test.*' ,test_data)
+
+        model_parameters =  {'n_estimators': 1200,
+                             'max_depth': 3,
+                             'subsample' : 0.5,
+                             'learning_rate': 0.01,
+                             'min_samples_leaf': 1}
+
+        result = segmentation.create_and_predict_segment(
+                'select * from training',
+                'target',
+                'select * from test',
+                model_parameters)
+
+        prediction = [r[1] for r in result]
+
+        accuracy =np.sqrt(np.mean( np.square( np.array(prediction) - np.array(test_y))))
+
+        self.assertEqual(len(result),len(test_data))
+        self.assertTrue( result[0][2] < 0.01)
+        self.assertTrue( accuracy < 0.5*np.mean(test_y)  )

src/py/crankshaft/test/test_space_time_dynamics.py

@@ -0,0 +1,324 @@
+import unittest
+import numpy as np
+
+import unittest
+
+
+# from mock_plpy import MockPlPy
+# plpy = MockPlPy()
+#
+# import sys
+# sys.modules['plpy'] = plpy
+from helper import plpy, fixture_file
+
+import crankshaft.space_time_dynamics as std
+from crankshaft import random_seeds
+import json
+
+class SpaceTimeTests(unittest.TestCase):
+    """Testing class for Markov Functions."""
+
+    def setUp(self):
+        plpy._reset()
+        self.params = {"id_col": "cartodb_id",
+                       "time_cols": ['dec_2013', 'jan_2014', 'feb_2014'],
+                       "subquery": "SELECT * FROM a_list",
+                       "geom_col": "the_geom",
+                       "num_ngbrs": 321}
+        self.neighbors_data = json.loads(open(fixture_file('neighbors_markov.json')).read())
+        self.markov_data = json.loads(open(fixture_file('markov.json')).read())
+
+        self.time_data = np.array([i * np.ones(10, dtype=float) for i in range(10)]).T
+
+        self.transition_matrix = np.array([
+                [[ 0.96341463, 0.0304878 , 0.00609756, 0.        , 0.        ],
+                 [ 0.06040268, 0.83221477, 0.10738255, 0.        , 0.        ],
+                 [ 0.        , 0.14      , 0.74      , 0.12      , 0.        ],
+                 [ 0.        , 0.03571429, 0.32142857, 0.57142857, 0.07142857],
+                 [ 0.        , 0.        , 0.        , 0.16666667, 0.83333333]],
+                [[ 0.79831933, 0.16806723, 0.03361345, 0.        , 0.        ],
+                 [ 0.0754717 , 0.88207547, 0.04245283, 0.        , 0.        ],
+                 [ 0.00537634, 0.06989247, 0.8655914 , 0.05913978, 0.        ],
+                 [ 0.        , 0.        , 0.06372549, 0.90196078, 0.03431373],
+                 [ 0.        , 0.        , 0.        , 0.19444444, 0.80555556]],
+                [[ 0.84693878, 0.15306122, 0.        , 0.        , 0.        ],
+                 [ 0.08133971, 0.78947368, 0.1291866 , 0.        , 0.        ],
+                 [ 0.00518135, 0.0984456 , 0.79274611, 0.0984456 , 0.00518135],
+                 [ 0.        , 0.        , 0.09411765, 0.87058824, 0.03529412],
+                 [ 0.        , 0.        , 0.        , 0.10204082, 0.89795918]],
+                [[ 0.8852459 , 0.09836066, 0.        , 0.01639344, 0.        ],
+                 [ 0.03875969, 0.81395349, 0.13953488, 0.        , 0.00775194],
+                 [ 0.0049505 , 0.09405941, 0.77722772, 0.11881188, 0.0049505 ],
+                 [ 0.        , 0.02339181, 0.12865497, 0.75438596, 0.09356725],
+                 [ 0.        , 0.        , 0.        , 0.09661836, 0.90338164]],
+                [[ 0.33333333, 0.66666667, 0.        , 0.        , 0.        ],
+                 [ 0.0483871 , 0.77419355, 0.16129032, 0.01612903, 0.        ],
+                 [ 0.01149425, 0.16091954, 0.74712644, 0.08045977, 0.        ],
+                 [ 0.        , 0.01036269, 0.06217617, 0.89637306, 0.03108808],
+                 [ 0.        , 0.        , 0.        , 0.02352941, 0.97647059]]]
+                 )
+
+    def test_spatial_markov(self):
+        """Test Spatial Markov."""
+        data = [ { 'id': d['id'],
+                   'attr1': d['y1995'],
+                   'attr2': d['y1996'],
+                   'attr3': d['y1997'],
+                   'attr4': d['y1998'],
+                   'attr5': d['y1999'],
+                   'attr6': d['y2000'],
+                   'attr7': d['y2001'],
+                   'attr8': d['y2002'],
+                   'attr9': d['y2003'],
+                   'attr10': d['y2004'],
+                   'attr11': d['y2005'],
+                   'attr12': d['y2006'],
+                   'attr13': d['y2007'],
+                   'attr14': d['y2008'],
+                   'attr15': d['y2009'],
+                   'neighbors': d['neighbors'] } for d in self.neighbors_data]
+        print(str(data[0]))
+        plpy._define_result('select', data)
+        random_seeds.set_random_seeds(1234)
+
+        result = std.spatial_markov_trend('subquery', ['y1995', 'y1996', 'y1997', 'y1998', 'y1999', 'y2000', 'y2001', 'y2002', 'y2003', 'y2004', 'y2005', 'y2006', 'y2007', 'y2008', 'y2009'], 5, 'knn', 5, 0, 'the_geom', 'cartodb_id')
+
+        self.assertTrue(result != None)
+        result = [(row[0], row[1], row[2], row[3], row[4]) for row in result]
+        print result[0]
+        expected = self.markov_data
+        for ([res_trend, res_up, res_down, res_vol, res_id],
+             [exp_trend, exp_up, exp_down, exp_vol, exp_id]
+             ) in zip(result, expected):
+            self.assertAlmostEqual(res_trend, exp_trend)
+
+    def test_get_time_data(self):
+        """Test get_time_data"""
+        data = [ { 'attr1': d['y1995'],
+                   'attr2': d['y1996'],
+                   'attr3': d['y1997'],
+                   'attr4': d['y1998'],
+                   'attr5': d['y1999'],
+                   'attr6': d['y2000'],
+                   'attr7': d['y2001'],
+                   'attr8': d['y2002'],
+                   'attr9': d['y2003'],
+                   'attr10': d['y2004'],
+                   'attr11': d['y2005'],
+                   'attr12': d['y2006'],
+                   'attr13': d['y2007'],
+                   'attr14': d['y2008'],
+                   'attr15': d['y2009'] } for d in self.neighbors_data]
+
+        result = std.get_time_data(data, ['y1995', 'y1996', 'y1997', 'y1998', 'y1999', 'y2000', 'y2001', 'y2002', 'y2003', 'y2004', 'y2005', 'y2006', 'y2007', 'y2008', 'y2009'])
+
+        ## expected was prepared from PySAL example:
+        ### f = ps.open(ps.examples.get_path("usjoin.csv"))
+        ### pci = np.array([f.by_col[str(y)] for y in range(1995, 2010)]).transpose()
+        ### rpci = pci / (pci.mean(axis = 0))
+
+        expected = np.array([[ 0.87654416, 0.863147, 0.85637567, 0.84811668, 0.8446154,  0.83271652
+        ,  0.83786314, 0.85012593, 0.85509656, 0.86416612, 0.87119375, 0.86302631
+        ,  0.86148267, 0.86252252, 0.86746356],
+         [ 0.9188951,  0.91757931, 0.92333258, 0.92517289, 0.92552388, 0.90746978
+        ,  0.89830489, 0.89431991, 0.88924794, 0.89815176, 0.91832091, 0.91706054
+        ,  0.90139505, 0.87897455, 0.86216858],
+         [ 0.82591007, 0.82548596, 0.81989793, 0.81503235, 0.81731522, 0.78964559
+        ,  0.80584442, 0.8084998,  0.82258551, 0.82668196, 0.82373724, 0.81814804
+        ,  0.83675961, 0.83574199, 0.84647177],
+         [ 1.09088176, 1.08537689, 1.08456418, 1.08415404, 1.09898841, 1.14506948
+        ,  1.12151133, 1.11160697, 1.10888621, 1.11399806, 1.12168029, 1.13164797
+        ,  1.12958508, 1.11371818, 1.09936775],
+         [ 1.10731446, 1.11373944, 1.13283638, 1.14472559, 1.15910025, 1.16898201
+        ,  1.17212488, 1.14752303, 1.11843284, 1.11024964, 1.11943471, 1.11736468
+        ,  1.10863242, 1.09642516, 1.07762337],
+         [ 1.42269757, 1.42118434, 1.44273502, 1.43577571, 1.44400684, 1.44184737
+        ,  1.44782832, 1.41978227, 1.39092208, 1.4059372,  1.40788646, 1.44052766
+        ,  1.45241216, 1.43306098, 1.4174431 ],
+         [ 1.13073885, 1.13110513, 1.11074708, 1.13364636, 1.13088149, 1.10888138
+        ,  1.11856629, 1.13062931, 1.11944984, 1.12446239, 1.11671008, 1.10880034
+        ,  1.08401709, 1.06959206, 1.07875225],
+         [ 1.04706124, 1.04516831, 1.04253372, 1.03239987, 1.02072545, 0.99854316
+        ,  0.9880258,  0.99669587, 0.99327676, 1.01400905, 1.03176742, 1.040511
+        ,  1.01749645, 0.9936394,  0.98279746],
+         [ 0.98996986, 1.00143564, 0.99491,  1.00188408, 1.00455845, 0.99127006
+        ,  0.97925917, 0.9683482,  0.95335147, 0.93694787, 0.94308213, 0.92232874
+        ,  0.91284091, 0.89689833, 0.88928858],
+         [ 0.87418391, 0.86416601, 0.84425695, 0.8404494,  0.83903044, 0.8578708
+        ,  0.86036185, 0.86107306, 0.8500772,  0.86981998, 0.86837929, 0.87204141
+        ,  0.86633032, 0.84946077, 0.83287146],
+         [ 1.14196118, 1.14660262, 1.14892712, 1.14909594, 1.14436624, 1.14450183
+        ,  1.12349752, 1.12596664, 1.12213996, 1.1119989,  1.10257792, 1.10491258
+        ,  1.11059842, 1.10509795, 1.10020097],
+         [ 0.97282463, 0.96700147, 0.96252588, 0.9653878,  0.96057687, 0.95831051
+        ,  0.94480909, 0.94804195, 0.95430286, 0.94103989, 0.92122519, 0.91010201
+        ,  0.89280392, 0.89298243, 0.89165385],
+         [ 0.94325468, 0.96436902, 0.96455242, 0.95243009, 0.94117647, 0.9480927
+        ,  0.93539182, 0.95388718, 0.94597005, 0.96918424, 0.94781281, 0.93466815
+        ,  0.94281559, 0.96520315, 0.96715441],
+         [ 0.97478408, 0.98169225, 0.98712809, 0.98474769, 0.98559897, 0.98687073
+        ,  0.99237486, 0.98209969, 0.9877653,  0.97399471, 0.96910087, 0.98416665
+        ,  0.98423613, 0.99823861, 0.99545704],
+         [ 0.85570269, 0.85575915, 0.85986132, 0.85693406, 0.8538012,  0.86191535
+        ,  0.84981451, 0.85472102, 0.84564835, 0.83998883, 0.83478547, 0.82803648
+        ,  0.8198736,  0.82265395, 0.8399404 ],
+         [ 0.87022047, 0.85996258, 0.85961813, 0.85689572, 0.83947136, 0.82785597
+        ,  0.86008789, 0.86776298, 0.86720209, 0.8676334,  0.89179317, 0.94202108
+        ,  0.9422231,  0.93902708, 0.94479184],
+         [ 0.90134907, 0.90407738, 0.90403991, 0.90201769, 0.90399238, 0.90906632
+        ,  0.92693339, 0.93695966, 0.94242697, 0.94338265, 0.91981796, 0.91108804
+        ,  0.90543476, 0.91737138, 0.94793657],
+         [ 1.1977611,  1.18222564, 1.18439158, 1.18267865, 1.19286723, 1.20172869
+        ,  1.21328691, 1.22624778, 1.22397075, 1.23857042, 1.24419893, 1.23929384
+        ,  1.23418676, 1.23626739, 1.26754398],
+         [ 1.24919678, 1.25754773, 1.26991161, 1.28020651, 1.30625667, 1.34790023
+        ,  1.34399863, 1.32575181, 1.30795492, 1.30544841, 1.30303302, 1.32107766
+        ,  1.32936244, 1.33001241, 1.33288462],
+         [ 1.06768004, 1.03799276, 1.03637303, 1.02768449, 1.03296093, 1.05059016
+        ,  1.03405057, 1.02747623, 1.03162734, 0.9961416,  0.97356208, 0.94241549
+        ,  0.92754547, 0.92549227, 0.92138102],
+         [ 1.09475614, 1.11526796, 1.11654299, 1.13103948, 1.13143264, 1.13889622
+        ,  1.12442212, 1.13367018, 1.13982256, 1.14029944, 1.11979401, 1.10905389
+        ,  1.10577769, 1.11166825, 1.09985155],
+         [ 0.76530058, 0.76612841, 0.76542451, 0.76722683, 0.76014284, 0.74480073
+        ,  0.76098396, 0.76156903, 0.76651952, 0.76533288, 0.78205934, 0.76842416
+        ,  0.77487118, 0.77768683, 0.78801192],
+         [ 0.98391336, 0.98075816, 0.98295341, 0.97386015, 0.96913803, 0.97370819
+        ,  0.96419154, 0.97209861, 0.97441313, 0.96356162, 0.94745352, 0.93965462
+        ,  0.93069645, 0.94020973, 0.94358232],
+         [ 0.83561828, 0.82298088, 0.81738502, 0.81748588, 0.80904801, 0.80071489
+        ,  0.83358256, 0.83451613, 0.85175032, 0.85954307, 0.86790024, 0.87170334
+        ,  0.87863799, 0.87497981, 0.87888675],
+         [ 0.98845573, 1.02092428, 0.99665283, 0.99141823, 0.99386619, 0.98733195
+        ,  0.99644997, 0.99669587, 1.02559097, 1.01116651, 0.99988024, 0.97906749
+        ,  0.99323123, 1.00204939, 0.99602148],
+         [ 1.14930913, 1.15241949, 1.14300962, 1.14265542, 1.13984683, 1.08312397
+        ,  1.05192626, 1.04230892, 1.05577278, 1.08569751, 1.12443486, 1.08891079
+        ,  1.08603695, 1.05997314, 1.02160943],
+         [ 1.11368269, 1.1057147,  1.11893431, 1.13778669, 1.1432272,  1.18257029
+        ,  1.16226243, 1.16009196, 1.14467789, 1.14820235, 1.12386598, 1.12680236
+        ,  1.12357937, 1.1159258,  1.12570828],
+         [ 1.30379431, 1.30752186, 1.31206366, 1.31532267, 1.30625667, 1.31210239
+        ,  1.29989156, 1.29203193, 1.27183516, 1.26830786, 1.2617743,  1.28656675
+        ,  1.29734097, 1.29390205, 1.29345446],
+         [ 0.83953719, 0.82701448, 0.82006005, 0.81188876, 0.80294864, 0.78772975
+        ,  0.82848011, 0.8259679,  0.82435705, 0.83108634, 0.84373784, 0.83891093
+        ,  0.84349247, 0.85637272, 0.86539395],
+         [ 1.23450087, 1.2426022,  1.23537935, 1.23581293, 1.24522626, 1.2256767
+        ,  1.21126648, 1.19377804, 1.18355337, 1.19674434, 1.21536573, 1.23653297
+        ,  1.27962009, 1.27968392, 1.25907738],
+         [ 0.9769662,  0.97400719, 0.98035944, 0.97581531, 0.95543282, 0.96480308
+        ,  0.94686376, 0.93679073, 0.92540049, 0.92988835, 0.93442917, 0.92100464
+        ,  0.91475304, 0.90249622, 0.9021363 ],
+         [ 0.84986886, 0.8986851,  0.84295997, 0.87280534, 0.85659368, 0.88937573
+        ,  0.894401, 0.90448993, 0.95495898, 0.92698333, 0.94745352, 0.92562488
+        ,  0.96635366, 1.02520312, 1.0394296 ],
+         [ 1.01922808, 1.00258203, 1.00974428, 1.00303417, 0.99765073, 1.00759019
+        ,  0.99192968, 0.99747298, 0.99550759, 0.97583768, 0.9610168,  0.94779638
+        ,  0.93759089, 0.93353431, 0.94121705],
+         [ 0.86367411, 0.85558932, 0.85544346, 0.85103025, 0.84336613, 0.83434854
+        ,  0.85813595, 0.84667961, 0.84374558, 0.85951183, 0.87194227, 0.89455097
+        ,  0.88283929, 0.90349491, 0.90600675],
+         [ 1.00947534, 1.00411055, 1.00698819, 0.99513687, 0.99291086, 1.00581626
+        ,  0.98850522, 0.99291168, 0.98983209, 0.97511924, 0.96134615, 0.96382634
+        ,  0.95011401, 0.9434686,  0.94637765],
+         [ 1.05712571, 1.05459419, 1.05753012, 1.04880786, 1.05103857, 1.04800023
+        ,  1.03024941, 1.04200483, 1.0402554,  1.03296979, 1.02191682, 1.02476275
+        ,  1.02347523, 1.02517684, 1.04359571],
+         [ 1.07084189, 1.06669497, 1.07937623, 1.07387988, 1.0794043,  1.0531801
+        ,  1.07452771, 1.09383478, 1.1052447,  1.10322136, 1.09167939, 1.08772756
+        ,  1.08859544, 1.09177338, 1.1096083 ],
+         [ 0.86719222, 0.86628896, 0.86675156, 0.86425632, 0.86511809, 0.86287327
+        ,  0.85169796, 0.85411285, 0.84886336, 0.84517414, 0.84843858, 0.84488343
+        ,  0.83374329, 0.82812044, 0.82878599],
+         [ 0.88389211, 0.92288667, 0.90282398, 0.91229186, 0.92023286, 0.92652175
+        ,  0.94278865, 0.93682452, 0.98655146, 0.992237, 0.9798497,  0.93869677
+        ,  0.96947771, 1.00362626, 0.98102351],
+         [ 0.97082064, 0.95320233, 0.94534081, 0.94215593, 0.93967,  0.93092109
+        ,  0.92662519, 0.93412152, 0.93501274, 0.92879506, 0.92110542, 0.91035556
+        ,  0.90430364, 0.89994694, 0.90073864],
+         [ 0.95861858, 0.95774543, 0.98254811, 0.98919472, 0.98684824, 0.98882205
+        ,  0.97662234, 0.95601578, 0.94905385, 0.94934888, 0.97152609, 0.97163004
+        ,  0.9700702,  0.97158948, 0.95884908],
+         [ 0.83980439, 0.84726737, 0.85747,  0.85467221, 0.8556751,  0.84818516
+        ,  0.85265681, 0.84502402, 0.82645665, 0.81743586, 0.83550406, 0.83338919
+        ,  0.83511679, 0.82136617, 0.80921874],
+         [ 0.95118156, 0.9466212,  0.94688098, 0.9508583,  0.9512441,  0.95440787
+        ,  0.96364363, 0.96804412, 0.97136214, 0.97583768, 0.95571724, 0.96895368
+        ,  0.97001634, 0.97082733, 0.98782366],
+         [ 1.08910044, 1.08248968, 1.08492895, 1.08656923, 1.09454249, 1.10558188
+        ,  1.1214086,  1.12292577, 1.13021031, 1.13342735, 1.14686068, 1.14502975
+        ,  1.14474747, 1.14084037, 1.16142926],
+         [ 1.06336033, 1.07365823, 1.08691496, 1.09764846, 1.11669863, 1.11856702
+        ,  1.09764283, 1.08815849, 1.08044313, 1.09278827, 1.07003204, 1.08398066
+        ,  1.09831768, 1.09298232, 1.09176125],
+         [ 0.79772065, 0.78829196, 0.78581151, 0.77615922, 0.77035744, 0.77751194
+        ,  0.79902974, 0.81437881, 0.80788828, 0.79603865, 0.78966436, 0.79949807
+        ,  0.80172182, 0.82168155, 0.85587911],
+         [ 1.0052447,  1.00007696, 1.00475899, 1.00613942, 1.00639561, 1.00162979
+        ,  0.99860739, 1.00814981, 1.00574316, 0.99030032, 0.97682565, 0.97292596
+        ,  0.96519561, 0.96173403, 0.95890284],
+         [ 0.95808419, 0.9382568,  0.9654441,  0.95561201, 0.96987289, 0.96608031
+        ,  0.99727185, 1.00781194, 1.03484236, 1.05333619, 1.0983263,  1.1704974
+        ,  1.17025154, 1.18730553, 1.14242645]])
+
+        self.assertTrue(np.allclose(result, expected))
+        self.assertTrue(type(result) == type(expected))
+        self.assertTrue(result.shape == expected.shape)
+
+    def test_rebin_data(self):
+        """Test rebin_data"""
+        ## sample in double the time (even case since 10 % 2 = 0):
+        ##   (0+1)/2, (2+3)/2, (4+5)/2, (6+7)/2, (8+9)/2
+        ## = 0.5,     2.5,     4.5,     6.5,     8.5
+        ans_even = np.array([(i + 0.5) * np.ones(10, dtype=float)
+                             for i in range(0, 10, 2)]).T
+
+        self.assertTrue(np.array_equal(std.rebin_data(self.time_data, 2), ans_even))
+
+        ## sample in triple the time (uneven since 10 % 3 = 1):
+        ##   (0+1+2)/3, (3+4+5)/3, (6+7+8)/3, (9)/1
+        ## = 1,         4,         7,         9
+        ans_odd  = np.array([i * np.ones(10, dtype=float)
+                             for i in (1, 4, 7, 9)]).T
+        self.assertTrue(np.array_equal(std.rebin_data(self.time_data, 3), ans_odd))
+
+    def test_get_prob_dist(self):
+        """Test get_prob_dist"""
+        lag_indices = np.array([1, 2, 3, 4])
+        unit_indices = np.array([1, 3, 2, 4])
+        answer = np.array([
+            [ 0.0754717 , 0.88207547, 0.04245283, 0.        , 0.        ],
+            [ 0.        , 0.        , 0.09411765, 0.87058824, 0.03529412],
+            [ 0.0049505 , 0.09405941, 0.77722772, 0.11881188, 0.0049505 ],
+            [ 0.        , 0.        , 0.        , 0.02352941, 0.97647059]
+        ])
+        result = std.get_prob_dist(self.transition_matrix, lag_indices, unit_indices)
+
+        self.assertTrue(np.array_equal(result, answer))
+
+    def test_get_prob_stats(self):
+        """Test get_prob_stats"""
+
+        probs = np.array([
+            [ 0.0754717 , 0.88207547, 0.04245283, 0.        , 0.        ],
+            [ 0.        , 0.        , 0.09411765, 0.87058824, 0.03529412],
+            [ 0.0049505 , 0.09405941, 0.77722772, 0.11881188, 0.0049505 ],
+            [ 0.        , 0.        , 0.        , 0.02352941, 0.97647059]
+        ])
+        unit_indices = np.array([1, 3, 2, 4])
+        answer_up = np.array([0.04245283, 0.03529412, 0.12376238, 0.])
+        answer_down = np.array([0.0754717, 0.09411765, 0.0990099, 0.02352941])
+        answer_trend = np.array([-0.03301887 / 0.88207547, -0.05882353 / 0.87058824,  0.02475248 / 0.77722772, -0.02352941 / 0.97647059])
+        answer_volatility = np.array([ 0.34221495,  0.33705421,  0.29226542,  0.38834223])
+
+        result = std.get_prob_stats(probs, unit_indices)
+        result_up = result[0]
+        result_down = result[1]
+        result_trend = result[2]
+        result_volatility = result[3]
+
+        self.assertTrue(np.allclose(result_up, answer_up))
+        self.assertTrue(np.allclose(result_down, answer_down))
+        self.assertTrue(np.allclose(result_trend, answer_trend))
+        self.assertTrue(np.allclose(result_volatility, answer_volatility))