def timings():
print 'loading'
data = pg.VectorData(
r"C:\Users\kimok\Desktop\gazetteer data\raw\global_urban_extent_polygons_v1.01.shp",
encoding='latin')
#data = list(pg.VectorData(r"C:\Users\kimok\Desktop\gazetteer data\raw\ne_10m_admin_0_countries.shp", encoding='latin')) * 3
print len(data)
print 'making shapely (no copy)'
t = time()
shapelys = [asShape(f.geometry) for f in data]
print time() - t
print 'making shapely (copy)'
t = time()
shapelys = [shape(f.geometry) for f in data]
print time() - t
print 'dump geoj (interface)'
t = time()
geojs = [s.__geo_interface__ for s in shapelys]
print time() - t
##print 'dump geoj (mapping)'
##t = time()
##geojs = [mapping(s) for s in shapelys]
##print time()-t
print 'load geoj asShape (no copy)'
t = time()
shapelys = [asShape(geoj) for geoj in geojs]
print time() - t
print 'load geoj shape (copy)'
t = time()
shapelys = [shape(geoj) for geoj in geojs]
print time() - t
print 'dump wkt'
t = time()
wkts = [s.wkt for s in shapelys]
print time() - t
print 'load wkt'
t = time()
shapelys = [wkt_loads(wkt) for wkt in wkts]
print time() - t
print 'dump wkb'
t = time()
wkbs = [s.wkb for s in shapelys]
print time() - t
print 'load wkb'
t = time()
shapelys = [wkb_loads(wkb) for wkb in wkbs]
print time() - t
def viewresult(res):
vz = pg.VectorData()
for r in res:
props = [] #r[:-1] if len(r) > 1 else []
if r[-1]:
geoj = json.loads(
r[-1]
) # assumes last value of row is returned as geojson string
vz.add_feature(props, geoj)
vz.view()
def get_data(path, iso):
d = pg.VectorData(fields=['adm_level'])
reader = shapefile.Reader(path)
max_level = 0
for rec in reader.iterRecords(fields=['iso_countr', 'adm_level']):
if rec['iso_countr'] == iso:
shape = reader.shape(rec.oid)
#rec = reader.record()
d.add_feature([rec['adm_level']], shape.__geo_interface__)
max_level = max(max_level, rec['adm_level'])
print('max level', max_level)
return d
def load(self):
# TODO: should prob be threaded
filepath = self.filepath.get()
data = None
for typ in pg.vector.loader.file_extensions:
if filepath.endswith(typ):
data = pg.VectorData(filepath, encoding=self.encoding.get())
break
for typ in pg.raster.loader.file_extensions:
if filepath.endswith(typ):
data = pg.RasterData(filepath)
break
self.onsuccess(data)
def accept(self):
# check geoj
geoj = self.draw_geoj
geoj['coordinates'].pop(
-1) # drop the latest one, either empty or unfinished
if geoj['coordinates']:
# make vectordata
d = pg.VectorData()
d.add_feature([], geoj)
# add to renderer
self.mapview.renderer.add_layer(d)
for cntr in self.mapview.controls:
if isinstance(cntr, StaticLayersControl):
cntr.update_layers()
# exit draw mode
self.cancel()
def view(results, text=False):
import pythongis as pg
# setup map
m = pg.renderer.Map()
m.add_layer(
r"C:\Users\kimok\Desktop\gazetteer data\raw\ne_10m_admin_0_countries.shp",
fillcolor='gray')
# options
kwargs = {}
if text:
kwargs.update(text=lambda f: f[text][:20],
textoptions=dict(textsize=3))
# add
d = pg.VectorData(fields=[])
for row in results:
d.add_feature([], row[-1].__geo_interface__)
m.add_layer(d, fillcolor='blue', **kwargs)
# view
m.view()
def sqlite_geoms():
print 'load shapefile'
t = time()
#data = pg.VectorData(r"C:\Users\kimok\Desktop\gazetteer data\raw\global_urban_extent_polygons_v1.01.shp", encoding='latin')
#data = pg.VectorData(r"C:\Users\kimok\Desktop\gazetteer data\raw\atlas_urban.geojson", encoding='latin')
data = pg.VectorData(
r"C:\Users\kimok\Desktop\gazetteer data\raw\global_settlement_points_v1.01.shp",
encoding='latin')
#data = pg.VectorData(r"C:\Users\kimok\Desktop\gazetteer data\raw\ne_10m_admin_0_countries.shp", encoding='latin')
print time() - t
print 'making shapely'
t = time()
shapelys = [
shape(f.geometry) for f in data
] # CRUCIAL SPEEDUP, SHAPELY SHOULD BE FROM SHAPE, NOT ASSHAPE WHICH IS INDIRECT REFERENCING
print time() - t
print 'dump wkb'
t = time()
wkbs = [s.wkb for s in shapelys]
print time() - t
print 'convert to binary'
from sqlite3 import Binary
t = time()
blobs = [Binary(wkb) for wkb in wkbs]
print time() - t
print 'insert wkb into db'
fields = ['ID', 'geom']
typs = ['int', 'BLOB']
w = Writer('testgeodb::data', fields=zip(fields, typs), replace=True)
t = time()
for i, blb in enumerate(blobs):
w.add([i, blb])
print time() - t
print 'load wkb from db'
t = time()
shapelys = [wkb_loads(bytes(blb)) for ID, blb in w.select('*')]
print time() - t
def view_footprints(self, mapIDs=None):
import pythongis as pg
render = pg.renderer.Map()
render.add_layer(r"C:\Users\kimok\Downloads\cshapes\cshapes.shp",
fillcolor=(222, 222, 222))
d = pg.VectorData(fields=['mapID', 'link'])
sql = 'select oid,link,xmin,ymin,xmax,ymax from maps'
if mapIDs is not None:
sql += ' where oid in ({})'.format(','.join(map(str, mapIDs)))
for oid, link, x1, y1, x2, y2 in self.query(sql):
if x1 is None: continue
row = [oid, link]
geoj = {
'type': 'Polygon',
'coordinates': [[(x1, y1), (x2, y1), (x2, y2), (x1, y2)]]
}
d.add_feature(row, geoj)
if len(d):
render.add_layer(d, fillcolor=(0, 200, 0, 100), outlinewidth=0.2)
render.zoom_bbox(*d.bbox)
render.view()
self._index.cur.execute('INSERT INTO links VALUES (?, ?)',
(self.nodeid, itemid))
######################################
# TESTING
if __name__ == '__main__':
import pythongis as pg
DEBUG = False
PROFILE = True
print 'loading'
d = pg.VectorData(
r"C:\Users\kimok\Downloads\ne_10m_admin_1_states_provinces (1)\ne_10m_admin_1_states_provinces.shp"
)
#d = pg.VectorData(r"C:\Users\kimok\Desktop\BIGDATA\gazetteer data\raw\global_settlement_points_v1.01.shp", encoding='latin')
items = [(f.id, f.bbox)
for f in d] # items = [(i+1, f.bbox) for i,f in enumerate(d)]
print len(items)
print 'building'
spindex = QuadTree(-180, -90, 180, 90)
if PROFILE:
import cProfile
prof = cProfile.Profile()
prof.enable()
spindex.build(items)
if PROFILE:
print prof.print_stats('cumtime')
import pythongis as pg
from time import time
# test distance
vect = pg.VectorData(r"C:\Users\kimo\Downloads\cshapes_0.6\cshapes.shp",
) #select=lambda f: f["GWCODE"]==666)
hist = vect.histogram("GWCODE")
#hist.view()
t = time()
distrast = pg.raster.analyzer.distance(vect,
bbox=[-180, 90, 180, -90],
width=72 * 10,
height=36 * 10)
#distrast = pg.RasterData("C:/Users/kimo/Desktop/world.jpg", bbox=[-180,90,180,-90], width=512, height=256)
print time() - t
hist = distrast.bands[0].histogram()
print hist
#hist.view()
#mapp = distrast.render()
mapp = pg.renderer.Map()
mapp.add_layer(distrast)
mapp.add_layer(vect, fillcolor=None)
#mapp.add_legend()
mapp.view()
import unittest
import pythongis as pg
# data
pointdata = pg.VectorData('data/ne_10m_populated_places_simple.shp',
encoding='latin')
linedata = pg.VectorData('data/ne_10m_railroads.shp', encoding='latin')
polygondata = pg.VectorData('data/ne_10m_admin_0_countries.shp',
encoding='latin')
rasterdata = pg.RasterData('data/land_shallow_topo_2048.png')
rasterdata.set_geotransform(width=2048,
height=1024,
affine=[0.175781250, 0, -180, 0, -0.175781250, 90])
# base class
class BaseTestCases:
class DrawShapes(unittest.TestCase):
width = 600
height = 300
kwargs = {'fillcolor': 'yellow', 'outlinecolor': 'black'}
output_prefix = 'render_projections'
crs = None
def create_map(self):
self.map = pg.renderer.Map(self.width,
self.height,
background='gray',
import pythongis as pg
poly = pg.VectorData("data/ne_10m_admin_0_countries.shp")
def diagonal_left(width, height):
import pyagg
c = pyagg.Canvas(width, height)
for frac in range(0, 100 + 1, 10):
frac = frac / 100.0
x = width * frac
y = height * frac
c.draw_line([(x, 0), (0, y)], fillcolor='black', outlinecolor=None)
c.draw_line([(x, height), (width, y)],
fillcolor='black',
outlinecolor=None)
return c
diagonal_left(1000, 1000).view()
dsfsd
mapp = pg.renderer.Map(1000, 500, background=None) #,background=(255,0,0))
lyr = mapp.add_layer(poly, fillcolor=filleffect)
mapp.view()
def get_table(filepath, page):
import os
#os.chdir(os.path.abspath(''))
#print os.getcwd()
os.system(
'''cd "{wdir}" & "xpdf-tools-win-4.00\\bin64\pdftopng.exe" -f {page} -l {page} -r 300 -gray "{filepath}" "output\page{pagestr}"'''
.format(filepath=filepath,
page=page,
pagestr=str(page).zfill(6),
wdir=os.path.abspath('')))
im = PIL.Image.open('output/page{pagestr}-{pagestr}.png'.format(
pagestr=str(page).zfill(6))).convert('1').point(lambda v: v == 0)
# grow to connect tiny gaps
global oim
oim = im
#im = PIL.ImageMorph.MorphOp(op_name='dilation4').apply(im)[1]
#im = PIL.ImageMorph.MorphOp(op_name='dilation4').apply(im)[1]
## im = PIL.ImageMorph.MorphOp(patterns=['1(... ... ...)->0',
## '4(000 111 000)->1',
## '4(111 111 000)->1',
## '4(000 110 010)->1',
## '4(000 110 110)->1',]).apply(im)[1]
# collect lines
hlines, vlines = find_lines(im, 50, mingap=2, maxthick=None)
# extend lines
## for i,hl in enumerate(hlines):
## (x1,y1),(x2,y2) = hl
## extend = 10 #int((x2-x1)*0.05) # 5 percent
## hlines[i] = (x1-extend,y1),(x2+extend,y2)
## for i,vl in enumerate(vlines):
## (x1,y1),(x2,y2) = vl
## extend = 10 #int((y2-y1)*0.05) # 5 percent
## vlines[i] = (x1,y1-extend),(x2,y2+extend)
# draw lines
lineimg = PIL.Image.new('1', im.size, 0)
draw = PIL.ImageDraw.Draw(lineimg)
for s, e in hlines:
draw.line(s + e, fill=255)
for s, e in vlines:
draw.line(s + e, fill=255)
if DEBUG: lineimg.show()
lineimg.save('output/page%s-lines.png' % str(page).zfill(6))
## # ALT1: thin lines to single pixel
## lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## #lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## #lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## #lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## #lineimg.show()
## for _ in range(30):
## print _
## lineimg = PIL.ImageMorph.MorphOp(patterns=['4(000 .1. 111)->0','4(.00 110 .1.)->0',
## '1(111 111 000)->1','1(110 110 110)->1',
## ]).apply(lineimg)[1]
## lineimg.show()
# ALT2: collect new lines, only 1 per maxthick pixels
# TODO: compare all neighbouring lines and get longest one
## lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## lineimg = PIL.ImageMorph.MorphOp(op_name='dilation8').apply(lineimg)[1]
## lineimg.show()
hlines, vlines = find_lines(lineimg,
150,
mingap=50,
maxthick=None,
move=True) #30, offset=0, move=1)
# group lines
ghlines, gvlines = group_lines(hlines, vlines)
# draw grouped lines
lineimg = PIL.Image.new('1', im.size, 0)
draw = PIL.ImageDraw.Draw(lineimg)
for s, e in ghlines:
draw.line(s + e, fill=255)
for s, e in gvlines:
draw.line(s + e, fill=255)
if DEBUG: lineimg.show()
# find corners
# ALT1: from image corners
## cornimg = PIL.ImageMorph.MorphOp(patterns=['1(... ... ...)->0',
## '4(010 110 000)->1', #right corner
## '4(000 110 000)->1',#right tip
## '4(010 111 000)->1',#3way junction
## '1(010 111 010)->1',#4way junction
## ]).apply(lineimg)[1]
##
## PIL.ImageMorph.MorphOp(op_name="dilation8").apply(cornimg)[1].show()
##
## # thin each corner to one pixel
## corners = []
## pixels = cornimg.load()
## for y in range(cornimg.height):
## for x in range(cornimg.width):
## val = pixels[x,y]
## if val:
## similar = any((abs(x-cx)<10 and abs(y-cy)<10
## for cx,cy in corners))
## # dont add if a similar one has already been added, < 10 pixels
## if similar:
## continue
## corners.append((x,y))
##
## newcornimg = PIL.Image.new('1', cornimg.size, 0)
## pixels = newcornimg.load()
## for x,y in corners:
## pixels[x,y] = 255
##
## PIL.ImageMorph.MorphOp(op_name="dilation8").apply(newcornimg)[1].show()
# ALT2: instead, if we already have all straight connected lines, then just get the list of all their geom intersections
isecs = []
for hline in ghlines:
for vline in gvlines:
if hline[0][0] - 10 <= vline[0][0] <= hline[1][0] + 10 and vline[
0][1] - 10 <= hline[0][1] <= vline[1][1] + 10:
x, y = vline[0][0], hline[0][1]
isecs.append((x, y))
corners = isecs
# add corners of really long lines
hlines = [l for l in ghlines if l[1][0] - l[0][0] > 1000]
for l in hlines:
corners.append(l[0])
corners.append(l[1])
# draw isecs
cornimg = PIL.Image.new('1', im.size, 0)
pixels = cornimg.load()
for x, y in corners:
pixels[x, y] = 255
if DEBUG:
PIL.ImageMorph.MorphOp(op_name="dilation8").apply(cornimg)[1].show()
# add lines that contains all isecs
xmin, ymin = min((c[0] for c in corners)), min((c[1] for c in corners))
xmax, ymax = max((c[0] for c in corners)), max((c[1] for c in corners))
ghlines.append([(xmin, ymin), (xmax, ymin)])
ghlines.append([(xmin, ymax), (xmax, ymax)])
gvlines.append([(xmin, ymin), (xmin, ymax)])
gvlines.append([(xmax, ymin), (xmax, ymax)])
# find boxes
boxes = find_boxes(corners, ghlines, gvlines)
# test
## for box in boxes:
## text = get_text(oim, box)
## print box, text
# find all lines that are really long
hlines = [l for l in ghlines if l[1][0] - l[0][0] > 1000]
print hlines
# draw lines
tablineimg = PIL.Image.new('1', im.size, 0)
draw = PIL.ImageDraw.Draw(tablineimg)
for s, e in hlines:
draw.line(s + e, fill=255)
if DEBUG: tablineimg.show()
# two top ones defines the fields region, bottom ones the column/data region
tabfieldtop = hlines[0][0][1]
tabfieldbottom = hlines[1][0][1]
tabdattop = hlines[1][0][1]
tabdatbottom = hlines[2][0][1]
# detect top text
toprows = detect_data(oim, bbox=[0, 0, oim.size[0], tabfieldtop])
# read table title
topmargin = 90 + 60
chapter = get_text(toprows, [(0, 0), (oim.size[0], topmargin)])[0]
titlerows = filter_data(toprows, [0, topmargin, oim.size[0], tabfieldtop])
title = ' '.join((r['text'] for r in titlerows if r['text'].strip()))
print chapter
print title
# detect all text
drows = detect_data(oim, bbox=[0, 0, oim.size[0], tabdatbottom])
# print all fields text for debugging
for line in filter_data(drows,
[0, tabfieldtop, oim.size[0], tabfieldbottom]):
if 'text' in line and line['text']:
print line
# loop through all topleft corners of boxes inside the fields region, left to right, ignoring ca duplicate x corners
fieldboxes = [bbox for bbox in boxes if bbox[1][1] <= tabfieldbottom + 20]
fieldxs = sorted(set((b[0][0] for b in fieldboxes)))
# for each corner, find ca duplicate x corners
fields = []
for x in fieldxs:
print x
g = [bbox for bbox in fieldboxes if bbox[0][0] <= x < bbox[1][0]]
# loop bboxes of those duplicate corners, top to bottom
g = sorted(g, key=lambda b: b[0][1])
print g
# define field name as their concatenated texts
names = []
for bbox in g:
text, notes = get_text(drows, bbox)
if not text:
# try detecting again, on subset
print 'no text in field', bbox[0] + bbox[1]
x1, y1, x2, y2 = bbox[0] + bbox[1]
trimbox = x1 + 5, y1 + 5, x2 - 5, y2 - 5
fielddata = detect_data(oim, trimbox)
if DEBUG: oim.crop(trimbox).show()
try:
text, notes = get_text(fielddata)
print 'second pass found', repr(text)
except Exception as err:
print 'UNKNOWN FAIL', err
names.append(text or '')
print names
name = '|'.join(reversed(names))
# define bbox width as the bottom bbox width
x1, x2 = g[-1][0][0], g[-1][1][0]
# define column bbox as same width, but extending from lower fields lines to lower table line
y1, y2 = tabdattop, tabdatbottom
bbox = [(x1, y1), (x2, y2)]
print bbox
# add as field obj
fields.append((name, bbox))
# for first field, consider its data column bbox
(x1, y1), (x2, y2) = fields[0][1]
croprows = filter_data(drows, [x1, y1, x2, y2])
# loop through column text lines via image_to_data, top to bottom
lines = []
line = dict(text='')
for row in croprows:
# process
if row['word_num'] == '0' and line['text'].strip():
lines.append(line)
line = dict(text='')
if line['text'] and 'text' in row and row['text'].strip():
# update
bbox = row['left'], row[
'top'], row['left'] + row['width'], row['top'] + row['height']
prevbbox = line['bbox']
line['bbox'] = min(bbox[0],
prevbbox[0]), min(bbox[1], prevbbox[1]), max(
bbox[2],
prevbbox[2]), max(bbox[3], prevbbox[3])
line['text'] += ' ' + row['text']
elif not line['text'] and 'text' in row and row['text'].strip():
# new line
bbox = row['left'], row[
'top'], row['left'] + row['width'], row['top'] + row['height']
line = dict(text=row['text'], bbox=bbox)
lines.append(line)
# for each text
rows = []
row = dict()
lines = sorted(lines, key=lambda l: l['bbox'][1])
for i in range(len(lines)):
line = lines[i]
print line
## rowtop = line['bbox'][1]
## try:rowbottom = lines[i+1]['bbox'][1]
## except:rowbottom = line['bbox'][3] # last line
##
## try:crop.crop([0,rowtop,crop.size[0]-1,rowbottom]).show()
## except:pass
# process
if not row:
# first new
row['rowtop'] = line['bbox'][1] - 3
row['text'] = line['text']
# if firstval is indented, consider a subunit
# ...
else:
# continued
row['text'] += '\n' + line['text']
# if ends with dot ".", add row as dict
if line['text'].endswith('.') or line['text'].endswith(u'\u2018') \
or (len(lines) >= i+1 and lines[i+1]['bbox'][1] - line['bbox'][3] > 30): # or if large gap until next line, colon, or is mostly upper caps, add this as header/grouping
row['rowbottom'] = line['bbox'][3]
# loop through all fields and check for text/data within the top/bottom range of the text
vals = []
for fname, fbox in fields:
bbox = [(fbox[0][0], row['rowtop']),
(fbox[1][0], row['rowbottom'])]
text, notes = get_text(drows, bbox)
vals.append(text)
row['vals'] = vals
# add and reset
print row
rows.append(row)
row = dict()
## for row in rows:
## # testview
## oim.crop([0,row['rowtop'],oim.size[0]-1,row['rowbottom']]).show()
import pythongis as pg
d = pg.VectorData(fields=[f[0] for f in fields]) #.split('|')[-1]
for r in rows:
d.add_feature(row=r['vals'])
return chapter, title, d
import pythongis as pg
import gc
countries = pg.VectorData(r"C:\Users\kimok\Desktop\ch est\data\cshapes.shp")
countries = countries.select(
lambda f: f["GWCODE"] != -1 and f["GWEYEAR"] == 2016)
print countries
# mapit
#rast = pg.RasterData(r'C:\Users\kimok\Downloads\F182013.v4c_web.stable_lights.avg_vis.tif')
rast = pg.RasterData(
r'C:\Users\kimok\Downloads\SVDNB_npp_20170701-20170731_75N060W_vcmcfg_v10_c201708061230.avg_rade9.tif'
)
print rast
for iso in ['NGA', 'COD', 'YEM']:
print iso
c = countries.select(lambda f: f["ISO1AL3"] == iso)
clip = rast.manage.clip(c, bbox=c.bbox)
print clip
clip.view(cutoff=(0.1, 99.9))
print 'finished!'
def test_example(self):
import pythongis as pg
DEBUG = False
PROFILE = True
print 'loading'
d = pg.VectorData(
r"C:\Users\kimok\Downloads\ne_10m_admin_1_states_provinces (1)\ne_10m_admin_1_states_provinces.shp"
)
#d = pg.VectorData(r"C:\Users\kimok\Desktop\BIGDATA\gazetteer data\raw\global_settlement_points_v1.01.shp", encoding='latin')
d = d.select(lambda f: f.id < 1000)
items = [(f.id, f.bbox)
for f in d] # items = [(i+1, f.bbox) for i,f in enumerate(d)]
print len(items)
#####################
# build
print 'building'
spindex = QuadTree(xmin=-180, ymin=-90, xmax=180, ymax=90)
spindex.save()
if PROFILE:
import cProfile
prof = cProfile.Profile()
prof.enable()
spindex.build(items)
if PROFILE:
print prof.print_stats('cumtime')
#fdsdfd
print 'items', Item.objects.all().count()
print 'nodes', Node.objects.all().count()
print 'links', ItemNodeLink.objects.all().count()
# visualize
m = pg.renderer.Map()
# quad structure
print 'explore nodes and items'
quads = pg.VectorData(fields=['nodeid', 'parent', 'depth', 'count'])
for node in spindex.nodes.all():
x1, y1, x2, y2 = node.xmin, node.ymin, node.xmax, node.ymax
box = {
'type': 'Polygon',
'coordinates': [[(x1, y1), (x1, y2), (x2, y2), (x2, y1)]]
}
quads.add_feature(
[node.pk, node.parent, node.depth, node.item_count], box)
items = pg.VectorData(fields=['item'])
for item in Item.objects.all():
x1, y1, x2, y2 = item.xmin, item.ymin, item.xmax, item.ymax
box = {
'type': 'Polygon',
'coordinates': [[(x1, y1), (x1, y2), (x2, y2), (x2, y1)]]
}
items.add_feature([item], box)
print(quads)
m.add_layer(d, fillcolor='red')
m.add_layer(items, fillcolor=None, outlinecolor='blue')
m.add_layer(
quads,
fillcolor=None,
outlinecolor='green',
) #text=lambda f: f['nodeid'], textoptions={'textcolor':'green','textsize':6})
m.render_all()
m.view()
###################
# intersection
print 'intersecting'
testbox = (0, 0, 90, 45)
#testbox = (100,1,120,20)
#testbox = (100,15,105,20)
if PROFILE:
import cProfile
prof = cProfile.Profile()
prof.enable()
matches = spindex.intersect(testbox)
if PROFILE:
print prof.print_stats('cumtime')
#fdsdfds
# visualize
m = pg.renderer.Map()
# result item boxes
boxes = pg.VectorData()
res = pg.VectorData()
for match in matches:
#print match
#oid,itemid,x1,y1,x2,y2,depth,path = match
oid, itemid, x1, y1, x2, y2 = [
getattr(match, k)
for k in 'pk,item_id,xmin,ymin,xmax,ymax'.split(',')
]
f = d[itemid]
res.add_feature([], f.geometry)
box = {
'type': 'Polygon',
'coordinates': [[(x1, y1), (x1, y2), (x2, y2), (x2, y1)]]
}
boxes.add_feature([], box)
m.add_layer(res, fillcolor='red')
m.add_layer(boxes, fillcolor=None, outlinecolor='green')
# result node boxes
## nodematches = spindex.intersect_nodes(testbox)
## nodedata = pg.VectorData(fields=['path','count'])
## for node in nodematches:
## #print match
## count = node[-3]
## x1,y1,x2,y2 = node[1:5] #f.bbox
## path = node[-1]
## box = {'type':'Polygon', 'coordinates':[[(x1,y1),(x1,y2),(x2,y2),(x2,y1)]]}
## nodedata.add_feature([path,count], box)
## m.add_layer(nodedata, fillcolor=None, outlinecolor='blue')
# testbox
testboxdata = pg.VectorData()
x1, y1, x2, y2 = testbox
geoj = {
'type': 'Polygon',
'coordinates': [[(x1, y1), (x1, y2), (x2, y2), (x2, y1)]]
}
testboxdata.add_feature([], geoj)
m.add_layer(testboxdata,
fillcolor=None,
outlinecolor='black',
outlinewidth='3px')
m.render_all()
m.view()
import pythongis as pg
rw = 6
rh = 3
w = 360
h = 180
xscale = w / float(rw)
yscale = h / float(rh)
affine = [xscale, 0, -180, 0, -yscale, 90]
r = pg.RasterData(width=rw, height=rh, mode='float32', affine=affine)
rb = r.add_band()
for y in range(rh):
for x in range(rw):
rb.set(x, y, (x + 1) * (y + 1))
v = pg.VectorData()
bounds = [(-180, -90), (180, -90), (180, 90), (-180, 90)]
bounds.append(bounds[0])
v.add_feature([], {'type': 'Polygon', 'coordinates': [bounds]})
m = pg.renderer.Map(2000, 1000, 'white')
m.add_layer(r)
m.add_layer(v, fillcolor=None, outlinecolor='red')
m.zoom_auto()
m.view()
import pythongis as pg
import pycrs
pg.vector.data.DEFAULT_SPATIAL_INDEX = 'quadtree'
data = pg.VectorData(r"P:\(Temp Backup)\priocountries\priocountries.shp")#"C:\Users\kimok\Desktop\BIGDATA\priocountries\priocountries.shp")
rast = pg.RasterData(r"C:\Users\kimok\Documents\GitHub\AutoMap\tests\testmaps\burkina_pol96_georeferenced.tif")
#testcrs = '+proj=robin +datum=WGS84 +ellps=WGS84 +a=6378137.0 +rf=298.257223563 +pm=0 +lon_0=0 +x_0=0 +y_0=0 +units=m +axis=enu +no_defs'
#testcrs = pycrs.parse.from_sr_code(6980).to_proj4() # space
#testcrs = pycrs.parse.from_sr_code(7619).to_proj4() # goode?
#testcrs = next(pycrs.utils.search('van der grinten'))['proj4']
testcrs = next(pycrs.utils.search('eckert iv'))['proj4']
#### original crs
#data.view()
#### test on-the-fly crs
#data.view(crs=testcrs)
#### raster crs
rast.save('C:/Users/kimok/Desktop/testmap.png')
#rast.view() #testcrs)
#rast.manage.reproject(testcrs, resample='nearest').view()
#rast.manage.reproject(testcrs, resample='bilinear').view()
dsadsads
import pythongis as pg
import math
points = pg.VectorData("data/ne_10m_populated_places_simple.shp",
encoding='latin')
#points.browse()
def radsize(f):
val = f['pop_max']
sz = round(val * 2 / 10000000.0,
2) # weird relative dist error due to small e- nr...
return sz
def areasize(f):
val = f['pop_max']
area = round(val * 2 / 10000000.0, 2)
sz = math.sqrt(area / math.pi)
return sz
# custom
#points.view(fillsize=radsize, fillcolor='yellow')
#points.view(fillsize=areasize, fillcolor='yellow')
# builtin
#points.view(fillsize=dict(breaks='proportional', key='pop_max', sizes=[0.1,1]), fillcolor='yellow')
mapp = pg.renderer.Map()
mapp.add_layer(points,
fillsize=dict(breaks='proportional',
node.update(itemid, parent=node.nodeid)
def add_item(self, item, bbox):
xmin, ymin, xmax, ymax = bbox
parent = self.nodeid
self._index.cur.execute('INSERT INTO items VALUES (?, ?, ?, ?, ?, ?)',
(parent, item, xmin, ymin, xmax, ymax))
if __name__ == '__main__':
import pythongis as pg
DEBUG = False
print 'loading'
d = pg.VectorData(
r"C:\Users\kimok\Downloads\ne_10m_admin_1_states_provinces (1)\ne_10m_admin_1_states_provinces.shp"
)
items = [(i, f.bbox) for i, f in enumerate(d)]
print 'building'
spindex = QuadTree(-180, -90, 180, 90)
spindex.build(items)
print 'intersecting'
matches = spindex.intersect((1, 1, 20, 20))
## print 'viewing'
## boxes = pg.VectorData(fields=['path','parent'])
## res = pg.VectorData()
## for match in matches:
## #print match
import pythongis as pg
data = pg.VectorData("data/ne_10m_admin_0_countries.shp")
print data
# vector
proj = data.manage.reproject("+proj=latlong +datum=WGS84",
"+proj=robin +lon_0=0 +x_0=0 +y_0=0 +ellps=WGS84 +datum=WGS84 +units=m +no_defs")
print proj
proj.view()
grid = pg.VectorData()
for x in range(-180,180+1,20):
line = [[x,y] for y in range(90,-90-1,-10)]
geoj = dict(type='LineString', coordinates=line)
grid.add_feature(geometry=geoj)
grid = grid.manage.reproject("+proj=latlong +datum=WGS84",
"+proj=robin +lon_0=0 +x_0=0 +y_0=0 +ellps=WGS84 +datum=WGS84 +units=m +no_defs")
grid.view(outlinecolor=None)
# raster
rast = pg.raster.manager.rasterize(data, bbox=[-170,80,170,-80], width=720, height=360)
#rast.view()
rast.crs = "+proj=latlong +datum=WGS84"
import pythongis as pg
data = pg.VectorData(r"C:\Users\kimo\Downloads\cshapes_0.6\cshapes.shp")
pt = pg.VectorData()
pt.add_feature([], dict(type="Point", coordinates=(10, 30)))
pt.add_feature([], dict(type="Point", coordinates=(11, 31)))
snap = pt.manage.snap(data, 0.5)
for f in snap:
print f.__geo_interface__
#snap.view()
mapp = pg.renderer.Map()
mapp.add_layer(data)
mapp.add_layer(pt.manage.buffer(lambda f: 0.5))
mapp.add_layer(pt)
mapp.add_layer(snap)
mapp.view()
# test speed (very slow...???)
from random import randrange
pt = pg.VectorData()
for _ in range(4):
print _
pt.add_feature([],
dict(type="Point",
coordinates=(randrange(180), randrange(90))))
snap = pt.manage.snap(data, 1)
import pythongis as pg
c = pg.VectorData(r"C:\Users\kimok\Desktop\ch est\data\cshapes.shp",
select=lambda f: f['GWEYEAR']==2016)
# normal colors
mapp = pg.renderer.Map(4000,2000,title='World Map')
mapp.add_layer(c, fillcolor=dict(breaks='natural',
key='GWCODE',
colors=['green','red']))
mapp.add_legend(legendoptions=dict(title='Legend'))
mapp.save("normal color breaks.png")
# normal sizes
mapp = pg.renderer.Map(2000,1000,title='World Map')
mapp.add_layer(c.convert.to_points(), fillsize=dict(breaks='natural',
key='GWCODE',
sizes=[0.2,2.0]))
mapp.add_legend()
mapp.save("normal size breaks.png")
# normal multiple breaks
mapp = pg.renderer.Map(2000,1000,title='World Map')
mapp.add_layer(c.convert.to_points(),
fillsize=dict(breaks='natural',
key='GWCODE',
sizes=[0.2,2.0]),
fillcolor=dict(breaks='natural',
key='GWCODE',
colors=['green','red']))
import pythongis as pg
from time import time
if 0:
# conflictsite
confs = pg.VectorData("C:/Users/kimo/Downloads/Conflict Site Dataset 2.0/Conflict Site 4-2006.xls",
xfield="Longitude", yfield="Latitude")
print confs, confs.fields
confs = confs.manage.buffer(lambda f: f["Radius"], geodetic=True, resolution=20)
#confs.view()
countries = pg.VectorData("C:/Users/kimo/Downloads/cshapes_0.6/cshapes.shp")
countries = countries.select(lambda f: f["GWCODE"]!=-1)
# clip to countries (hacky for now, what happened to the intersection method?)
from shapely.ops import cascaded_union
for f in confs:
print ["vector intersecting",f,f.id,len(confs)]
fg = f.get_shapely()
intsecs = [fg.intersection(gw.get_shapely())
for gw in countries
if gw["GWSYEAR"] <= f["Year"] <= gw["GWEYEAR"] \
and int(float(gw["GWCODE"])) in [int(float(cod)) for cod in str(f["Conflict site"]).split(",")]]
if len(intsecs) == 1:
newg = intsecs[0].__geo_interface__
elif len(intsecs) > 1:
newg = cascaded_union(intsecs).__geo_interface__
else:
continue # no gw match
import pythongis as pg
W = 2000
H = 1000
SPACING = 330
RIGHTSPIN = 500
DOWNTILT = 700
c = pg.VectorData("ch est/data/cshapes.shp")
import PIL, PIL.Image
def tilt(img, oldplane, newplane):
pb,pa = oldplane,newplane
grid = []
for p1,p2 in zip(pa, pb):
grid.append([p1[0], p1[1], 1, 0, 0, 0, -p2[0]*p1[0], -p2[0]*p1[1]])
grid.append([0, 0, 0, p1[0], p1[1], 1, -p2[1]*p1[0], -p2[1]*p1[1]])
import advmatrix as mt
A = mt.Matrix(grid)
B = mt.Vec([xory for xy in pb for xory in xy])
AT = A.tr()
ATA = AT.mmul(A)
gridinv = ATA.inverse()
invAT = gridinv.mmul(AT)
res = invAT.mmul(B)
transcoeff = res.flatten()
#then calculate new coords, thanks to http://math.stackexchange.com/questions/413860/is-perspective-transform-affine-if-it-is-why-its-impossible-to-perspective-a"
new = img.transform(img.size, PIL.Image.PERSPECTIVE, transcoeff, PIL.Image.BILINEAR)
return new
or (len(lines) >= i+1 and lines[i+1]['bbox'][1] - line['bbox'][3] > 30): # or if large gap until next line, colon, or is mostly upper caps, add this as header/grouping
row['rowbottom'] = line['bbox'][3]
# loop through all fields and check for text/data within the top/bottom range of the text
vals = []
for fname,fbox in fields:
bbox = [(fbox[0][0],row['rowtop']), (fbox[1][0], row['rowbottom'])]
text,notes = get_text(drows, bbox)
vals.append(text)
row['vals'] = vals
# add and reset
print row
rows.append(row)
row = dict()
## for row in rows:
## # testview
## oim.crop([0,row['rowtop'],oim.size[0]-1,row['rowbottom']]).show()
import pythongis as pg
d=pg.VectorData(fields=[f[0].split('|')[-1] for f in fields])
for r in rows:
d.add_feature(row=r['vals'])
d.browse()
import pythongis as pg
from time import time
poly = pg.VectorData("data/ne_10m_admin_1_states_provinces.shp",
encoding="latin")
points = pg.VectorData("data/ne_10m_populated_places_simple.shp",
encoding="latin")
print points
##t=time()
##join = points.manage.spatial_join(points, "distance",
## radius=10, n=3, # 3 nearest within 10k
## key=lambda f1,f2: f1.geometry != f2.geometry) # not self
##print time()-t, join
t = time()
join = poly.manage.spatial_join(
points, "intersects",
clip=lambda f1, f2: f2.geometry) # poly contains points
print "fast join country to each point", time() - t, join
# HEAVY ONES
# Slow point-poly
##t=time()
##join = points.manage.spatial_join(poly, "intersects") # point in polys
##print time()-t, join
t = time()
join = points.manage.spatial_join(poly, "distance", radius=60, n=1)
import pythongis as pg
conf = pg.VectorData('rosling/data/ged171_excelcsv.csv',
encoding='latin',
xfield='longitude',
yfield='latitude',
select=lambda f: f['year'] == 2016)
conf.compute('dum', 1)
width = 720
height = 360
agg = pg.raster.manager.rasterize(conf,
valuekey=lambda f: f['dum'],
stat='sum',
width=width,
height=height,
bbox=[-180, 90, 180, -90])
mapp = agg.render(cutoff=(0, 100))
mapp.add_legend()
mapp.save('agg.png')
smooth = pg.raster.analyzer.smooth(
conf,
rasterdef=dict(mode='float32',
width=width,
height=height,
bbox=[-180, 90, 180, -90]),
algorithm='radial',
radius=3,
)
import pythongis as pg
cshapes = pg.VectorData("cshapes.shp", select=lambda f: f["GWCODE"] != -1)
cshapes = pg.VectorData("selfisec.geojson")
print cshapes
mapp = pg.renderer.Map(width=1000)
mapp.add_layer(cshapes, fillcolor="blue")
selfint = cshapes.intersections()
print selfint
selfint.view(1000,
1000,
flipy=1,
fillcolor=pg.renderer.Color("red", opacity=155))
selfint = selfint.duplicates(fieldmapping=[("count", lambda f: 1, "count")])
print selfint
mapp.add_layer(selfint, fillcolor=pg.renderer.Color("red", opacity=155))
mapp.render_all()
mapp.view()
import pythongis as pg
data = pg.VectorData('data/ne_10m_admin_0_countries.shp')
data.view(fillcolor='green',
fillopacity={
'key': 'POP_EST',
'opacities': [0.3, 1.0]
})
##data.view(fillcolor={'key':'POP_EST', 'colors':['beige','red']},
## fillopacity={'key':lambda f: f.get_shapely().area, 'opacities':[1.0,0.3]},
## )
import unittest
import pythongis as pg
# data
pointdata = pg.VectorData('data/ne_10m_populated_places_simple.shp',
encoding='latin')
# base class
class BaseTestCases:
class TestInitSpindex(unittest.TestCase):
def create_spindex(self, **kwargs):
self.data.create_spatial_index(**kwargs)
def test_default_rtree(self):
# on local pc should fail and fallback to quadtree
pg.vector.data.DEFAULT_SPATIAL_INDEX = 'rtree'
self.create_spindex()
self.assertTrue(hasattr(self.data, 'spindex'))
self.assertTrue(not isinstance(self.data.spindex, pg.vector.spindex.Rtree))
def test_default_quadtree(self):
pg.vector.data.DEFAULT_SPATIAL_INDEX = 'quadtree'
self.create_spindex()
self.assertTrue(hasattr(self.data, 'spindex'))
self.assertTrue(isinstance(self.data.spindex, pg.vector.spindex.QuadTree))
