Skip to content

Commit

Permalink
BUG: return np.nan if feature not in spatial_weights (#131)
Browse files Browse the repository at this point in the history 
* check if uID is in spatial_weights

* fix distribution.py

* fix

* tests

* warning formatting
  • Loading branch information
@martinfleis
martinfleis authored Dec 14, 2019
1 parent 14cd876 commit ba46d4a
Show file tree
Hide file tree
Showing 9 changed files with 254 additions and 125 deletions.
73 changes: 41 additions & 32 deletions momepy/dimension.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -399,26 +399,29 @@ def __init__(self, gdf, values, spatial_weights, unique_id, rng=None, mode="mean

results_list = []
for index, row in tqdm(data.iterrows(), total=data.shape[0]):
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]
if index in spatial_weights.neighbors.keys():
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]

values_list = data.loc[neighbours][values]
values_list = data.loc[neighbours][values]

if rng:
from momepy import limit_range
if rng:
from momepy import limit_range

values_list = limit_range(values_list, rng=rng)
if mode == "mean":
results_list.append(np.mean(values_list))
elif mode == "median":
results_list.append(np.median(values_list))
elif mode == "mode":
results_list.append(sp.stats.mode(values_list)[0][0])
values_list = limit_range(values_list, rng=rng)
if mode == "mean":
results_list.append(np.mean(values_list))
elif mode == "median":
results_list.append(np.median(values_list))
elif mode == "mode":
results_list.append(sp.stats.mode(values_list)[0][0])
else:
raise ValueError("{} is not supported as mode.".format(mode))
else:
raise ValueError("{} is not supported as mode.".format(mode))
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand Down Expand Up @@ -764,16 +767,19 @@ def __init__(self, gdf, values, spatial_weights, unique_id, areas=None):

results_list = []
for index, row in tqdm(data.iterrows(), total=data.shape[0]):
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]
if index in spatial_weights.neighbors.keys():
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]

subset = data.loc[neighbours]
results_list.append(
(sum(subset[values] * subset[areas])) / (sum(subset[areas]))
)
subset = data.loc[neighbours]
results_list.append(
(sum(subset[values] * subset[areas])) / (sum(subset[areas]))
)
else:
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand Down Expand Up @@ -824,14 +830,17 @@ def __init__(self, gdf, spatial_weights, unique_id):

results_list = []
for index, row in tqdm(data.iterrows(), total=data.shape[0]):
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]
if index in spatial_weights.neighbors.keys():
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]

areas = data.loc[neighbours].geometry.area
results_list.append(sum(areas))
areas = data.loc[neighbours].geometry.area
results_list.append(sum(areas))
else:
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand Down
87 changes: 52 additions & 35 deletions momepy/distribution.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -485,15 +485,17 @@ def __init__(self, gdf, spatial_weights, unique_id, orientations):

# iterating over rows one by one
for index, row in tqdm(data.iterrows(), total=data.shape[0]):

neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
orientation = data.loc[neighbours][orientations]
deviations = abs(orientation - row[orientations])

results_list.append(statistics.mean(deviations))
if index in spatial_weights.neighbors.keys():
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
orientation = data.loc[neighbours][orientations]
deviations = abs(orientation - row[orientations])

results_list.append(statistics.mean(deviations))
else:
results_list.append(np.nan)
else:
results_list.append(0)
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand All @@ -502,6 +504,8 @@ class NeighborDistance:
"""
Calculate the mean distance to adjacent buildings (based on spatial_weights)
If no neighbours are found, return np.nan.
.. math::
\\frac{1}{n}\\sum_{i=1}^n dist_i=\\frac{dist_1+dist_2+\\cdots+dist_n}{n}
Expand All @@ -510,7 +514,7 @@ class NeighborDistance:
gdf : GeoDataFrame
GeoDataFrame containing objects to analyse
spatial_weights : libpysal.weights
spatial weights matrix
spatial weights matrix based on unique_id
unique_id : str
name of the column with unique id used as spatial_weights index.
Expand Down Expand Up @@ -549,14 +553,17 @@ def __init__(self, gdf, spatial_weights, unique_id):

# iterating over rows one by one
for index, row in tqdm(data.iterrows(), total=data.shape[0]):
neighbours = spatial_weights.neighbors[index]
building_neighbours = data.loc[neighbours]
if len(building_neighbours) > 0:
results_list.append(
np.mean(building_neighbours.geometry.distance(row["geometry"]))
)
if index in spatial_weights.neighbors.keys():
neighbours = spatial_weights.neighbors[index]
building_neighbours = data.loc[neighbours]
if len(building_neighbours) > 0:
results_list.append(
np.mean(building_neighbours.geometry.distance(row["geometry"]))
)
else:
results_list.append(np.nan)
else:
results_list.append(0)
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand Down Expand Up @@ -667,13 +674,16 @@ def __init__(
# id to match spatial weights
uid = row[unique_id]
# define neighbours based on weights matrix defining analysis area
neighbours = spatial_weights_higher.neighbors[uid].copy()
neighbours.append(uid)
if neighbours:
selection = adj_list[adj_list.focal.isin(neighbours)][
adj_list.neighbor.isin(neighbours)
]
results_list.append(np.nanmean(selection.distance))
if uid in spatial_weights_higher.neighbors.keys():
neighbours = spatial_weights_higher.neighbors[uid].copy()
neighbours.append(uid)
if neighbours:
selection = adj_list[adj_list.focal.isin(neighbours)][
adj_list.neighbor.isin(neighbours)
]
results_list.append(np.nanmean(selection.distance))
else:
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand Down Expand Up @@ -855,16 +865,19 @@ def __init__(self, gdf, spatial_weights_higher, unique_id, spatial_weights=None)

print("Calculating adjacency...")
for index, row in tqdm(gdf.iterrows(), total=gdf.shape[0]):
neighbours = spatial_weights_higher.neighbors[row[unique_id]].copy()
if neighbours:
neighbours.append(row[unique_id])
if row[unique_id] in spatial_weights_higher.neighbors.keys():
neighbours = spatial_weights_higher.neighbors[row[unique_id]].copy()
if neighbours:
neighbours.append(row[unique_id])

patches_sub = [patches[x] for x in neighbours]
patches_nr = len(set(patches_sub))
patches_sub = [patches[x] for x in neighbours]
patches_nr = len(set(patches_sub))

results_list.append(patches_nr / len(neighbours))
results_list.append(patches_nr / len(neighbours))
else:
results_list.append(np.nan)
else:
results_list.append(0)
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand Down Expand Up @@ -928,11 +941,15 @@ def __init__(self, gdf, spatial_weights, unique_id, weighted=False):

neighbours = []
for index, row in tqdm(gdf.iterrows(), total=gdf.shape[0]):
if weighted is True:
neighbours.append(
spatial_weights.cardinalities[row[unique_id]] / row.geometry.length
)
if row[unique_id] in spatial_weights.neighbors.keys():
if weighted is True:
neighbours.append(
spatial_weights.cardinalities[row[unique_id]]
/ row.geometry.length
)
else:
neighbours.append(spatial_weights.cardinalities[row[unique_id]])
else:
neighbours.append(spatial_weights.cardinalities[row[unique_id]])
neighbours.append(np.nan)

self.series = pd.Series(neighbours, index=gdf.index)
82 changes: 47 additions & 35 deletions momepy/diversity.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -86,14 +86,17 @@ def __init__(self, gdf, values, spatial_weights, unique_id, rng=(0, 100), **kwar

results_list = []
for index, row in tqdm(data.iterrows(), total=data.shape[0]):
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
if index in spatial_weights.neighbors.keys():
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]

values_list = data.loc[neighbours][values]
results_list.append(sp.stats.iqr(values_list, rng=rng, **kwargs))
else:
neighbours = [index]

values_list = data.loc[neighbours][values]
results_list.append(sp.stats.iqr(values_list, rng=rng, **kwargs))
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand Down Expand Up @@ -168,19 +171,22 @@ def __init__(self, gdf, values, spatial_weights, unique_id, rng=None):

results_list = []
for index, row in tqdm(data.iterrows(), total=data.shape[0]):
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]
if index in spatial_weights.neighbors.keys():
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]

values_list = data.loc[neighbours][values]
values_list = data.loc[neighbours][values]

if rng:
from momepy import limit_range
if rng:
from momepy import limit_range

values_list = limit_range(values_list, rng=rng)
results_list.append(Theil(values_list).T)
values_list = limit_range(values_list, rng=rng)
results_list.append(Theil(values_list).T)
else:
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand Down Expand Up @@ -292,16 +298,19 @@ def __init__(
data = data.set_index(unique_id)
results_list = []
for index, row in tqdm(data.iterrows(), total=data.shape[0]):
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
if index in spatial_weights.neighbors.keys():
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
else:
neighbours = [index]
values_list = data.loc[neighbours][values]

sample_bins = classifiers.UserDefined(values_list, self.bins)
counts = dict(zip(self.bins, sample_bins.counts))
results_list.append(self._simpson_di(counts))
else:
neighbours = [index]
values_list = data.loc[neighbours][values]

sample_bins = classifiers.UserDefined(values_list, self.bins)
counts = dict(zip(self.bins, sample_bins.counts))
results_list.append(self._simpson_di(counts))
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)

Expand Down Expand Up @@ -399,18 +408,21 @@ def __init__(self, gdf, values, spatial_weights, unique_id, rng=None):

results_list = []
for index, row in tqdm(data.iterrows(), total=data.shape[0]):
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)
if index in spatial_weights.neighbors.keys():
neighbours = spatial_weights.neighbors[index].copy()
if neighbours:
neighbours.append(index)

values_list = data.loc[neighbours][values].values
values_list = data.loc[neighbours][values].values

if rng:
from momepy import limit_range
if rng:
from momepy import limit_range

values_list = np.array(limit_range(values_list, rng=rng))
results_list.append(Gini(values_list).g)
values_list = np.array(limit_range(values_list, rng=rng))
results_list.append(Gini(values_list).g)
else:
results_list.append(0)
else:
results_list.append(0)
results_list.append(np.nan)

self.series = pd.Series(results_list, index=gdf.index)
Loading

0 comments on commit ba46d4a

Please sign in to comment.