fixed sort_by to respect updated index (upon subsequent adding of new data)

This commit is contained in:
Philip Eisenlohr 2019-12-05 12:00:59 -05:00
parent 3caf2c1296
commit 2648a67dcd
2 changed files with 44 additions and 35 deletions

View File

@ -1,4 +1,3 @@
import random
import re import re
import pandas as pd import pandas as pd
@ -24,15 +23,24 @@ class Table():
self.comments = [] if comments is None else [c for c in comments] self.comments = [] if comments is None else [c for c in comments]
self.data = pd.DataFrame(data=data) self.data = pd.DataFrame(data=data)
self.shapes = shapes self.shapes = shapes
self.__label_condensed()
def __label_flat(self):
"""Label data individually, e.g. v v v ==> 1_v 2_v 3_v."""
labels = []
for label,shape in self.shapes.items():
size = np.prod(shape)
labels += ['{}{}'.format('' if size == 1 else '{}_'.format(i+1),label) for i in range(size)]
self.data.columns = labels
def __label_condensed(self):
"""Label data condensed, e.g. 1_v 2_v 3_v ==> v v v."""
labels = [] labels = []
for label,shape in self.shapes.items(): for label,shape in self.shapes.items():
labels += [label] * np.prod(shape) labels += [label] * np.prod(shape)
self.data.columns = labels
if len(labels) != self.data.shape[1]:
raise IndexError('Mismatch between shapes and data')
self.data.rename(columns=dict(zip(range(len(labels)),labels)),inplace=True)
def __add_comment(self,label,shape,info): def __add_comment(self,label,shape,info):
@ -87,7 +95,7 @@ class Table():
return Table(data,shapes,comments) return Table(data,shapes,comments)
@property
def labels(self): def labels(self):
"""Return the labels of all columns.""" """Return the labels of all columns."""
return list(self.shapes.keys()) return list(self.shapes.keys())
@ -105,9 +113,11 @@ class Table():
""" """
if re.match(r'[0-9]*?_',label): if re.match(r'[0-9]*?_',label):
idx,key = label.split('_',1) idx,key = label.split('_',1)
return self.data[key].to_numpy()[:,int(idx)-1].reshape((-1,1)) data = self.data[key].to_numpy()[:,int(idx)-1].reshape((-1,1))
else: else:
return self.data[label].to_numpy().reshape((-1,)+self.shapes[label]) # better return shape (N) instead of (N,1), i.e. no reshaping? data = self.data[label].to_numpy().reshape((-1,)+self.shapes[label])
return data.astype(type(data.flatten()[0]))
def set(self,label,data,info=None): def set(self,label,data,info=None):
@ -152,10 +162,11 @@ class Table():
self.shapes[label] = data.shape[1:] if len(data.shape) > 1 else (1,) self.shapes[label] = data.shape[1:] if len(data.shape) > 1 else (1,)
size = np.prod(data.shape[1:],dtype=int) size = np.prod(data.shape[1:],dtype=int)
self.data = pd.concat([self.data, new = pd.DataFrame(data=data.reshape(-1,size),
pd.DataFrame(data=data.reshape(-1,size), columns=[label]*size,
columns=[label]*size)], )
axis=1) new.index = self.data.index
self.data = pd.concat([self.data,new],axis=1)
def delete(self,label): def delete(self,label):
@ -201,24 +212,15 @@ class Table():
Parameters Parameters
---------- ----------
label : list of str or str label : str or list
Column labels. Column labels.
ascending : bool, optional ascending : bool or list, optional
Set sort order. Set sort order.
""" """
_temp = [] self.__label_flat()
_labels = [] self.data.sort_values(labels,axis=0,inplace=True,ascending=ascending)
for label in labels if isinstance(labels,list) else [labels]: self.__label_condensed()
if re.match(r'[0-9]*?_',label):
_temp.append(str(random.getrandbits(128)))
self.add(_temp[-1],self.get(label))
_labels.append(_temp[-1])
else:
_labels.append(label)
self.data.sort_values(_labels,axis=0,inplace=True,ascending=ascending)
for t in _temp: self.delete(t)
self.comments.append('sorted by [{}]'.format(', '.join(labels))) self.comments.append('sorted by [{}]'.format(', '.join(labels)))

View File

@ -9,7 +9,7 @@ from damask import Table
@pytest.fixture @pytest.fixture
def default(): def default():
"""Simple Table.""" """Simple Table."""
x = np.ones((5,13)) x = np.ones((5,13),dtype=float)
return Table(x,{'F':(3,3),'v':(3,),'s':(1,)},['test data','contains only ones']) return Table(x,{'F':(3,3),'v':(3,),'s':(1,)},['test data','contains only ones'])
@pytest.fixture @pytest.fixture
@ -58,7 +58,7 @@ class TestTable:
assert np.allclose(d,0.0) and d.shape[1:] == (3,3) assert np.allclose(d,0.0) and d.shape[1:] == (3,3)
def test_labels(self,default): def test_labels(self,default):
assert default.labels() == ['F','v','s'] assert default.labels == ['F','v','s']
def test_add(self,default): def test_add(self,default):
d = np.random.random((5,9)) d = np.random.random((5,9))
@ -82,9 +82,9 @@ class TestTable:
default.get('v') default.get('v')
def test_invalid_initialization(self,default): def test_invalid_initialization(self):
x = default.get('v') x = np.random.random((5,10))
with pytest.raises(IndexError): with pytest.raises(ValueError):
Table(x,{'F':(3,3)}) Table(x,{'F':(3,3)})
def test_invalid_set(self,default): def test_invalid_set(self,default):
@ -115,7 +115,14 @@ class TestTable:
def test_sort_revert(self): def test_sort_revert(self):
x = np.random.random((5,12)) x = np.random.random((5,12))
t = Table(x,{'F':(3,3),'v':(3,)},['random test data']) t = Table(x,{'F':(3,3),'v':(3,)},['random test data'])
t.sort_by('4_F',False) t.sort_by('4_F',ascending=False)
sort = t.get('4_F') sort = t.get('4_F')
assert np.all(np.sort(sort,0)==sort[::-1,:]) assert np.all(np.sort(sort,0)==sort[::-1,:])
def test_sort(self):
t = Table(np.array([[0,1,],[2,1,]]),
{'v':(2,)},
['test data'])
t.add('s',np.array(['b','a']))
t.sort_by('s')
assert np.all(t.get('1_v') == np.array([2,0]).reshape((2,1)))