LocateGoodSectionsNode actually appears to work... needs more testing...

nilmtk · May 1, 2014 · 1bdd079 · 1bdd079
1 parent 1250f47
commit 1bdd079
Show file tree

Hide file tree

Showing 5 changed files with 26 additions and 15 deletions.
diff --git a/data/energy_complex.h5 b/data/energy_complex.h5
diff --git a/nilmtk/pipeline/__init__.py b/nilmtk/pipeline/__init__.py
@@ -2,4 +2,4 @@
 from nilmtk.pipeline.pipeline import Pipeline
 from nilmtk.pipeline.energyresults import EnergyResults
 from nilmtk.pipeline.energynode import EnergyNode
-from nilmtk.pipeline.locategapsnode import LocateGapsNode
+from nilmtk.pipeline.locategoodsectionsnode import LocateGoodSectionsNode
diff --git a/nilmtk/pipeline/locategapsnode.py → nilmtk/pipeline/locategoodsectionsnode.py b/nilmtk/pipeline/locategapsnode.py → nilmtk/pipeline/locategoodsectionsnode.py
@@ -14,9 +14,9 @@ def reframe_index(index, window_start=None, window_end=None):
 
     window_start, window_end : pd.Timestamp
         The start and end of the window of interest. If this window
-        is larger than the duration of `data` then a single zero will be
+        is larger than the duration of `data` then a single timestamp will be
         inserted at `window_start` or `window_end` as necessary. If this window
-        is shorter than the duration of `data` data will be cropped.
+        is shorter than the duration of `data` then `data` will be cropped.
 
     Returns
     -------
@@ -48,14 +48,13 @@ def reframe_index(index, window_start=None, window_end=None):
 
     return index
 
-class LocateGapsNode(Node):
+class LocateGoodSectionsNode(Node):
 
-    requirements = {'device': {
-        'max_sample_period': 'ANY VALUE'}}
-    postconditions =  {'preprocessing': {'gaps_located':True}}
+    requirements = {'device': {'max_sample_period': 'ANY VALUE'}}
+    postconditions =  {'preprocessing': {'good_sections_located':True}}
 
-    def __init__(self, name='locate_gaps'):
-        super(LocateGapsNode, self).__init__(name)
+    def __init__(self, name='locate_good_sections'):
+        super(LocateGoodSectionsNode, self).__init__(name)
 
     def process(self, df, metadata=None):
         assert metadata is not None
@@ -66,8 +65,19 @@ def process(self, df, metadata=None):
         index = reframe_index(index, df.timeframe.start, df.timeframe.end)
         timedeltas_sec = timedelta64_to_secs(np.diff(index.values))
         overlong_timedeltas = timedeltas_sec > max_sample_period
-        gap_starts = index[:-1][overlong_timedeltas]
-        gap_ends = index[1:][overlong_timedeltas] 
+        good_sect_starts = index[1:][overlong_timedeltas] 
+        good_sect_starts = good_sect_starts.insert(0, index[0])
+        good_sect_ends = index[:-1][overlong_timedeltas]
+        good_sect_ends = good_sect_ends.insert(len(good_sect_ends), index[-1])
+
+        assert len(good_sect_starts) == len(good_sect_ends)
+
+        mask = []
+        for start, end in zip(good_sect_starts, good_sect_ends):
+            try:
+                mask.append(TimeFrame(start, end))
+            except ValueError:
+                pass # silently ignore good sections of zero length
 
         results = getattr(df, 'results', {})
 #        locate_gaps_results = LocateGapsResults()

diff --git a/nilmtk/pipeline/tests/test_locategaps.py → ...pipeline/tests/test_locategoodsections.py b/nilmtk/pipeline/tests/test_locategaps.py → ...pipeline/tests/test_locategoodsections.py
@@ -1,7 +1,7 @@
 #!/usr/bin/python
 from __future__ import print_function, division
 import unittest
-from nilmtk.pipeline import Pipeline, EnergyNode, LocateGapsNode
+from nilmtk.pipeline import Pipeline, EnergyNode, LocateGoodSectionsNode
 from nilmtk.pipeline.energynode import _energy_per_power_series
 from nilmtk import TimeFrame, EMeter, HDFDataStore, Loader
 from nilmtk.consts import JOULES_PER_KWH
@@ -15,14 +15,14 @@ class TestLocateGaps(unittest.TestCase):
 
     @classmethod
     def setUpClass(cls):
-        filename = join(data_dir(), 'energy.h5')
+        filename = join(data_dir(), 'energy_complex.h5')
         cls.datastore = HDFDataStore(filename)
         cls.loader = Loader(store=cls.datastore, key='/building1/electric/meter1')
 
     def test_pipeline(self):
         meter = EMeter()
         meter.load(self.loader)
-        nodes = [LocateGapsNode()]
+        nodes = [LocateGoodSectionsNode()]
         pipeline = Pipeline(nodes)
         pipeline.run(meter)
 

diff --git a/nilmtk/tests/generate_test_data.py b/nilmtk/tests/generate_test_data.py
@@ -112,7 +112,8 @@ def create_random_hdf5():
 
 
 def create_energy_hdf5(simple=True):
-    FILENAME = join(data_dir(), 'energy.h5')
+    fname = 'energy.h5' if simple else 'energy_complex.h5'
+    FILENAME = join(data_dir(), fname)
 
     df = power_data(simple=simple)