Error with NuPIC Geospatial Tracking Tutorial Example

[David Waldrop]

The example on GIT does NOT work.
I am pretty sure I have NUPIC installed correctly as I been able to
interact with various other Nupic examples and encoders directly.  All
seems to be working fine. 
The Geospatial Tracking Tutorial is not working for me though.

Here are the exact steps I just undertook and the corresponding system
response:


1, downloaded zip file from https://github.com/numenta/nupic.geospatial
to ensure I had the latest code.

2. extracted to /home/davidwaldrop/nupic.geospatial-master

3. ran  python tools/convert_gpx.py tests/data/sample.gpx

    Wrote output file output/sample.csv.

4. ran python run.py output/sample.csv -- System responded with:

    Traceback (most recent call last):
      File "run.py", line 120, in <module>
        not options.manualSequence)
      File "run.py", line 98, in run
        verbose=verbose)
      File
    "/home/davidwaldrop/nupic.geospatial-master/model/geospatial_anomaly.py",
    line 142, in runGeospatialAnomaly
        result = model.run(modelInput)
      File
    "/usr/local/lib/python2.7/dist-packages/nupic/frameworks/opf/clamodel.py",
    line 400, in run
        inferences = self._anomalyCompute()
      File
    "/usr/local/lib/python2.7/dist-packages/nupic/frameworks/opf/clamodel.py",
    line 618, in _anomalyCompute
        inputValue=self._input[self._predictedFieldName])
    KeyError: None


5. I ran again with -V option and saw the preprocess step discard a
bunch of the rows and emit the following:

    Running NuPIC on
    /home/davidwaldrop/nupic.geospatial-master/output/preprocessed_data.csv...
    Model parameters:
    {'model': 'CLA', 'version': 1, 'predictAheadTime': None,
    'modelParams': {'sensorParams': {'verbosity': 0, 'encoders':
    {u'vector': {'scale': 5, 'name': u'vector', 'timestep': 10, 'n':
    2048, 'fieldname': u'vector', 'w': 51, 'type':
    'GeospatialCoordinateEncoder'}}, 'sensorAutoReset': None},
    'clEnable': False, 'spParams': {'columnCount': 2048, 'spVerbosity':
    0, 'maxBoost': 1.0, 'spatialImp': 'cpp', 'synPermConnected': 0.1,
    'seed': 1956, 'numActiveColumnsPerInhArea': 40, 'globalInhibition':
    1, 'inputWidth': 0, 'synPermInactiveDec': 0.0005,
    'synPermActiveInc': 0.0001, 'potentialPct': 0.8}, 'spEnable': True,
    'clParams': None, 'inferenceType': 'TemporalAnomaly', 'tpEnable':
    True, 'anomalyParams': {u'anomalyCacheRecords': None,
    u'autoDetectThreshold': None, u'autoDetectWaitRecords': 2184},
    'tpParams': {'columnCount': 2048, 'activationThreshold': 6,
    'pamLength': 3, 'cellsPerColumn': 32, 'connectedPerm': 0.5,
    'permanenceInc': 0.1, 'minThreshold': 3, 'verbosity': 0,
    'maxSynapsesPerSegment': 32, 'outputType': 'normal', 'initialPerm':
    0.21, 'globalDecay': 0.0, 'maxAge': 0, 'permanenceDec': 0.1, 'seed':
    1960, 'newSynapseCount': 20, 'maxSegmentsPerCell': 128,
    'temporalImp': 'cpp', 'inputWidth': 2048},
    'trainSPNetOnlyIfRequested': False}}
    Traceback (most recent call last):
      File "run.py", line 120, in <module>
        not options.manualSequence)
      File "run.py", line 98, in run
        verbose=verbose)
      File
    "/home/davidwaldrop/nupic.geospatial-master/model/geospatial_anomaly.py",
    line 142, in runGeospatialAnomaly
        result = model.run(modelInput)
      File
    "/usr/local/lib/python2.7/dist-packages/nupic/frameworks/opf/clamodel.py",
    line 400, in run
        inferences = self._anomalyCompute()
      File
    "/usr/local/lib/python2.7/dist-packages/nupic/frameworks/opf/clamodel.py",
    line 618, in _anomalyCompute
        inputValue=self._input[self._predictedFieldName])
    KeyError: None
    davidwaldrop@dlwlinuxdesktop:~/nupic.geospatial-master$ python
    run.py output/sample.csv
    Traceback (most recent call last):
      File "run.py", line 120, in <module>
        not options.manualSequence)
      File "run.py", line 98, in run
        verbose=verbose)
      File
    "/home/davidwaldrop/nupic.geospatial-master/model/geospatial_anomaly.py",
    line 142, in runGeospatialAnomaly
        result = model.run(modelInput)
      File
    "/usr/local/lib/python2.7/dist-packages/nupic/frameworks/opf/clamodel.py",
    line 400, in run
        inferences = self._anomalyCompute()
      File
    "/usr/local/lib/python2.7/dist-packages/nupic/frameworks/opf/clamodel.py",
    line 618, in _anomalyCompute
        inputValue=self._input[self._predictedFieldName])
    KeyError: None




As you can see the modle is trying to run but is failing because the
predictedFieldName is None.
At this point I cannot tell if the code is broken or the model_parms.py
has errors or maybe the sample is out of sync with the Nupic code base.
I now have about 30 hours invested in nuPic and have not been able to
get this  packaged tutorial to run properly; any assistance would be
greatly appreciated.

David Waldrop