Merge pull request #294 from AllenInstitute/fix/running

missing intervalsms workaround

Author: neuromusic

setup.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.4.2
+current_version = 0.4.3
 commit = True
 tag = False
 parse = (?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(\.(?P<release>[a-z]+)(?P<n>\d+))?

setup.py

@@ -4,7 +4,7 @@
 
 setuptools.setup(
     name="visual-behavior",
-    version="0.4.2",
+    version="0.4.3",
     author="Justin Kiggins",
     author_email="[email protected]",
     description="analysis package for visual behavior",

tests/validation/test_core.py

@@ -2,6 +2,48 @@
 import pandas as pd
 from visual_behavior.validation.core import *
 
+def test_parse_log():
+
+    EXPECTED = dict(
+        levelname="ERROR",
+        name="package.module",
+        message="This is the error"
+    )
+
+    parsed = parse_log("{levelname}::{name}::{message}".format(**EXPECTED))
+    assert parsed == EXPECTED
+
+def test_count_read_errors():
+
+    EMPTY = dict(
+        log=[]
+    )
+    results = count_read_errors(EMPTY)
+    print(results)
+    assert 'ERROR' not in results
+
+    INFO = dict(
+        log=["INFO::package.module::informative message"]
+    )
+
+    results = count_read_errors(INFO)
+    print(results)
+    assert 'ERROR' not in results
+    assert results['INFO']==1
+
+def test_validate_no_read_errors():
+
+    WITH_ERRORS = dict(
+        log=["ERROR::package.module::error message"]
+    )
+
+    WITHOUT_ERRORS = dict(
+        log=["INFO::package.module::informative message"]
+    )
+
+    assert validate_no_read_errors(WITH_ERRORS)==False
+    assert validate_no_read_errors(WITHOUT_ERRORS)==True
+
 
 def test_validate_running_data():
     # good data: length matches time and not all values the same

visual_behavior/__init__.py

@@ -1 +1 @@
-__version__ = "0.4.2"
+__version__ = "0.4.3"

visual_behavior/translator/foraging/__init__.py

@@ -3,7 +3,11 @@
 import numpy as np
 from scipy.signal import medfilt
 from .extract import get_end_time
-from ...utilities import calc_deriv, rad_to_dist, local_time
+from ...utilities import calc_deriv, rad_to_dist, local_time, ListHandler, DoubleColonFormatter
+
+import logging
+
+logger = logging.getLogger(__name__)
 
 warnings.warn(
     "support for the loading stimulus_code outputs will be deprecated in a future version",
@@ -29,10 +33,20 @@ def data_to_change_detection_core(data, time=None):
     - currently doesn't require or check that the `task` field in the
     experiment data is "DoC" (Detection of Change)
     """
+
+    log_messages = []
+    handler = ListHandler(log_messages)
+    handler.setFormatter(
+        DoubleColonFormatter
+    )
+    logger.addHandler(
+        handler
+    )
+
     if time is None:
         time = load_time(data)
 
-    return {
+    core_data = {
         "time": time,
         "metadata": load_metadata(data),
         "licks": load_licks(data, time=time),
@@ -42,6 +56,9 @@ def data_to_change_detection_core(data, time=None):
         "visual_stimuli": load_visual_stimuli(data, time=time),
     }
 
+    core_data['log'] = log_messages
+    return core_data
+
 
 def load_metadata(data):
     fields = (

visual_behavior/translator/foraging2/__init__.py

@@ -1,5 +1,5 @@
 import pandas as pd
-from ...utilities import local_time
+from ...utilities import local_time, ListHandler, DoubleColonFormatter
 
 from ...devices import get_rig_id
 from .extract import get_trial_log, get_stimuli, get_pre_change_time, \
@@ -18,6 +18,10 @@
 
 from .extract_stimuli import get_visual_stimuli
 
+import logging
+
+logger = logging.getLogger(__name__)
+
 
 def data_to_change_detection_core(data):
     """Core data structure to be used across all analysis code?
@@ -37,7 +41,18 @@ def data_to_change_detection_core(data):
     - currently doesn't require or check that the `task` field in the
     experiment data is "DoC" (Detection of Change)
     """
-    return {
+
+    log_messages = []
+    handler = ListHandler(log_messages)
+    handler.setFormatter(
+        DoubleColonFormatter
+    )
+
+    logger.addHandler(
+        handler
+    )
+
+    core_data = {
         "metadata": data_to_metadata(data),
         "time": data_to_time(data),
         "licks": data_to_licks(data),
@@ -47,6 +62,10 @@ def data_to_change_detection_core(data):
         "visual_stimuli": data_to_visual_stimuli(data),
     }
 
+    core_data['log'] = log_messages
+
+    return core_data
+
 
 def expand_dict(out_dict, from_dict, index):
     """there is obviously a better way...

visual_behavior/translator/foraging2/extract.py

@@ -709,8 +709,9 @@ def get_running_speed(exp_data, smooth=False, time=None):
     dx = medfilt(dx, kernel_size=5)  # remove big, single frame spikes in encoder values
     dx = np.cumsum(dx)  # wheel rotations
 
-    if len(time) != len(dx):
-        raise ValueError("dx and time must be the same length")
+    if len(time) < len(dx):
+        logger.error('intervalsms record appears to be missing entries')
+        dx = dx[:len(time)]
 
     speed = calc_deriv(dx, time)
     speed = rad_to_dist(speed)

visual_behavior/utilities.py

@@ -1,5 +1,6 @@
 from __future__ import print_function
 from dateutil import tz
+import logging
 import numpy as np
 import pandas as pd
 from scipy.stats import norm
@@ -159,3 +160,19 @@ def local_time(iso_timestamp, timezone=None):
     if not datetime.tzinfo:
         datetime = datetime.replace(tzinfo=tz.gettz('America/Los_Angeles'))
     return datetime.isoformat()
+
+
+class ListHandler(logging.Handler):
+    """docstring for ListHandler."""
+    def __init__(self, log_list):
+        super(ListHandler, self).__init__()
+        self.log_list = log_list
+
+    def emit(self, record):
+        entry = self.format(record)
+        self.log_list.append(entry)
+
+
+DoubleColonFormatter = logging.Formatter(
+    "%(levelname)s::%(name)s::%(message)s",
+)

visual_behavior/validation/core.py

@@ -3,6 +3,29 @@
 from .extended_trials import get_first_lick_relative_to_scheduled_change
 
 
+def parse_log(log_record):
+    levelname, name, message = log_record.split('::')
+    return dict(
+        levelname=levelname,
+        name=name,
+        message=message,
+    )
+
+
+def count_read_errors(core_data):
+    log = [parse_log(log_record) for log_record in core_data['log']]
+    log = pd.DataFrame(log, columns=['levelname', 'name', 'message'])
+    return log.groupby('levelname').size().to_dict()
+
+
+def validate_no_read_errors(core_data):
+    error_counts = count_read_errors(core_data)
+
+    n_errors = error_counts.get('ERROR', 0) + error_counts.get('CRITICAL', 0)
+
+    return (n_errors == 0)
+
+
 def validate_running_data(core_data):
     '''
     for each sampling frame, the value of the encoder should be known

visual_behavior/validation/qc.py

@@ -94,6 +94,7 @@ def define_validation_functions(core_data):
         cd.validate_licks: (core_data,),  # this one doesn't take trials
         cd.validate_minimal_dropped_frames: (core_data,),  # this one doesn't take trials
         # f2.validate_frame_intervals_exists:(data), # this one doesn't take trials
+        cd.validate_no_read_errors: (core_data,),
     }
 
     return validation_functions