bAnalysis

The bAnalysis class represents a whole-cell recording and provides functions for analysis.

A bAnalysis object can be created in a number of ways: (i) From a file path including .abf and .csv (ii) From a pandas DataFrame when loading from a h5 file. (iii) From a byteStream abf when working in the cloud.

Once loaded, a number of operations can be performed including: Spike detection, Error checking, Plotting, and Saving.

Examples:

path = 'data/19114001.abf'
ba = bAnalysis(path)
dDict = sanpy.bDetection().getDetectionDict('SA Node')
ba.spikeDetect(dDict)
print(ba)

Source code in sanpy/bAnalysis_.py

class bAnalysis:
    """
    The bAnalysis class represents a whole-cell recording and provides functions for analysis.

    A bAnalysis object can be created in a number of ways:
    (i) From a file path including .abf and .csv
    (ii) From a pandas DataFrame when loading from a h5 file.
    (iii) From a byteStream abf when working in the cloud.

    Once loaded, a number of operations can be performed including:
    Spike detection, Error checking, Plotting, and Saving.

    Examples:

    ```python
    path = 'data/19114001.abf'
    ba = bAnalysis(path)
    dDict = sanpy.bDetection().getDetectionDict('SA Node')
    ba.spikeDetect(dDict)
    print(ba)
    ```
    """

    # def getNewUuid():
    #     return 't' + str(uuid.uuid4()).replace('-', '_')

    def __init__(
        self,
        filepath: str = None,
        byteStream=None,
        loadData: bool = True,
        fileLoaderDict: dict = None,
        stimulusFileFolder: str = None,
        verbose: bool = False,
    ):
        """
        Args:
            filepath (str): Path to either .abf or .csv with time/mV columns.
            byteStream (io.BytesIO): Binary stream for use in the cloud.
            loadData: If true, load raw data, otherwise just load header
            fileLoaderDict (dict)
                If None then fetch from sanpy.fileloaders.getFileLoaders()
                Do this if running in a script.
                If running an SanPy app, we pass the dict
            stimulusFileFolder:
        """

        """
        self._path = file  # todo: change this to filePath
        """
        """str: File path."""

        self._detectionDict: dict = None  # corresponds to an item in sanpy.bDetection

        # sept 9, moving this to file loader
        # fileloader holds meta data
        #self._metaData = MetaData(self)  #self.getMetaDataDict()

        self._isAnalyzed: bool = False

        self.loadError: bool = False
        """bool: True if error loading file/stream."""

        # self.detectionDict = None  # remember the parameters of our last detection
        """dict: Dictionary specifying detection parameters, see bDetection.getDefaultDetection."""

        # self._abf = None
        """pyAbf: If loaded from binary .abf file"""

        self.dateAnalyzed: str = None
        """str: Date Time of analysis. TODO: make a property."""

        # self.detectionType = None
        """str: From ('dvdt', 'mv')"""

        self.spikeDict: sanpy.bAnalysisResults.analysisResultList = (
            sanpy.bAnalysisResults.analysisResultList()
        )
        # class to store all analysis results

        # self._spikesPerSweep : int = None

        self.spikeClips = []  # created in self.spikeDetect()
        self.spikeClips_x = []  #
        self.spikeClips_x2 = []  #

        self.dfError = None  # dataframe with a list of detection errors
        self._dfReportForScatter = None  # dataframe to be used by scatterplotwidget

        self._detectionDirty = False

        # will be overwritten by existing uuid in self._loadFromDf()
        self.uuid = sanpy._util.getNewUuid()

        # self.tifData = None
        # when we have a tif kymograph

        # self.isBytesIO = False
        # when we are running in the cloud

        # TODO (cudmore) need to parse folder of file loaders in fileloders/ and determine
        # class to use to load file (using fileLoader.filetype
        self._fileLoader = None
        if filepath is not None and not os.path.isfile(filepath):
            logger.error(f'File does not exist: "{filepath}"')
            self.loadError = True
        else:

            if fileLoaderDict is None:
                fileLoaderDict = (
                    sanpy.fileloaders.getFileLoaders()
                )  # EXPENSIVE, to do, pass in from app
            # print('2 fileLoaderDict:', fileLoaderDict)

            # print('1 fileLoaderDict:')
            # for _k,_v in fileLoaderDict.items():
            #     print(_k,_v)

            _ext = os.path.splitext(filepath)[1]
            # _ext = _ext[1:]
            try:
                if verbose:
                    logger.info(f"Loading file with extension: {_ext}")
                constructorObject = fileLoaderDict[_ext]["constructor"]
                self._fileLoader = constructorObject(filepath)
                # may 2, 2023
                if self._fileLoader._loadError:
                    logger.error(f'load error in file loader for ext: "{_ext}"')
                    self.loadError = True

            except KeyError as e:
                logger.error(f'did not find a file loader for extension "{_ext}", available loaders are: {fileLoaderDict.keys()}')
                self.loadError = True

            self._kymAnalysis : sanpy.kymAnalysis = None
            if (self.fileLoader is not None) and (self.fileLoader.recordingMode == recordingModes.kymograph):
                if verbose:
                    logger.info('creating kymAnalysis')
                    logger.info(f'    self.fileLoader.filepath:{self.fileLoader.filepath}')
                    logger.info(f'    self.fileLoader.tifData:{self.fileLoader.tifData.shape}')
                    logger.info(f'    self.fileLoader.tifHeader:{self.fileLoader.tifHeader}')
                self._kymAnalysis = sanpy.kymAnalysis(self.fileLoader.filepath,
                                                      self.fileLoader.tifData,
                                                      self.fileLoader.tifHeader)

            if self._fileLoader is not None:
                # we need to so file loader meta data can set ba (Self) dirty when changed
                self.fileLoader.metadata._ba = self


        """
        if byteStream is not None:
            self._loadAbf(byteStream=byteStream,
                    loadData=loadData,
                    stimulusFileFolder=stimulusFileFolder)
        elif file is not None and file.endswith('.abf'):
            self._loadAbf(loadData=loadData)
        elif file is not None and file.endswith('.atf'):
            self._loadAtf(loadData=loadData)
        elif file is not None and file.endswith('.tif'):
            self._loadTif()
        elif file is not None and file.endswith('.csv'):
            self._loadCsv()
        else:
            pass
            #logger.error(f'Can only open abf/csv/tif/stream files: {file}')
            #self.loadError = True
        """

        # get default derivative
        if loadData and not self.loadError:
            self._rebuildFiltered()

        self._detectionDirty = False

        """
        self.setSweep()
        """

    @property
    def metaData(self):
        # sept 9, moved to file loader
        # return self._metaData
        return self.fileLoader.metadata

    @property
    def kymAnalysis(self):
        """Get the kymAnalysis object (if it exists).
        """
        return self._kymAnalysis

    @property
    def fileLoader(self) -> "sanpy.fileLoader_base":
        """ """
        return self._fileLoader

    def getFileName(self):
        return self.fileLoader.filename

    def asDataFrame(self, regenerateAnalysisDataFrame=False):
        """Return analysis as a Pandas DataFrame.

        Important:
            This is a df copy of our self.spikeDict
            Do not modify and expect changes to stick
        """

        # re-add file metadata, it may have changed
        if regenerateAnalysisDataFrame:
            self.regenerateAnalysisDataFrame()

        return self._dfReportForScatter
        # return self.spikeDict.asDataFrame()

    def getDetectionDict(self, asCopy: bool = False):
        """Get the detection dictionary that was used for detect()."""
        if asCopy:
            return copy.deepcopy(self._detectionDict)
        else:
            return self._detectionDict

    def __str__(self):
        """Get a brief str representation. Usefull for print()."""
        # if self.isBytesIO:
        #      filename = '<BytesIO>'
        # else:
        #     filename = self.getFileName()
        fileLoadStr = self.fileLoader.__str__()
        txt = f"fileLoader: {fileLoadStr} spikes:{self.numSpikes}"
        return txt

    def _saveHdf_pytables(self, hdfPath):
        """Save detection parameters and analysis into an hdf5 file.
        """

        # save kym diameter analysis
        if self._kymAnalysis is not None:
            if self._kymAnalysis.hasDiamAnalysis() and self._kymAnalysis._analysisDirty:
                self._kymAnalysis.saveAnalysis()


        if not self.detectionDirty:
            # Do not save it detection has not changed
            logger.info(f"NOT SAVING, is not dirty {self}")
            return False

        # always save as csv
        self.saveAnalysis_tocsv()

        # when making df from dict, need to pass it a list
        # o.w. key values that are lists get expanded into rows
        if self._detectionDict is not None:
            dfDetection = pd.DataFrame([self._detectionDict])

        dfMetaData = pd.DataFrame([self.metaData])

        # convert spikeList (list of dict) to json
        # spikeList = self.spikeDict.asList()
        # dataJson = json.dumps(spikeList, cls=NumpyEncoder)  # list of dict
        # dfAnalysis = pd.DataFrame(spikeList)
        if len(self.spikeDict) > 0:
            dfAnalysis = self.spikeDict.asDataFrame()

        uuid = self.uuid

        logger.info(
            f"    Saving {self.numSpikes} spikes to uuid {uuid} in h5 file {hdfPath}"
        )

        with pd.HDFStore(hdfPath) as hdfStore:
            if self._detectionDict is not None:
                key = uuid + "/" + "detectionDict"
                dfDetection.to_hdf(hdfStore, key)  # default mode='a'

            # always save meta data
            key = uuid + "/" + "metaDataDict"
            dfMetaData.to_hdf(hdfStore, key)  # default mode='a'

            # logger.warning('=== saving dfMetaData')
            # print(dfMetaData)

            if len(self.spikeDict) > 0:
                key = uuid + "/" + "analysisList"
                dfAnalysis.to_hdf(hdfStore, key)

        # we saved, detection is not dirty
        self._detectionDirty = False

        return True

    def _findUuid(self, hdfPath):
        """Find this analysis uuid in an h5 file. If analysis is not saved, it will not exists.
        """

        # load the database
        detectionDictKey = 'sanpy_recording_db'
        dfDetection = pd.read_hdf(hdfPath, key=detectionDictKey)

        fileList = dfDetection['File'].to_list()

        filename = self.fileLoader.filename
        try:
            idx = fileList.index(filename)
        except (ValueError) as e:
            #logger.warning(f'did not find file {filename} in file list {fileList}')
            return

        uuid = dfDetection['uuid'].to_list()[idx]

        return uuid

    def _loadHdf_pytables(self, hdfPath, uuid = None):
        """Load analysis from an h5 file using key 'uuid'.

        Parameters
        ----------
        hdfPath : str
            path to h5 file
        uuid : uuid
            Unique uuid for the file, if None then will try and find the file in hdfPath

        Notes
        -----
        If uuid is None, this only work for 'flat' directories,
        the ba has to be in same folder as h5 file
        """

        # df.to_dict() requires into=OrderedDict, o.w. column order is sorted
        # Error report needs to be generated (is not in h5 file) use getErrorReport()

        # cant use pd.HDFStore(<path>) as read_hdf does not understand file pointer

        if uuid is None:
            uuid = self._findUuid(hdfPath)
            if uuid is None:
                logger.warning(f'did not find a uuid for {self.fileLoader.filename} in h5 file {hdfPath}')
                logger.warning(f'this usually happens when the analysis was not saved')
                return

        # logger.info(f"loading {uuid} from {hdfPath}")

        # load pandas dataframe(s) from h5 file
        loadedDetection = False
        loadedMetaData = False
        loadedAnalysis = False
        try:
            detectionDictKey = uuid + "/" + "detectionDict"  # group
            dfDetection = pd.read_hdf(hdfPath, detectionDictKey)
            loadedDetection = True
        except KeyError as e:
            logger.error(f'detectionDict: {e}')
            # didLoad = False

        try:
            metaDataDictKey = uuid + "/" + "metaDataDict"  # group
            dfMetaData = pd.read_hdf(hdfPath, metaDataDictKey)
            loadedMetaData = True
        except KeyError as e:
            logger.error(f'metaDataDict: {e}')
            # didLoad = False

        try:
            analysisListKey = uuid + "/" + "analysisList"
            dfAnalysis = pd.read_hdf(hdfPath, analysisListKey)
            loadedAnalysis = True
        except KeyError as e:
            logger.error(f'analysisList: {e}')
            # didLoad = False

        # if didLoad:
        if 1:
            # we take on the uuid we were loaded from
            self.uuid = uuid

            # convert to a dict
            if loadedDetection:
                detectionDict = dfDetection.to_dict("records", into=OrderedDict)[
                    0
                ]  # one dict
                self._detectionDict = detectionDict

            if loadedMetaData:
                # create a child MEtaData object
                #logger.info('creating child MetaData')
                #self.metaData = sanpy.MetaData(self)
                #self.fileLoader.metadata = sanpy.MetaData()

                #metaDataDict = self.metaData.getMetaDataDict()  # default
                metaDataDict = sanpy.MetaData.getMetaDataDict()

                loadedMetaDataDict = dfMetaData.to_dict("records", into=OrderedDict)[
                    0
                ]  # one dict
                # we need to load current meta data dict with all current keys
                # saved file may be out of date

                # bug during implementing meta data code
                # if loadedMetaDataDict['sex'] == '':
                #     loadedMetaDataDict['sex'] = 'unknown'

                # logger.info('loadedMetaDataDict')
                # logger.info(loadedMetaDataDict)

                for k,v in loadedMetaDataDict.items():
                    if not k in metaDataDict.keys():
                        logger.error(f'  did not find loaded meta data key "{k}" in meta data keys {metaDataDict.keys()}')
                        continue
                    metaDataDict[k] = v
                self.metaData.fromDict(metaDataDict, triggerDirty=False)

                # logger.warning(f'LOADED META DATA:')
                # print('self.metaData:', self.metaData)

            # convert to a list of dict
            if loadedAnalysis:
                analysisList = dfAnalysis.to_dict(
                    "records", into=OrderedDict
                )  # list of dict
                self.spikeDict.setFromListDict(analysisList)
                # pprint(analysisList[0])

                # recreate spike analysis dataframe
                # self._dfReportForScatter = dfAnalysis
                self.regenerateAnalysisDataFrame()

                # regenerate error report
                self.dfError = self.getErrorReport()

                # dec 2022
                self._isAnalyzed = True

            # logger.info(
            #     f"    loaded {len(detectionDict.keys())} detection keys and {len(self.spikeDict)} spikes"
            # )
        else:
            logger.error(f"    LOAD FAILED")

    @property
    def detectionDirty(self):
        return self._detectionDirty

    @property
    def numSpikes(self):
        """Get the total number of detected spikes (all sweeps).

        See getNumSpikes(sweep)
        """
        return len(self.spikeDict)  # spikeDict has all spikes for all sweeps

    def getNumSpikes(self, sweep: int = 0):
        """Get number of spikes in a sweep.

        See property numSpikes
        """
        thresholdSec = self.getStat("thresholdSec", sweepNumber=sweep)
        return len(thresholdSec)
        # return self._spikesPerSweep[sweep]

    @property
    def numErrors(self) -> int:
        """Get number of detection errors.
        """
        if self.dfError is None:
            # no analysis
            return None
        else:
            return len(self.dfError)

    def _old_getAbsSpikeFromSweep(self, sweepSpikeIdx: int, sweep: int) -> int:
        """Given a spike index within a sweep, get the absolute spike index.

        See getSweepSpikeFromAbsolute()
        """
        absIdx = 0
        for sweepIdx in range(sweep):
            absIdx += self._spikesPerSweep[sweepIdx]
        absIdx += sweepSpikeIdx
        return absIdx

    def getSweepSpikeFromAbsolute(self, absSpikeIdx: int, sweep: int) -> int:
        """Get sweep spike from absolute spike.

        See getAbsSpikeFromSweep()
        """
        sweepSpikeNum = self.spikeDict[absSpikeIdx]["sweepSpikeNumber"]
        return sweepSpikeNum

        # absIdx = 0
        # for oneSweep in range(sweep):
        #     absIdx += self._spikesPerSweep[oneSweep]
        # sweepSpike = absSpikeIdx - absIdx
        # return sweepSpike

    def isDirty(self):
        """Return True if analysis has been modified but not save."""
        return self._detectionDirty

    def isAnalyzed(self):
        """Return True if this bAnalysis has been analyzed, False otherwise."""
        return self._isAnalyzed

    def getStatMean(self, statName: str, sweepNumber: int = None):
        """
        Get the mean of an analysis parameter.

        Args:
            statName (str): Name of the statistic to retreive.
                For a list of available stats use bDetection.defaultDetection.
        """
        theMean = None
        x = self.getStat(statName, sweepNumber=sweepNumber)
        if x is not None and len(x) > 1:
            theMean = np.nanmean(x)
        return theMean

    def getSpikeStat(self, spikeList : List[int], stat : str):
        """Get one stat from a list of spikes

        Parameters
        ----------
        spikeList : List[int]
        stat : str
        """

        # if isinstance(spikeList, int):
        #     spikeList = [spikeList]

        if len(spikeList) == 0:
            return None

        # logger.info(f'spikeList: {spikeList} stat:{stat}')

        retList = []
        # count = 0
        for idx, spike in enumerate(self.spikeDict):
            # logger.info(f'  idx:{idx}')
            if idx in spikeList:
                try:
                    val = spike[stat]
                    retList.append(val)
                    # count += 1
                except KeyError as e:
                    logger.error(e)
        # logger.info(f'  retList: {retList}')
        return retList

    def setSpikeStat_time(self, startSec: int, stopSec: int, stat: str, value):
        """Set a spike stat for spikes in a range of time."""

        # get spike list in range [startSec, stopSec]
        spikeSeconds = self.getSpikeSeconds()
        spikeList = [
            idx for idx, x in enumerate(spikeSeconds) if x >= startSec and x < stopSec
        ]
        self.setSpikeStat(spikeList, stat, value)

    def setSpikeStat(self, spikeList: Union[list, int], stat: str, value):
        """Set a spike stat for one spike or a list of spikes.

        Used to set things like ('isBad', 'userType1', 'condition', ...)
        """
        if isinstance(spikeList, int):
            spikeList = [spikeList]
            # else:
            #     logger.error(f'Expecting list[int] or int but got spikeList type {type(spikeList)}')
            return

        if len(spikeList) == 0:
            return

        now = datetime.datetime.now()
        modDate = now.strftime("%Y%m%d")
        modTime = now.strftime("%H:%M:%S")

        for spike in spikeList:
            self.spikeDict[spike][stat] = value
            self.spikeDict[spike]["modDate"] = modDate
            self.spikeDict[spike]["modTime"] = modTime

        self._detectionDirty = True

        logger.info(f'set spikes {spikeList} stat "{stat}" to value "{value}"')

        """
        count = 0
        for idx, spike in enumerate(self.spikeDict):
            if idx in spikeList:
                try:
                    spike[stat] = value
                    count += 1
                except (KeyError) as e:
                    logger.info(e)
        #
        logger.info(f'Given {len(spikeList)} and set {count}')
        """

    def _old_getSweepStats(
        self, statName: str, decimals=3, asDataFrame=False, df: pd.DataFrame = None
    ):
        """

        Args:
            df (pd.DataFrame): For kymograph we sometimes have to convert (peak) values to molar
        """

        if df is None:
            df = self.spikeDict.asDataFrame()

        sweepStatList = []

        for sweep in range(self.fileLoader.numSweeps):
            oneDf = df[df["sweep"] == sweep]
            theValues = oneDf[statName]

            theCount = np.count_nonzero(~np.isnan(theValues))
            theMin = np.min(theValues)
            theMax = np.max(theValues)
            theMean = np.nanmean(theValues)

            theMin = round(theMin, decimals)
            theMax = round(theMax, decimals)
            theMean = round(theMean, decimals)

            if theCount > 2:
                theMedian = np.nanmedian(theValues)
                theSEM = scipy.stats.sem(theValues)
                theSD = np.nanstd(theValues)
                theVar = np.nanvar(theValues)
                theCV = theSD / theVar

                theMedian = round(theMedian, decimals)
                theSEM = round(theSEM, decimals)
                theSD = round(theSD, decimals)
                theVar = round(theVar, decimals)
                theCV = round(theCV, decimals)

            else:
                theMedian = None
                theSEM = None
                theSD = None
                theVar = None
                theCV = None

            oneDict = {
                statName + "_sweep": sweep,
                statName + "_count": theCount,
                statName + "_min": theMin,
                statName + "_max": theMax,
                statName + "_mean": theMean,
                statName + "_median": theMedian,
                statName + "_sem": theSEM,
                statName + "_std": theSD,
                statName + "_var": theVar,
                statName + "_cv": theCV,
            }

            sweepStatList.append(oneDict)

        #
        if asDataFrame:
            return pd.DataFrame(sweepStatList)
        else:
            return sweepStatList

    def getStat(
        self,
        statName1,
        statName2: Optional[str] = None,
        sweepNumber: Optional[int] = None,
        epochNumber: Optional[int] = None,
        asArray: Optional[bool] = False,
        getFullList : Optional[bool] = False
    ):
        """Get a list of values for one or two analysis results.

        Parameters
        ----------
        statName1 : str
            Name of the first analysis parameter to retreive.
        statName2 : str
            Optional name of the second analysis parameter to retreive.
        sweepNumber : int str or None
            Optional sweep number, if None or 'All' then get all sweeps
        epochNumber : int str or None
            Optional epoch number, if None or 'All' then get all epochs
        asArray : bool
            If True then return as np.array(), otherwise return as a list

        Notes
        -----
        For a list of available analysis results,
            see [bDetection.getDefaultDetection()][sanpy.bDetection.bDetection]

        If the returned list of analysis results are in points,
            convert to seconds or ms using: pnt2Sec_(pnt) or pnt2Ms_(pnt).

        Returns
        -------
        list or np.array
            List of analysis parameter values, None if error.
            Returns a np.array is asArray is True
        """

        def clean(val):
            """Convert None to float('nan')"""
            if val is None:
                val = float("nan")
            return val

        x = []  # None
        y = []  # None
        error = False
        if len(self.spikeDict) == 0:
            # logger.error(f'Did not find any spikes in spikeDict')
            error = True
        elif statName1 not in self.spikeDict[0].keys():
            logger.error(f'Did not find statName1: "{statName1}" in spikeDict')
            # print('available stat names are:', self.spikeDict[0].keys())
            error = True
        elif statName2 is not None and statName2 not in self.spikeDict[0].keys():
            logger.error(f'Did not find statName2: "{statName2}" in spikeDict')
            error = True

        if sweepNumber is None:
            sweepNumber = "All"

        if epochNumber is None:
            epochNumber = "All"

        if not error:
            # original
            # x = [clean(spike[statName1]) for spike in self.spikeDict]

            if getFullList:
                # April 15, 2023, trying to fix bug in scatter plugin when we are
                # using sweep and epoch
                # strategy is to return all spikes, just nan out the ones we 
                # are not interested in
                x = []
                for spike in self.spikeDict:
                    _include = \
                        (sweepNumber == "All" or spike["sweep"] == sweepNumber) \
                            and (epochNumber == "All" or spike["epoch"] == epochNumber)
                    if _include:
                        x.append(clean(spike[statName1]))
                    else:
                        x.append(float("nan"))

            else:
                # only current sweep and epoch
                # (1) was this
                # was causing errors with kym diam analysis
                x = [
                    clean(spike[statName1])
                    for spike in self.spikeDict
                    if (sweepNumber == "All" or spike["sweep"] == sweepNumber)
                    and (epochNumber == "All" or spike["epoch"] == epochNumber)
                ]
                # for _idx, spike in enumerate(self.spikeDict):
                #     if (sweepNumber == "All" or spike["sweep"] == sweepNumber) and (epochNumber == "All" or spike["epoch"] == epochNumber):
                #         try:
                #             val = spike[statName1]
                #         except (KeyError) as e:
                #             logger.error(f'did not find key "{statName1}" at spike {_idx}')
                #         clean(val)

            if statName2 is not None:
                # original
                # y = [clean(spike[statName2]) for spike in self.spikeDict]
                # only current spweek
                y = [
                    clean(spike[statName2])
                    for spike in self.spikeDict
                    if sweepNumber == "All" or spike["sweep"] == sweepNumber
                ]

        if asArray:
            x = np.array(x)
            if statName2 is not None:
                y = np.array(y)

        if statName2 is not None:
            return x, y
        else:
            return x

    def getSpikeTimes(self, sweepNumber=None, epochNumber='All'):
        """Get spike times (points) for current sweep"""
        # theRet = [spike['thresholdPnt'] for spike in self.spikeDict if spike['sweep']==self.currentSweep]
        theRet = self.getStat("thresholdPnt", sweepNumber=sweepNumber, epochNumber=epochNumber)
        return theRet

    def getSpikeSeconds(self, sweepNumber=None):
        """Get spike times (seconds) for current sweep"""
        # theRet = [spike['thresholdSec'] for spike in self.spikeDict if spike['sweep']==self.currentSweep]
        theRet = self.getStat("thresholdSec", sweepNumber=sweepNumber)
        return theRet

    def getSpikeDictionaries(self, sweepNumber=None):
        """Get spike dictionaries for current sweep
        """
        if sweepNumber is None:
            sweepNumber = "All"
        # logger.info(f'sweepNumber:{sweepNumber}')
        theRet = [
            spike
            for spike in self.spikeDict
            if sweepNumber == "All" or spike["sweep"] == sweepNumber
        ]
        return theRet

    def getOneSpikeDict(self, spikeNumber: int):
        return self.spikeDict[spikeNumber]

    def _rebuildFiltered(self):
        if self.fileLoader.sweepX is None:
            # no data
            logger.warning("not getting derivative ... sweepX was none?")
            return

        if (
            self.fileLoader.recordingMode == recordingModes.iclamp
            or self.fileLoader.recordingMode == recordingModes.kymograph
        ):
            self.fileLoader._getDerivative()
        elif self.fileLoader.recordingMode == recordingModes.vclamp:
            self.fileLoader._getDerivative()
        else:
            logger.warning(
                f'Did not take derivative, unknown recording mode "{self.fileLoader.recordingMode}"'
            )

    def _getFilteredRecording(self):
        """
        Get a filtered version of recording, used for both V-Clamp and I-Clamp.

        Args:
            dDict (dict): Default detection dictionary. See bDetection.defaultDetection
        """

        if self._detectionDict is not None:
            medianFilter = self._detectionDict["medianFilter"]
            SavitzkyGolay_pnts = self._detectionDict["SavitzkyGolay_pnts"]
            SavitzkyGolay_poly = self._detectionDict["SavitzkyGolay_poly"]
        else:
            # we have not been analyzed, impose some defaults
            medianFilter = 0  # no median filter
            SavitzkyGolay_pnts = 5
            SavitzkyGolay_poly = 2

        self.fileLoader._getDerivative(
            medianFilter, SavitzkyGolay_pnts, SavitzkyGolay_poly
        )

        # if medianFilter > 0:
        #     if not medianFilter % 2:
        #         medianFilter += 1
        #         logger.warning(f'Please use an odd value for the median filter, set medianFilter: {medianFilter}')
        #     medianFilter = int(medianFilter)
        #     self._filteredVm = scipy.signal.medfilt2d(self.sweepY(), [medianFilter,1])
        # elif SavitzkyGolay_pnts > 0:
        #     self._filteredVm = scipy.signal.savgol_filter(self.sweepY(),
        #                         SavitzkyGolay_pnts, SavitzkyGolay_poly,
        #                         mode='nearest', axis=0)
        # else:
        #     self._filteredVm = self.sweepY

    def _backupSpikeVm(self, spikeTimes, sweepNumber, medianFilter=None):
        """
        Backup spike time using deminishing SD and diff b/w vm at pnt[i]-pnt[i-1]
        Used when detecting with just mV threshold (not dv/dt)

        Args:
            spikeTimes (list of float):
            medianFilter (int): bin width
        """
        # realSpikeTimePnts = [np.nan] * self.numSpikes
        realSpikeTimePnts = [np.nan] * len(spikeTimes)

        medianFilter = 5
        sweepY = self.fileLoader.sweepY
        if medianFilter > 0:
            myVm = scipy.signal.medfilt(sweepY, medianFilter)
        else:
            myVm = sweepY

        #
        # TODO: this is going to fail if spike is at start/stop of recorrding
        #

        maxNumPntsToBackup = 20  # todo: add _ms
        bin_ms = 1
        bin_pnts = round(bin_ms * self.fileLoader.dataPointsPerMs)
        half_bin_pnts = math.floor(bin_pnts / 2)
        for idx, spikeTimePnts in enumerate(spikeTimes):
            foundRealThresh = False
            thisMean = None
            thisSD = None
            backupNumPnts = 0
            atBinPnt = spikeTimePnts
            while not foundRealThresh:
                thisWin = myVm[atBinPnt - half_bin_pnts : atBinPnt + half_bin_pnts]
                if thisMean is None:
                    thisMean = np.mean(thisWin)
                    thisSD = np.std(thisWin)

                nextStart = atBinPnt - 1 - bin_pnts - half_bin_pnts
                nextStop = atBinPnt - 1 - bin_pnts + half_bin_pnts
                nextWin = myVm[nextStart:nextStop]
                nextMean = np.mean(nextWin)
                nextSD = np.std(nextWin)

                meanDiff = thisMean - nextMean
                # logic
                sdMult = 0.7  # 2
                if (meanDiff < nextSD * sdMult) or (
                    backupNumPnts == maxNumPntsToBackup
                ):
                    # second clause will force us to terminate (this recording has a very slow rise time)
                    # bingo!
                    foundRealThresh = True
                    # not this xxx but the previous
                    moveForwardPnts = 4
                    backupNumPnts = backupNumPnts - 1  # the prev is thresh
                    if backupNumPnts < moveForwardPnts:
                        logger.warning(
                            f"spike {idx} backupNumPnts:{backupNumPnts} < moveForwardPnts:{moveForwardPnts}"
                        )
                        # print('  -->> not adjusting spike time')
                        realBackupPnts = backupNumPnts - 0
                        realPnt = spikeTimePnts - (realBackupPnts * bin_pnts)

                    else:
                        realBackupPnts = backupNumPnts - moveForwardPnts
                        realPnt = spikeTimePnts - (realBackupPnts * bin_pnts)
                    #
                    realSpikeTimePnts[idx] = realPnt

                # increment
                thisMean = nextMean
                thisSD = nextSD

                atBinPnt -= bin_pnts
                backupNumPnts += 1
                """
                if backupNumPnts>maxNumPntsToBackup:
                    print(f'  WARNING: _backupSpikeVm() exiting spike {idx} ... reached maxNumPntsToBackup:{maxNumPntsToBackup}')
                    print('  -->> not adjusting spike time')
                    foundRealThresh = True # set this so we exit the loop
                    realSpikeTimePnts[idx] = spikeTimePnts
                """

        #
        return realSpikeTimePnts

    def _throwOutRefractory(self, spikeTimes0, goodSpikeErrors, refractory_ms=20):
        """
        spikeTimes0: spike times to consider
        goodSpikeErrors: list of errors per spike, can be None
        refractory_ms:
        """
        before = len(spikeTimes0)

        # if there are doubles, throw-out the second one
        # refractory_ms = 20 #10 # remove spike [i] if it occurs within refractory_ms of spike [i-1]
        lastGood = 0  # first spike [0] will always be good, there is no spike [i-1]
        for i in range(len(spikeTimes0)):
            if i == 0:
                # first spike is always good
                continue
            dPoints = spikeTimes0[i] - spikeTimes0[lastGood]
            if dPoints < self.fileLoader.dataPointsPerMs * refractory_ms:
                # remove spike time [i]
                spikeTimes0[i] = 0
            else:
                # spike time [i] was good
                lastGood = i
        # regenerate spikeTimes0 by throwing out any spike time that does not pass 'if spikeTime'
        # spikeTimes[i] that were set to 0 above (they were too close to the previous spike)
        # will not pass 'if spikeTime', as 'if 0' evaluates to False
        if goodSpikeErrors is not None:
            goodSpikeErrors = [
                goodSpikeErrors[idx]
                for idx, spikeTime in enumerate(spikeTimes0)
                if spikeTime
            ]
        spikeTimes0 = [spikeTime for spikeTime in spikeTimes0 if spikeTime]

        # TODO: put back in and log if detection ['verbose']
        after = len(spikeTimes0)
        if self._detectionDict["verbose"]:
            logger.info(
                f"From {before} to {after} spikes with refractory_ms:{refractory_ms}"
            )

        return spikeTimes0, goodSpikeErrors

    def _getHalfWidth(
        self,
        vm,
        iIdx,
        spikeDict,
        thresholdPnt,
        peakPnt,
        hwWindowPnts,
        dataPointsPerMs,
        halfHeightList,
        verbose=False,
    ):
        """
        Get half-widhts for one spike.

        Note: Want to make this standalone function outside of class but we need self._getErrorDict()

        Args:
            vm ():
            iIdx (int):
            spikeDict (): new 20210928
            #dictNumber (int):
            thresholdPnt (int): AP threshold crossing
            peakPnt (int): AP peak
            hwWindowPnts (int): Window to look after peakPnt for falling vm
            dataPointsPerMs (int):
            halfHeightList (list): List of half-height [10,20,50,80,90]
        """

        halfWidthWindow_ms = hwWindowPnts / dataPointsPerMs

        thresholdVal = vm[thresholdPnt]
        peakVal = vm[peakPnt]
        spikeHeight = peakVal - thresholdVal

        spikeSecond = thresholdPnt / dataPointsPerMs / 1000
        peakSec = peakPnt / dataPointsPerMs / 1000

        widthDictList = []
        errorList = []

        # clear out any existing list
        spikeDict[iIdx]["widths"] = []

        tmpErrorType = None
        for j, halfHeight in enumerate(halfHeightList):
            # halfHeight in [20, 50, 80]

            # search rising/falling phae of vm for this vm
            thisVm = thresholdVal + spikeHeight * (halfHeight * 0.01)

            # todo: logic is broken, this get over-written in following try
            widthDict = {
                "halfHeight": halfHeight,
                "risingPnt": None,
                #'risingVal': defaultVal,
                "fallingPnt": None,
                #'fallingVal': defaultVal,
                "widthPnts": None,
                "widthMs": float("nan"),
            }
            widthMs = float("nan")
            try:
                postRange = vm[peakPnt : peakPnt + hwWindowPnts]
                fallingPnt = np.where(postRange < thisVm)[0]  # less than
                if len(fallingPnt) == 0:
                    # no falling pnts found within hwWindowPnts
                    tmpErrorType = "falling point"
                    raise IndexError
                fallingPnt = fallingPnt[0]  # first falling point
                fallingPnt += peakPnt
                fallingVal = vm[fallingPnt]

                # use the post/falling to find pre/rising
                preRange = vm[thresholdPnt:peakPnt]
                risingPnt = np.where(preRange > fallingVal)[0]  # greater than
                if len(risingPnt) == 0:
                    tmpErrorType = "rising point"
                    raise IndexError
                risingPnt = risingPnt[0]  # first rising point
                risingPnt += thresholdPnt
                # risingVal = vm[risingPnt]

                # width (pnts)
                widthPnts = fallingPnt - risingPnt
                widthMs = widthPnts / dataPointsPerMs
                # 20210825 may want to add this to analysis
                # widthPnts2 = fallingPnt - thresholdPnt
                # assign
                widthDict["halfHeight"] = halfHeight
                widthDict["risingPnt"] = risingPnt
                # widthDict['risingVal'] = risingVal
                widthDict["fallingPnt"] = fallingPnt
                # widthDict['fallingVal'] = fallingVal
                widthDict["widthPnts"] = widthPnts
                widthDict["widthMs"] = widthMs
                # widthMs = widthPnts / dataPointsPerMs # abb 20210125

                # may want to add this
                # widthDict['widthPnts2'] = widthPnts2
                # widthDict['widthMs2'] = widthPnts2 / dataPointsPerMs

            except IndexError as e:
                errorType = "Spike Width"
                errorStr = (
                    f'Half width {halfHeight} error in "{tmpErrorType}" '
                    f"with halfWidthWindow_ms:{halfWidthWindow_ms} "
                    f"searching for Vm:{round(thisVm,2)} from peak sec {round(peakSec,2)}"
                )

                # was this
                # eDict = self._getErrorDict(spikeNumber, thresholdPnt, errorType, errorStr) # spikeTime is in pnts
                eDict = self._getErrorDict(
                    iIdx, thresholdPnt, errorType, errorStr
                )  # spikeTime is in pnts
                # self.spikeDict[dictNumber]['errors'].append(eDict)
                spikeDict[iIdx]["errors"].append(eDict)
                if verbose:
                    print(
                        f"_getHalfWidth() error iIdx:{iIdx} j:{j} halfHeight:{halfHeight} eDict:{eDict}"
                    )
            #
            # self.spikeDict[dictNumber]['widths_'+str(halfHeight)] = widthMs
            # self.spikeDict[dictNumber]['widths'][j] = widthDict

            # logger.info('================')
            # print(f'len(spikeDict):{len(spikeDict)} iIdx:{iIdx} j:{j} widthDict:{widthDict}')

            spikeDict[iIdx]["widths_" + str(halfHeight)] = widthMs
            # spikeDict[iIdx]['widths'][j] = widthDict
            spikeDict[iIdx]["widths"].append(widthDict)

        #
        # return widthDictList, errorList

    def _getErrorDict(self, spikeNumber, pnt, _type : str, detailStr) -> dict:
        """Get error dict for one spike

        Notes
        -----
        Can't use self.getSpikeStat() because it is not created yet.
            We are in the middle of analysis
        """
        sec = self.fileLoader.pnt2Sec_(pnt)  # pnt / self.dataPointsPerMs / 1000
        sec = round(sec, 4)

        # print(f'  spikeNumber: {spikeNumber} {type(spikeNumber)}')
        # print('    sweep:', self.getSpikeStat([spikeNumber], 'sweep'))

        eDict = {
            "Spike": spikeNumber,
            "Seconds": sec,
            "Sweep": '',  # self.getSpikeStat([spikeNumber], 'sweep')[0],
            "Epoch": '',  # self.getSpikeStat([spikeNumber], 'epoch')[0],
            "Type": _type,
            "Details": detailStr,
        }
        return eDict

    def _spikeDetect_dvdt(self, dDict: dict, sweepNumber: int, verbose: bool = False):
        """
        Search for threshold crossings (dvdtThreshold) in first derivative (dV/dt) of membrane potential (Vm)
        append each threshold crossing (e.g. a spike) in self.spikeTimes list

        Returns:
            self.spikeTimes (pnts): the time before each threshold crossing when dv/dt crosses 15% of its max
            self.filteredVm:
            self.filtereddVdt:
        """

        #
        # analyze full recording
        filteredDeriv = self.fileLoader.filteredDeriv
        Is = np.where(filteredDeriv > dDict["dvdtThreshold"])[0]
        Is = np.concatenate(([0], Is))
        Ds = Is[:-1] - Is[1:] + 1
        spikeTimes0 = Is[np.where(Ds)[0] + 1]

        #
        # reduce spike times based on start/stop
        # logger.error('THIS IS a BUg if start sec is none then set to 0 !!!')
        # THIS IS ABUG ... FIX
        if dDict["startSeconds"] is not None and dDict["stopSeconds"] is not None:
            startPnt = self.fileLoader.dataPointsPerMs * (
                dDict["startSeconds"] * 1000
            )  # seconds to pnt
            stopPnt = self.fileLoader.dataPointsPerMs * (
                dDict["stopSeconds"] * 1000
            )  # seconds to pnt
            tmpSpikeTimes = [
                spikeTime
                for spikeTime in spikeTimes0
                if (spikeTime >= startPnt and spikeTime <= stopPnt)
            ]
            spikeTimes0 = tmpSpikeTimes

        #
        # throw out all spikes that are below a threshold Vm (usually below -20 mV)
        peakWindow_pnts = self.fileLoader.ms2Pnt_(dDict["peakWindow_ms"])
        # peakWindow_pnts = self.dataPointsPerMs * dDict['peakWindow_ms']
        # peakWindow_pnts = round(peakWindow_pnts)
        goodSpikeTimes = []
        sweepY = self.fileLoader.sweepY
        for spikeTime in spikeTimes0:
            # wu-lab-stanford data
            try:
                peakVal = np.max(sweepY[spikeTime : spikeTime + peakWindow_pnts])
                if peakVal > dDict["mvThreshold"]:
                    goodSpikeTimes.append(spikeTime)
            except (ValueError) as e:
                logger.error(e)
                logger.error(f'   spikeTime:{spikeTime} peakWindow_pnts:{peakWindow_pnts}')
                logger.error(f'   _dataPointsPerMs: {self.fileLoader._dataPointsPerMs}')
        spikeTimes0 = goodSpikeTimes

        #
        # throw out spike that are not upward deflections of Vm
        """
        prePntUp = 7 # pnts
        goodSpikeTimes = []
        for spikeTime in spikeTimes0:
            preAvg = np.average(self.abf.sweepY[spikeTime-prePntUp:spikeTime-1])
            postAvg = np.average(self.abf.sweepY[spikeTime+1:spikeTime+prePntUp])
            #print(preAvg, postAvg)
            if preAvg < postAvg:
                goodSpikeTimes.append(spikeTime)
        spikeTimes0 = goodSpikeTimes
        """

        #
        # if there are doubles, throw-out the second one
        spikeTimeErrors = None
        spikeTimes0, ignoreSpikeErrors = self._throwOutRefractory(
            spikeTimes0, spikeTimeErrors, refractory_ms=dDict["refractory_ms"]
        )

        # logger.warning('REMOVED SPIKE TOP AS % OF DVDT')
        # return spikeTimes0, [None] * len(spikeTimes0)

        #
        # for each threshold crossing, search backwards in dV/dt for a % of maximum (about 10 ms)
        # dvdt_percentOfMax = 0.1
        # window_ms = 2
        window_pnts = dDict["dvdtPreWindow_ms"] * self.fileLoader.dataPointsPerMs
        # abb 20210130 lcr analysis
        window_pnts = round(window_pnts)
        spikeTimes1 = []
        spikeErrorList1 = []
        filteredDeriv = self.fileLoader.filteredDeriv
        for i, spikeTime in enumerate(spikeTimes0):
            # get max in derivative

            preDerivClip = filteredDeriv[
                spikeTime - window_pnts : spikeTime
            ]  # backwards
            postDerivClip = filteredDeriv[
                spikeTime : spikeTime + window_pnts
            ]  # forwards

            if len(preDerivClip) == 0:
                print(
                    "FIX ERROR: spikeDetect_dvdt()",
                    "spike",
                    i,
                    "at pnt",
                    spikeTime,
                    "window_pnts:",
                    window_pnts,
                    "dvdtPreWindow_ms:",
                    dDict["dvdtPreWindow_ms"],
                    "len(preDerivClip)",
                    len(preDerivClip),
                )  # preDerivClip = np.flip(preDerivClip)

            # look for % of max in dvdt
            try:
                # peakPnt = np.argmax(preDerivClip)
                peakPnt = np.argmax(postDerivClip)
                # peakPnt += spikeTime-window_pnts
                peakPnt += spikeTime
                peakVal = filteredDeriv[peakPnt]

                percentMaxVal = (
                    peakVal * dDict["dvdt_percentOfMax"]
                )  # value we are looking for in dv/dt
                preDerivClip = np.flip(preDerivClip)  # backwards
                tmpWhere = np.where(preDerivClip < percentMaxVal)
                # print('tmpWhere:', type(tmpWhere), tmpWhere)
                tmpWhere = tmpWhere[0]
                if len(tmpWhere) > 0:
                    threshPnt2 = np.where(preDerivClip < percentMaxVal)[0][0]
                    threshPnt2 = (spikeTime) - threshPnt2
                    # print('i:', i, 'spikeTime:', spikeTime, 'peakPnt:', peakPnt, 'threshPnt2:', threshPnt2)
                    threshPnt2 -= 1  # backup by 1 pnt
                    spikeTimes1.append(threshPnt2)
                    spikeErrorList1.append(None)

                else:
                    errorType = "dvdt Percent"
                    errStr = f"Did not find dvdt_percentOfMax: {dDict['dvdt_percentOfMax']} peak dV/dt is {round(peakVal,2)}"
                    eDict = self._getErrorDict(
                        i, spikeTime, errorType, errStr
                    )  # spikeTime is in pnts
                    spikeErrorList1.append(eDict)
                    # always append, do not REJECT spike if we can't find % in dv/dt
                    spikeTimes1.append(spikeTime)
            except (IndexError, ValueError) as e:
                ##
                print(
                    "   FIX ERROR: bAnalysis.spikeDetect_dvdt() looking for dvdt_percentOfMax"
                )
                print("      ", "IndexError for spike", i, spikeTime)
                print("      ", e)
                # always append, do not REJECT spike if we can't find % in dv/dt
                spikeTimes1.append(spikeTime)

        return spikeTimes1, spikeErrorList1

    def _spikeDetect_vm(self, dDict: dict, sweepNumber: int, verbose: bool = False):
        """
        spike detect using Vm threshold and NOT dvdt
        append each threshold crossing (e.g. a spike) in self.spikeTimes list

        Returns:
            self.spikeTimes (pnts): the time before each threshold crossing when dv/dt crosses 15% of its max
            self.filteredVm:
            self.filtereddVdt:
        """

        filteredVm = self.fileLoader.sweepY_filtered
        Is = np.where(filteredVm > dDict["mvThreshold"])[0]  # returns boolean array
        Is = np.concatenate(([0], Is))
        Ds = Is[:-1] - Is[1:] + 1
        spikeTimes0 = Is[np.where(Ds)[0] + 1]

        #
        # reduce spike times based on start/stop
        if dDict["startSeconds"] is not None and dDict["stopSeconds"] is not None:
            startPnt = self.fileLoader.dataPointsPerMs * (
                dDict["startSeconds"] * 1000
            )  # seconds to pnt
            stopPnt = self.fileLoader.dataPointsPerMs * (
                dDict["stopSeconds"] * 1000
            )  # seconds to pnt
            tmpSpikeTimes = [
                spikeTime
                for spikeTime in spikeTimes0
                if (spikeTime >= startPnt and spikeTime <= stopPnt)
            ]
            spikeTimes0 = tmpSpikeTimes

        spikeErrorList = [None] * len(spikeTimes0)

        #
        # throw out all spikes that are below a threshold Vm (usually below -20 mV)
        # spikeTimes0 = [spikeTime for spikeTime in spikeTimes0 if self.abf.sweepY[spikeTime] > self.mvThreshold]
        # 20190623 - already done in this vm threshold funtion
        """
        peakWindow_ms = 10
        peakWindow_pnts = self.abf.dataPointsPerMs * peakWindow_ms
        goodSpikeTimes = []
        for spikeTime in spikeTimes0:
            peakVal = np.max(self.abf.sweepY[spikeTime:spikeTime+peakWindow_pnts])
            if peakVal > self.mvThreshold:
                goodSpikeTimes.append(spikeTime)
        spikeTimes0 = goodSpikeTimes
        """

        #
        # throw out spike that are NOT upward deflections of Vm
        tmpLastGoodSpike_pnts = None
        # minISI_pnts = 5000 # at 20 kHz this is 0.25 sec
        minISI_ms = 75  # 250
        minISI_pnts = self.fileLoader.ms2Pnt_(minISI_ms)

        prePntUp = 10  # pnts
        goodSpikeTimes = []
        goodSpikeErrors = []
        sweepY = self.fileLoader.sweepY
        for tmpIdx, spikeTime in enumerate(spikeTimes0):
            tmpFuckPreClip = sweepY[
                spikeTime - prePntUp : spikeTime
            ]  # not including the stop index
            tmpFuckPostClip = sweepY[
                spikeTime + 1 : spikeTime + prePntUp + 1
            ]  # not including the stop index
            preAvg = np.average(tmpFuckPreClip)
            postAvg = np.average(tmpFuckPostClip)
            if postAvg > preAvg:
                # tmpSpikeTimeSec = self.fileLoader.pnt2Sec_(spikeTime)
                if (
                    tmpLastGoodSpike_pnts is not None
                    and (spikeTime - tmpLastGoodSpike_pnts) < minISI_pnts
                ):
                    continue
                goodSpikeTimes.append(spikeTime)
                goodSpikeErrors.append(spikeErrorList[tmpIdx])
                tmpLastGoodSpike_pnts = spikeTime
            else:
                tmpSpikeTimeSec = self.fileLoader.pnt2Sec_(spikeTime)

        # todo: add this to spikeDetect_dvdt()
        goodSpikeTimes, goodSpikeErrors = self._throwOutRefractory(
            goodSpikeTimes, goodSpikeErrors, refractory_ms=dDict["refractory_ms"]
        )
        spikeTimes0 = goodSpikeTimes
        spikeErrorList = goodSpikeErrors

        #
        return spikeTimes0, spikeErrorList

    def spikeDetect(self, detectionDict: dict):
        """Run spike detection for all sweeps.

        Each spike is a row and has 'sweep'

        Args:
            detectionDict: From sanpy.bDetection
        """

        rememberSweep = (
            self.fileLoader.currentSweep
        )  # This is BAD we are mixing analysis with interface !!!

        startTime = time.time()

        #
        # todo: ask user if they want to remove their settings for (isBad, userType)
        #

        self._detectionDict = detectionDict

        if detectionDict["verbose"]:
            logger.info("=== detectionDict is:")
            for k in detectionDict.keys():
                v = detectionDict[k]
                print(f'  {k} value:"{v}" is type {type(v)}')

        self._isAnalyzed = True

        self.spikeDict = sanpy.bAnalysisResults.analysisResultList()
        # we are filling this in, one dict for each spike
        # self.spikeDict = [] # we are filling this in, one dict for each spike

        # self._spikesPerSweep = [0] * self.fileLoader.numSweeps

        for sweepNumber in self.fileLoader.sweepList:
            # self.setSweep(sweep)
            self._spikeDetect2(sweepNumber)

        #
        self.fileLoader.setSweep(rememberSweep)

        stopTime = time.time()

        if detectionDict["verbose"]:
            logger.info(
                f"Detected {len(self.spikeDict)} spikes in {round(stopTime-startTime,3)} seconds"
            )

    def _spikeDetect2(self, sweepNumber: int):
        """Detect all spikes in one sweep.

         Populate bAnalysisResult.py.

        Notes
        -----
        First spike in a sweep cannot have interval statistics like freq or isi

        Parameters
        ----------
        sweepNumber : int
        """
        dDict = self._detectionDict

        # a list of dict of sanpy.bAnalysisResults.analysisResult (one dict per spike)
        spikeDict = sanpy.bAnalysisResults.analysisResultList()

        verbose = dDict["verbose"]

        #
        self.fileLoader.setSweep(sweepNumber)
        #

        # in case dDict has new filter values
        self._getFilteredRecording()

        #
        # spike detect
        detectionType = dDict["detectionType"]

        # detect all spikes either with dvdt or mv
        if detectionType == sanpy.bDetection.detectionTypes["mv"].value:
            # detect using mV threshold
            spikeTimes, spikeErrorList = self._spikeDetect_vm(dDict, sweepNumber)

            # TODO: get rid of this and replace with foot
            # backup childish vm threshold
            if dDict["doBackupSpikeVm"]:
                spikeTimes = self._backupSpikeVm(
                    spikeTimes, sweepNumber, dDict["medianFilter"]
                )
        elif detectionType == sanpy.bDetection.detectionTypes["dvdt"].value:
            # detect using dv/dt threshold AND min mV
            spikeTimes, spikeErrorList = self._spikeDetect_dvdt(dDict, sweepNumber)
        else:
            logger.error(f'Unknown detection type "{detectionType}"')
            return

        #
        # backup thrshold to zero crossing in dvdt
        if 0:
            tmp_window_ms = dDict["dvdtPreWindow_ms"]
            tmp_window_pnts = self.fileLoader.ms2Pnt_(tmp_window_ms)
            spikeTimes = self._getFeet(spikeTimes, tmp_window_pnts)

        #
        # set up
        sweepX = self.fileLoader.sweepX  # sweepNumber is not optional
        filteredVm = self.fileLoader.sweepY_filtered  # sweepNumber is not optional
        filteredDeriv = self.fileLoader.filteredDeriv
        # sweepC = self.fileLoader.sweepC

        #
        now = datetime.datetime.now()
        dateStr = now.strftime("%Y%m%d")
        timeStr = now.strftime("%H:%M:%S")
        self.dateAnalyzed = dateStr

        #
        # look in a window after each threshold crossing to get AP peak
        peakWindow_pnts = self.fileLoader.ms2Pnt_(dDict["peakWindow_ms"])

        #
        # look in a window after each peak to get 'fast ahp'
        fastAhpWindow_pnts = self.fileLoader.ms2Pnt_(dDict["fastAhpWindow_ms"])

        #
        # throw out spikes that have peak BELOW onlyPeaksAbove_mV
        # throw out spikes that have peak ABOVE onlyPeaksBelow_mV
        onlyPeaksAbove_mV = dDict["onlyPeaksAbove_mV"]
        onlyPeaksBelow_mV = dDict["onlyPeaksBelow_mV"]
        (
            spikeTimes,
            spikeErrorList,
            newSpikePeakPnt,
            newSpikePeakVal,
        ) = sanpy.analysisUtil.throwOutAboveBelow(
            filteredVm,
            spikeTimes,
            spikeErrorList,
            peakWindow_pnts,
            onlyPeaksAbove_mV=onlyPeaksAbove_mV,
            onlyPeaksBelow_mV=onlyPeaksBelow_mV,
        )

        #
        # small window to average Vm to calculate MDP (itself in a window before spike)
        avgWindow_pnts = self.fileLoader.ms2Pnt_(dDict["avgWindow_ms"])
        avgWindow_pnts = math.floor(avgWindow_pnts / 2)  # can be 0 !!!

        #
        # for each spike
        # numSpikes = len(spikeTimes)
        for i, spikeTime in enumerate(spikeTimes):
            # spikeTime units is ALWAYS points

            # new, add a spike dict for this spike time
            spikeDict.appendDefault()

            # get the AP peak
            peakPnt = newSpikePeakPnt[i]
            peakVal = newSpikePeakVal[i]
            peakSec = (newSpikePeakPnt[i] / self.fileLoader.dataPointsPerMs) / 1000

            # create one spike dictionary
            # spikeDict = OrderedDict() # use OrderedDict so Pandas output is in the correct order

            # spikeDict[i]['isBad'] = False
            spikeDict[i]["analysisDate"] = dateStr
            spikeDict[i]["analysisTime"] = timeStr
            spikeDict[i]["analysisVersion"] = sanpy.analysisVersion
            spikeDict[i]["interfaceVersion"] = sanpy.interfaceVersion
            spikeDict[i]["file"] = self.fileLoader.filename

            spikeDict[i]["detectionType"] = detectionType

            spikeDict[i]["cellType"] = dDict["cellType"]
            spikeDict[i]["sex"] = dDict["sex"]
            spikeDict[i]["condition"] = dDict["condition"]

            spikeDict[i]["sweep"] = sweepNumber

            epoch = float("nan")
            epochLevel = float("nan")
            epochTable = self.fileLoader.getEpochTable(sweepNumber)
            if epochTable is not None:
                epoch = epochTable.findEpoch(spikeTime)
                epochLevel = epochTable.getLevel(epoch)
            spikeDict[i]["epoch"] = epoch
            spikeDict[i]["epochLevel"] = epochLevel

            # keep track of per sweep spike and total spike
            spikeDict[i]["sweepSpikeNumber"] = i
            spikeDict[i]["spikeNumber"] = self.numSpikes + i

            spikeDict[i]["include"] = True

            # todo: make this a byte encoding so we can have multiple user tyes per spike
            spikeDict[i]["userType"] = 0  # One userType (int) that can have values

            # using bAnalysisResults will already be []
            spikeDict[i]["errors"] = []

            # append existing spikeErrorList from spikeDetect_dvdt() or spikeDetect_mv()
            tmpError = spikeErrorList[i]
            if tmpError is not None and tmpError != np.nan:
                spikeDict[i]["errors"].append(tmpError)  # tmpError is from:
                if verbose:
                    print(f"  spike:{i} error:{tmpError}")
            #
            # detection params
            spikeDict[i]["dvdtThreshold"] = dDict["dvdtThreshold"]
            spikeDict[i]["mvThreshold"] = dDict["mvThreshold"]
            spikeDict[i]["medianFilter"] = dDict["medianFilter"]
            spikeDict[i]["halfHeights"] = dDict["halfHeights"]

            spikeDict[i]["thresholdPnt"] = spikeTime
            spikeDict[i]["thresholdSec"] = (
                spikeTime / self.fileLoader.dataPointsPerMs
            ) / 1000
            spikeDict[i]["thresholdVal"] = filteredVm[spikeTime]  # in vm
            spikeDict[i]["thresholdVal_dvdt"] = filteredDeriv[
                spikeTime
            ]  # in dvdt, spikeTime is points

            # TODO: revamp this for 'Plot FI' plugin
            # spikeTime falls into wrong epoch for first fast spike
            # DAC command at the precise spike point
            # spikeDict[i]['dacCommand'] = sweepC[spikeTime]  # spikeTime is in points
            # spikeDict[i]['dacCommand'] = sweepC[peakPnt]  # spikeTime is in points

            spikeDict[i]["peakPnt"] = peakPnt
            spikeDict[i]["peakSec"] = peakSec
            spikeDict[i]["peakVal"] = peakVal

            spikeDict[i]["peakHeight"] = (
                spikeDict[i]["peakVal"] - spikeDict[i]["thresholdVal"]
            )

            tmpThresholdSec = spikeDict[i]["thresholdSec"]
            spikeDict[i]["timeToPeak_ms"] = (peakSec - tmpThresholdSec) * 1000

            # only append to spikeDict after we are done (accounting for spikes within a sweep)
            # self.spikeDict.append(spikeDict)
            # iIdx = len(self.spikeDict) - 1

            iIdx = i

            # fast ahp, fastAhpWindow_pnts
            if peakPnt+fastAhpWindow_pnts < len(sweepX):
                fastAhpClip = filteredVm[peakPnt : peakPnt+fastAhpWindow_pnts]
                fastAhpPnt = np.argmin(fastAhpClip)

                fastAhpError = fastAhpPnt == len(fastAhpClip)-1

                fastAhpPnt += peakPnt
                fastAhpSec = self.fileLoader.pnt2Sec_(fastAhpPnt)
                fastAhpValue = filteredVm[fastAhpPnt]

                spikeDict[i]["fastAhpPnt"] = fastAhpPnt
                spikeDict[i]["fastAhpSec"] = fastAhpSec
                spikeDict[i]["fastAhpValue"] = fastAhpValue

                # log error
                if fastAhpError:
                    errorType = "Fast AHP was detected at end of fast AHP window"
                    errorStr = "Consider increasing the fast AHP window with fastAhpWindow_ms"
                    eDict = self._getErrorDict(
                        i, spikeTimes[i], errorType, errorStr
                    )  # spikeTime is in pnts
                    spikeDict[iIdx]["errors"].append(eDict)

            # todo: get rid of this
            defaultVal = float("nan")

            """
            # get pre/post spike minima
            self.spikeDict[iIdx]['preMinPnt'] = None
            self.spikeDict[iIdx]['preMinVal'] = defaultVal

            # early diastolic duration
            # 0.1 to 0.5 of time between pre spike min and spike time
            self.spikeDict[iIdx]['preLinearFitPnt0'] = None
            self.spikeDict[iIdx]['preLinearFitPnt1'] = None
            self.spikeDict[iIdx]['earlyDiastolicDuration_ms'] = defaultVal # seconds between preLinearFitPnt0 and preLinearFitPnt1
            self.spikeDict[iIdx]['preLinearFitVal0'] = defaultVal
            self.spikeDict[iIdx]['preLinearFitVal1'] = defaultVal
            # m,b = np.polyfit(x, y, 1)
            self.spikeDict[iIdx]['earlyDiastolicDurationRate'] = defaultVal # fit of y=preLinearFitVal 0/1 versus x=preLinearFitPnt 0/1
            self.spikeDict[iIdx]['lateDiastolicDuration'] = defaultVal #

            self.spikeDict[iIdx]['preSpike_dvdt_max_pnt'] = None
            self.spikeDict[iIdx]['preSpike_dvdt_max_val'] = defaultVal # in units mV
            self.spikeDict[iIdx]['preSpike_dvdt_max_val2'] = defaultVal # in units dv/dt
            self.spikeDict[iIdx]['postSpike_dvdt_min_pnt'] = None
            self.spikeDict[iIdx]['postSpike_dvdt_min_val'] = defaultVal # in units mV
            self.spikeDict[iIdx]['postSpike_dvdt_min_val2'] = defaultVal # in units dv/dt

            self.spikeDict[iIdx]['isi_pnts'] = defaultVal # time between successive AP thresholds (thresholdSec)
            self.spikeDict[iIdx]['isi_ms'] = defaultVal # time between successive AP thresholds (thresholdSec)
            self.spikeDict[iIdx]['spikeFreq_hz'] = defaultVal # time between successive AP thresholds (thresholdSec)
            self.spikeDict[iIdx]['cycleLength_pnts'] = defaultVal # time between successive MDPs
            self.spikeDict[iIdx]['cycleLength_ms'] = defaultVal # time between successive MDPs

            # Action potential duration (APD) was defined as the interval between the TOP and the subsequent MDP
            #self.spikeDict[iIdx]['apDuration_ms'] = defaultVal
            self.spikeDict[iIdx]['diastolicDuration_ms'] = defaultVal

            # any number of spike widths
            #print('spikeDetect__() appending widths list to spike iIdx:', iIdx)
            # was this
            #self.spikeDict[iIdx]['widths'] = []
            # debug 20210929, self._getHalfWidth() will assign spikeDict[iIdx]['widths'] = []
            for halfHeight in dDict['halfHeights']:
                widthDict = {
                    'halfHeight': halfHeight,
                    'risingPnt': None,
                    'risingVal': defaultVal,
                    'fallingPnt': None,
                    'fallingVal': defaultVal,
                    'widthPnts': None,
                    'widthMs': defaultVal
                }
                # was this
                #spikeDict[iIdx]['widths_' + str(halfHeight)] = defaultVal
                spikeDict[iIdx]['widths'].append(widthDict)
            """

            #
            mdp_ms = dDict["mdp_ms"]
            mdp_pnts = self.fileLoader.ms2Pnt_(mdp_ms)  # mdp_ms * self.dataPointsPerMs
            mdp_pnts = int(mdp_pnts)

            # pre spike min
            # other algorithms look between spike[i-1] and spike[i]
            # here we are looking in a predefined window
            startPnt = spikeTimes[i] - mdp_pnts
            if startPnt < 0:
                # logger.info('TODO: add an official warning, we went past 0 for pre spike mdp ms window')
                startPnt = 0
                # log error
                errorType = "Pre spike min under-run (mdp)"
                errorStr = "Went past startPnt 0 searching for pre-spike min"
                eDict = self._getErrorDict(
                    i, spikeTimes[i], errorType, errorStr
                )  # spikeTime is in pnts
                spikeDict[iIdx]["errors"].append(eDict)
                if verbose:
                    logger.error(f"  spike:{iIdx} error:{eDict}")
                    logger.error(f'  going to index vmFiltered {startPnt} to spike time i {i} with value {spikeTimes[i]}')

            preRange = filteredVm[startPnt : spikeTimes[i]]  # EXCEPTION
            try:
                preMinPnt = np.argmin(preRange)
            except ValueError as e:
                # 20220926, happend when we have no scale and mdp_pnts=0
                # print(f'xxx i:{i} mdp_pnts:{mdp_pnts} len:{len(filteredVm)} startPnt:{startPnt} spikeTimes[i]:{spikeTimes[i]}')
                # 20220926, we really just want ot bail on this error
                # lots of code below relies on this
                # TODO: fix this mess
                preMinPnt = startPnt
                errorType = "Pre spike min 0 (mdp)"
                errorStr = f"Did not find preMinPnt mdp_pnts:{mdp_pnts} startPnt:{startPnt} spikeTimes[i]:{spikeTimes[i]}"
                eDict = self._getErrorDict(
                    i, spikeTimes[i], errorType, errorStr
                )  # spikeTime is in pnts
                spikeDict[iIdx]["errors"].append(eDict)
                if verbose:
                    print(f"  spike:{iIdx} error:{eDict}")
            if preMinPnt is not None:
                # 20230924, avgMinPnts is coming up zero now that we have sampling dt for kymographs that are slow !!!
                if avgWindow_pnts < 1:
                    # error
                    errorType = "mdp error"
                    errorStr = "avgWindow_pnts"
                    eDict = self._getErrorDict(
                        i, spikeTimes[i], errorType, errorStr
                    )  # spikeTime is in pnts
                    spikeDict[iIdx]["errors"].append(eDict)
                    if verbose:
                        print(f"  spike:{iIdx} error:{eDict}")

                else:
                    preMinPnt += startPnt
                    # the pre min is actually an average around the real minima
                    avgRange = filteredVm[
                        preMinPnt - avgWindow_pnts : preMinPnt + avgWindow_pnts
                    ]
                    # print('  avgRange is:' , avgRange)
                    preMinVal = np.average(avgRange)

                    # search backward from spike to find when vm reaches preMinVal (avg)
                    preRange = filteredVm[preMinPnt : spikeTimes[i]]
                    preRange = np.flip(preRange)  # we want to search backwards from peak
                    try:
                        preMinPnt2 = np.where(preRange < preMinVal)[0][0]
                        preMinPnt = spikeTimes[i] - preMinPnt2
                        spikeDict[iIdx]["preMinPnt"] = preMinPnt
                        spikeDict[iIdx]["preMinVal"] = preMinVal

                    except IndexError as e:
                        errorType = "Pre spike min (mdp)"
                        errorStr = "Did not find preMinVal: " + str(
                            round(preMinVal, 3)
                        )  # + ' postRange min:' + str(np.min(postRange)) + ' max ' + str(np.max(postRange))
                        eDict = self._getErrorDict(
                            i, spikeTimes[i], errorType, errorStr
                        )  # spikeTime is in pnts
                        spikeDict[iIdx]["errors"].append(eDict)
                        if verbose:
                            print(f"  spike:{iIdx} error:{eDict}")

            #
            # The nonlinear late diastolic depolarization phase was
            # estimated as the duration between 1% and 10% dV/dt
            # linear fit on 10% - 50% of the time from preMinPnt to self.spikeTimes[i]
            startLinearFit = 0.1  # percent of time between pre spike min and AP peak
            stopLinearFit = 0.5  #
            timeInterval_pnts = spikeTimes[i] - preMinPnt
            # taking round() so we always get an integer # points
            preLinearFitPnt0 = preMinPnt + round(timeInterval_pnts * startLinearFit)
            preLinearFitPnt1 = preMinPnt + round(timeInterval_pnts * stopLinearFit)
            preLinearFitVal0 = filteredVm[preLinearFitPnt0]
            preLinearFitVal1 = filteredVm[preLinearFitPnt1]

            # linear fit before spike
            spikeDict[iIdx]["preLinearFitPnt0"] = preLinearFitPnt0
            spikeDict[iIdx]["preLinearFitPnt1"] = preLinearFitPnt1
            spikeDict[iIdx]["earlyDiastolicDuration_ms"] = self.fileLoader.pnt2Ms_(
                preLinearFitPnt1 - preLinearFitPnt0
            )
            spikeDict[iIdx]["preLinearFitVal0"] = preLinearFitVal0
            spikeDict[iIdx]["preLinearFitVal1"] = preLinearFitVal1

            # a linear fit where 'm,b = np.polyfit(x, y, 1)'
            # m*x+b"
            xFit = sweepX[preLinearFitPnt0:preLinearFitPnt1]  # abb added +1
            yFit = filteredVm[preLinearFitPnt0:preLinearFitPnt1]

            # sometimes xFit/yFit have 0 length -->> TypeError
            # print(f' {iIdx} preLinearFitPnt0:{preLinearFitPnt0}, preLinearFitPnt1:{preLinearFitPnt1}')
            # print(f'    xFit:{len(xFit)} yFit:{len(yFit)}')

            # TODO: somehow trigger following errors to confirm code works (pytest)
            with warnings.catch_warnings():
                warnings.filterwarnings("error")
                try:
                    mLinear, bLinear = np.polyfit(
                        xFit, yFit, 1
                    )  # m is slope, b is intercept
                    spikeDict[iIdx]["earlyDiastolicDurationRate"] = mLinear
                    # todo: make an error if edd rate is too low
                    lowestEddRate = dDict["lowEddRate_warning"]  # 8
                    if mLinear <= lowestEddRate:
                        errorType = "Fit EDD"
                        errorStr = f"Early diastolic duration rate fit - Too low {round(mLinear,3)}<={lowestEddRate}"
                        eDict = self._getErrorDict(
                            i, spikeTimes[i], errorType, errorStr
                        )  # spikeTime is in pnts
                        # print('fit edd start num error:', 'iIdx:', iIdx, 'num error:', len(spikeDict[iIdx]['errors']))
                        spikeDict[iIdx]["errors"].append(eDict)
                        # print('  after num error:', len(spikeDict[iIdx]['errors']))
                        if verbose:
                            print(f"  spike:{iIdx} error:{eDict}")

                except (TypeError, RuntimeWarning) as e:
                    # catching exception:  expected non-empty vector for x
                    # xFit/yFit turn up empty when mdp and TOP points are within 1 point
                    spikeDict[iIdx]["earlyDiastolicDurationRate"] = defaultVal
                    errorType = "Fit EDD"
                    # errorStr = 'Early diastolic duration rate fit - TypeError'
                    errorStr = (
                        "Early diastolic duration rate fit - preMinPnt == spikePnt"
                    )
                    eDict = self._getErrorDict(i, spikeTimes[i], errorType, errorStr)
                    spikeDict[iIdx]["errors"].append(eDict)
                    if verbose:
                        print(f"  spike:{iIdx} error:{eDict}")
                except np.RankWarning as e:
                    # logger.error('== FIX preLinearFitPnt0/preLinearFitPnt1 RankWarning')
                    # logger.error(f'  error is: {e}')
                    # print('RankWarning')
                    # also throws: RankWarning: Polyfit may be poorly conditioned
                    spikeDict[iIdx]["earlyDiastolicDurationRate"] = defaultVal
                    errorType = "Fit EDD"
                    errorStr = "Early diastolic duration rate fit - RankWarning"
                    eDict = self._getErrorDict(i, spikeTimes[i], errorType, errorStr)
                    spikeDict[iIdx]["errors"].append(eDict)
                    if verbose:
                        print(f"  spike:{iIdx} error:{eDict}")
                # 20230422, don't ever catch an unknown exception
                # except:
                #     logger.error(
                #         f" !!!!!!!!!!!!!!!!!!!!!!!!!!! UNKNOWN EXCEPTION DURING EDD LINEAR FIT for spike {i}"
                #     )
                #     spikeDict[iIdx]["earlyDiastolicDurationRate"] = defaultVal
                #     errorType = "Fit EDD"
                #     errorStr = "Early diastolic duration rate fit - Unknown Exception"
                #     eDict = self._getErrorDict(i, spikeTimes[i], errorType, errorStr)
                #     if verbose:
                #         print(f"  spike:{iIdx} error:{eDict}")

            # not implemented
            # self.spikeDict[i]['lateDiastolicDuration'] = ???

            #
            # maxima in dv/dt before spike (between TOP and peak)
            try:
                preRange = filteredDeriv[spikeTimes[i] : peakPnt + 1]
                preSpike_dvdt_max_pnt = np.argmax(preRange)
                preSpike_dvdt_max_pnt += spikeTimes[i]
                spikeDict[iIdx]["preSpike_dvdt_max_pnt"] = preSpike_dvdt_max_pnt
                spikeDict[iIdx]["preSpike_dvdt_max_val"] = filteredVm[
                    preSpike_dvdt_max_pnt
                ]  # in units mV
                spikeDict[iIdx]["preSpike_dvdt_max_val2"] = filteredDeriv[
                    preSpike_dvdt_max_pnt
                ]  # in units mV
            except ValueError as e:
                # sometimes preRange is empty, don't try and put min/max in error
                errorType = "Pre Spike dvdt"
                errorStr = "Searching for dvdt max - ValueError"
                eDict = self._getErrorDict(
                    i, spikeTimes[i], errorType, errorStr
                )  # spikeTime is in pnts
                spikeDict[iIdx]["errors"].append(eDict)
                if verbose:
                    print(f"  spike:{iIdx} error:{eDict}")

            #
            # minima in dv/dt after spike
            # postRange = dvdt[self.spikeTimes[i]:postMinPnt]
            # postSpike_ms = 20 # 10
            # postSpike_pnts = self.ms2Pnt_(postSpike_ms)
            dvdtPostWindow_ms = dDict["dvdtPostWindow_ms"]
            dvdtPostWindow_pnts = self.fileLoader.ms2Pnt_(dvdtPostWindow_ms)
            postRange = filteredDeriv[
                peakPnt : peakPnt + dvdtPostWindow_pnts
            ]  # fixed window after spike

            postSpike_dvdt_min_pnt = np.argmin(postRange)
            postSpike_dvdt_min_pnt += peakPnt
            spikeDict[iIdx]["postSpike_dvdt_min_pnt"] = postSpike_dvdt_min_pnt
            spikeDict[iIdx]["postSpike_dvdt_min_val"] = filteredVm[
                postSpike_dvdt_min_pnt
            ]
            spikeDict[iIdx]["postSpike_dvdt_min_val2"] = filteredDeriv[
                postSpike_dvdt_min_pnt
            ]

            #
            # diastolic duration was defined as the interval between MDP and TOP
            # one off error when preMinPnt is not defined
            spikeDict[iIdx]["diastolicDuration_ms"] = self.fileLoader.pnt2Ms_(
                spikeTime - preMinPnt
            )

            #
            # calculate instantaneous spike frequency and ISI, for first spike this is not defined
            spikeDict[iIdx]["cycleLength_ms"] = float("nan")
            if iIdx > 0:
                isiPnts = (
                    spikeDict[iIdx]["thresholdPnt"]
                    - spikeDict[iIdx - 1]["thresholdPnt"]
                )
                isi_ms = self.fileLoader.pnt2Ms_(isiPnts)
                isi_hz = 1 / (isi_ms / 1000)
                spikeDict[iIdx]["isi_pnts"] = isiPnts
                spikeDict[iIdx]["isi_ms"] = self.fileLoader.pnt2Ms_(isiPnts)
                spikeDict[iIdx]["spikeFreq_hz"] = 1 / (
                    self.fileLoader.pnt2Ms_(isiPnts) / 1000
                )

                # Cycle length was defined as the interval between MDPs in successive APs
                prevPreMinPnt = spikeDict[iIdx - 1]["preMinPnt"]  # can be nan
                thisPreMinPnt = spikeDict[iIdx]["preMinPnt"]
                if prevPreMinPnt is not None and thisPreMinPnt is not None:
                    cycleLength_pnts = thisPreMinPnt - prevPreMinPnt
                    spikeDict[iIdx]["cycleLength_pnts"] = cycleLength_pnts
                    spikeDict[iIdx]["cycleLength_ms"] = self.fileLoader.pnt2Ms_(
                        cycleLength_pnts
                    )
                else:
                    # error
                    prevPreMinSec = self.fileLoader.pnt2Sec_(prevPreMinPnt)
                    thisPreMinSec = self.fileLoader.pnt2Sec_(thisPreMinPnt)
                    # errorStr = f'Previous spike preMinPnt is {prevPreMinPnt} and this preMinPnt: {thisPreMinPnt}'
                    errorType = "Cycle Length"
                    errorStr = f"Previous spike preMinPnt (s) is {prevPreMinSec} and this preMinPnt: {thisPreMinSec}"
                    eDict = self._getErrorDict(
                        i, spikeTimes[i], errorType, errorStr
                    )  # spikeTime is in pnts
                    spikeDict[iIdx]["errors"].append(eDict)
                    if verbose:
                        print(f"  spike:{iIdx} error:{eDict}")

            #
            # TODO: Move half-width to a function !!!
            #
            hwWindowPnts = dDict["halfWidthWindow_ms"] * self.fileLoader.dataPointsPerMs
            hwWindowPnts = round(hwWindowPnts)
            halfHeightList = dDict["halfHeights"]
            # was this
            # self._getHalfWidth(filteredVm, i, iIdx, spikeTime, peakPnt, hwWindowPnts, self.dataPointsPerMs, halfHeightList)
            self._getHalfWidth(
                filteredVm,
                iIdx,
                spikeDict,
                spikeTime,
                peakPnt,
                hwWindowPnts,
                self.fileLoader.dataPointsPerMs,
                halfHeightList,
                verbose=verbose,
            )

        #
        # look between threshold crossing to get minima
        # we will ignore the first and last spike

        #
        # spike clips
        self.spikeClips = None
        self.spikeClips_x = None
        self.spikeClips_x2 = None

        # SUPER important, previously our self.spikeDict was simple list of dict
        # now it is a list of class xxx
        # print('=== addind', len(spikeDict))
        self.spikeDict.appendAnalysis(spikeDict)
        # print('   now have', len(self.spikeDict))
        # print(self.spikeDict)

        # keep track of spikes per sweep (expensive to calculate)
        # self._spikesPerSweep[sweepNumber] = len(spikeDict)

        # run all user analysis ... what if this fails ???
        sanpy.user_analysis.baseUserAnalysis.runAllUserAnalysis(self)

        #
        # generate a df holding stats (used by scatterplotwidget)
        # startSeconds = dDict['startSeconds']
        # stopSeconds = dDict['stopSeconds']
        # if self.numSpikes > 0:
        #     # exportObject = sanpy.bExport(self)
        #     # self.dfReportForScatter = exportObject.report(startSeconds, stopSeconds)
        #     self._dfReportForScatter = self.spikeDict.asDataFrame()
        # else:
        #     self.dfReportForScatter = None
        self.regenerateAnalysisDataFrame()

        # generate error report
        self.dfError = self.getErrorReport()

        # bAnalysis needs to be saved
        self._detectionDirty = True

        ## done

    def regenerateAnalysisDataFrame(self):
        if self.numSpikes > 0:
            # exportObject = sanpy.bExport(self)
            # self.dfReportForScatter = exportObject.report(startSeconds, stopSeconds)
            self._dfReportForScatter = self.spikeDict.asDataFrame()

            # get rid of analysis results columns, we get these from file metadata
            #  - include
            #  - cellType
            #  - sex
            #  - condition
            self._dfReportForScatter = self._dfReportForScatter.drop('include', axis=1)
            self._dfReportForScatter = self._dfReportForScatter.drop('cellType', axis=1)
            self._dfReportForScatter = self._dfReportForScatter.drop('sex', axis=1)
            # 202401 removed
            # self._dfReportForScatter = self._dfReportForScatter.drop('condition', axis=1)

            # add all file meta data to df
            for k,v in self.metaData.items():
                # logger.info(f'   adding metadata {k} {v}')
                self._dfReportForScatter[k] = v

        else:
            self.dfReportForScatter = None

    def _getFeet(self, thresholdPnts: List[int], prePnts: int) -> List[int]:
        """

        Args:
            thresholdPnts (list of int)
            prePnts (int): pre point window to search for zero crossing

        Notes:
            Will need to calculate new (height, half widths)
        """

        # prePnts = int(prePnts)

        logger.info(f"num thresh:{len(thresholdPnts)} prePnts:{prePnts}")

        # df = self.asDataFrame()
        # peaks = df['peakVal']
        # thresholdPnts = df['thresholdPnt']

        verbose = self._detectionDict["verbose"]

        # using the derivstive to find zero crossing before
        # original full width left point
        # TODO: USer self.filteredDeriv
        # yFull = self.filteredVm
        # yDiffFull = np.diff(yFull)
        # yDiffFull = np.insert(yDiffFull, 0, np.nan)
        yDiffFull = self.fileLoader.filteredDeriv

        secondDeriv = np.diff(yDiffFull, axis=0)
        secondDeriv = np.insert(secondDeriv, 0, np.nan)

        n = len(thresholdPnts)
        footPntList = [None] * n
        footSec = [None] * n  # not used
        yFoot = [None] * n  # not used
        # myHeight = []

        # todo: add this to bAnalysis
        # preMs = self._detectionParams['preFootMs']
        # prePnts = self._sec2Pnt(preMs/1000)

        # TODO: add to bDetection
        logger.warning("ADD preMs AS PARAMETER !!!")
        # preWinMs = 50  # sa-node
        # prePnts = self.ms2Pnt_(preMs)

        for idx, footPnt in enumerate(thresholdPnts):
            # footPnt = round(footPnt)  # footPnt is in fractional points
            lastCrossingPnt = footPnt
            # move forwared a bit in case we are already in a local minima ???
            logger.warning("REMOVED WHEN WORKING ON NEURON DETECTION")
            footPnt += 2  # TODO: add as param
            preStart = footPnt - prePnts
            preClip = yDiffFull[preStart:footPnt]

            zero_crossings = np.where(np.diff(np.sign(preClip)))[
                0
            ]  # find where derivative flips sign (crosses 0)
            xLastCrossing = self.fileLoader.pnt2Sec_(footPnt)  # defaults
            yLastCrossing = self.fileLoader.sweepY_filtered[footPnt]
            if len(zero_crossings) == 0:
                if verbose:
                    tmpSec = round(self.fileLoader.pnt2Sec_(footPnt), 3)
                    logger.error(
                        f"  no foot for peak {idx} at sec {tmpSec} ... did not find zero crossings"
                    )
            else:
                # print(idx, 'footPnt:', footPnt, zero_crossings, preClip)
                lastCrossingPnt = preStart + zero_crossings[-1]
                xLastCrossing = self.fileLoader.pnt2Sec_(lastCrossingPnt)
                # get y-value (pA) from filtered. This removes 'pops' in raw data
                yLastCrossing = self.fileLoader.sweepY_filtered[lastCrossingPnt]

            # find peak in second derivative
            """
            preStart2 = lastCrossingPnt
            footMs2 = 20
            footPnt2 = preStart2 + self.ms2Pnt_(footMs2)
            preClip2 = secondDeriv[preStart2:footPnt2]
            #zero_crossings = np.where(np.diff(np.sign(preClip2)))[0]
            peakPnt2 = np.argmax(preClip2)
            peakPnt2 += preStart2

            #
            footPntList[idx] = peakPnt2
            """

            footPntList[idx] = lastCrossingPnt  # was this and worked, a bit too early

            footSec[idx] = xLastCrossing
            yFoot[idx] = yLastCrossing

            """
            peakPnt = df.loc[idx, 'peak_pnt']
            peakVal = self.sweepY_filtered[peakPnt]
            height = peakVal - yLastCrossing
            #print(f'idx {idx} {peakPnt} {peakVal} - {yLastCrossing} = {height}')
            myHeight[idx] = (height)
            """

        #
        # df =self._analysisList[self._analysisIdx]['results_full']
        """
        df['foot_pnt'] = footPntList  # sec
        df['foot_sec'] = footSec  # sec
        df['foot_val'] = yFoot  # pA
        """
        # df['myHeight'] = myHeight

        # return footPntList, footSec, yFoot
        return footPntList

    def printSpike(self, idx):
        """
        Print values in one spike analysis using self.spikeDict (sanpy.bAnalysisResults).
        """
        spike = self.spikeDict[idx]
        for k, v in spike.items():
            if k == "widths":
                widths = v
                print(f"  spike:{idx} has {len(widths)} widths...")
                for wIdx, width in enumerate(widths):
                    print(f"    spike:{idx} width:{wIdx}: {width}")
            elif k == "errors":
                errors = v
                print(f"  spike:{idx} has {len(errors)} errors...")
                for eIdx, error in enumerate(errors):
                    print(f"    spike:{idx} error #:{eIdx}: {error}")
            else:
                print(f"{k}: {v}")

    def printErrors(self):
        for idx, spike in enumerate(self.spikeDict):
            print(f"spike {idx} has {len(spike['errors'])} errors")
            for eIdx, error in enumerate(spike["errors"]):
                print(f"  error # {eIdx} is: {error}")

    def _makeSpikeClips(
        self,
        preSpikeClipWidth_ms,
        postSpikeClipWidth_ms=None,
        theseTime_sec=None,
        sweepNumber=None,
        epochNumber='All'
    ):
        """
        (Internal) Make small clips for each spike.

        Args:
            preSpikeClipWidth_ms (int): Width of each spike clip in milliseconds.
            postSpikeClipWidth_ms (int): Width of each spike clip in milliseconds.
            theseTime_sec (list of float): [NOT USED] List of seconds to make clips from.

        Returns:
            spikeClips_x2: ms
            self.spikeClips (list): List of spike clips
        """

        verbose = self._detectionDict["verbose"]

        if preSpikeClipWidth_ms is None:
            preSpikeClipWidth_ms = self._detectionDict["preSpikeClipWidth_ms"]
        if postSpikeClipWidth_ms is None:
            postSpikeClipWidth_ms = self._detectionDict["postSpikeClipWidth_ms"]

        if sweepNumber is None:
            sweepNumber = "All"

        # print('makeSpikeClips() spikeClipWidth_ms:', spikeClipWidth_ms, 'theseTime_sec:', theseTime_sec)
        if theseTime_sec is None:
            theseTime_pnts = self.getSpikeTimes(sweepNumber=sweepNumber, epochNumber=epochNumber)
        else:
            # convert theseTime_sec to pnts
            theseTime_ms = [x * 1000 for x in theseTime_sec]
            theseTime_pnts = [x * self.fileLoader.dataPointsPerMs for x in theseTime_ms]
            theseTime_pnts = [round(x) for x in theseTime_pnts]

        preClipWidth_pnts = self.fileLoader.ms2Pnt_(preSpikeClipWidth_ms)
        # if preClipWidth_pnts % 2 == 0:
        #    pass # Even
        # else:
        #    clipWidth_pnts += 1 # Make odd even
        postClipWidth_pnts = self.fileLoader.ms2Pnt_(postSpikeClipWidth_ms)

        # halfClipWidth_pnts = int(clipWidth_pnts/2)

        # print('  makeSpikeClips() clipWidth_pnts:', clipWidth_pnts, 'halfClipWidth_pnts:', halfClipWidth_pnts)
        # make one x axis clip with the threshold crossing at 0
        # was this, in ms
        # self.spikeClips_x = [(x-halfClipWidth_pnts)/self.dataPointsPerMs for x in range(clipWidth_pnts)]

        # in ms
        self.spikeClips_x = [
            (x - preClipWidth_pnts) / self.fileLoader.dataPointsPerMs
            for x in range(preClipWidth_pnts)
        ]
        self.spikeClips_x += [
            (x) / self.fileLoader.dataPointsPerMs for x in range(postClipWidth_pnts)
        ]

        # 20190714, added this to make all clips same length, much easier to plot in MultiLine
        numPointsInClip = len(self.spikeClips_x)

        self.spikeClips = []
        self.spikeClips_x2 = []

        sweepY = self.fileLoader.sweepY_filtered

        # when there are no spikes getStat() will not return anything
        # For 'All' sweeps, we need to know column
        sweepNum = self.getStat("sweep", sweepNumber=sweepNumber)

        # logger.info(f'sweepY: {sweepY.shape} {len(sweepY.shape)}')
        # logger.info(f'theseTime_pnts: {theseTime_pnts}')

        for idx, spikeTime in enumerate(theseTime_pnts):
            sweep = sweepNum[idx]

            if len(sweepY.shape) == 1:
                # 1D case where recording has only oone sweep
                # currentClip = sweepY[spikeTime-halfClipWidth_pnts:spikeTime+halfClipWidth_pnts]
                currentClip = sweepY[
                    spikeTime - preClipWidth_pnts : spikeTime + postClipWidth_pnts
                ]
            else:
                # 2D case where recording has multiple sweeps
                # currentClip = sweepY[spikeTime-halfClipWidth_pnts:spikeTime+halfClipWidth_pnts, sweep]
                try:
                    currentClip = sweepY[
                        spikeTime - preClipWidth_pnts : spikeTime + preClipWidth_pnts,
                        sweep,
                    ]
                except IndexError as e:
                    logger.error(e)
                    print(f"sweep: {sweep}")
                    print(f"sweepY.shape: {sweepY.shape}")

            if len(currentClip) == numPointsInClip:
                self.spikeClips.append(currentClip)
                self.spikeClips_x2.append(
                    self.spikeClips_x
                )  # a 2D version to make pyqtgraph multiline happy
            else:
                # pass
                if verbose:
                    logger.warning(
                        f"Did not add clip for spike index: {idx} at time: {spikeTime} len(currentClip): {len(currentClip)} != numPointsInClip: {numPointsInClip}"
                    )

        #
        return self.spikeClips_x2, self.spikeClips

    def getSpikeClips(
        self,
        theMin,
        theMax,
        spikeSelection=[],
        preSpikeClipWidth_ms=None,
        postSpikeClipWidth_ms=None,
        sweepNumber=None,
        epochNumber='All',
        ignoreMinMax=False  # added 20230418
    ):
        """Get 2d list of spike clips, spike clips x, and 1d mean spike clip.

        Args:
            theMin (float): Start seconds.
            theMax (float): Stop seconds.
            spikeSelection (list): List of spike numbers
            preSpikeClipWidth_ms (float):
            postSpikeClipWidth_ms (float):

        Requires: self.spikeDetect() and self._makeSpikeClips()

        Returns:
            theseClips (list): List of clip
            theseClips_x (list): ms
            meanClip (list)
        """

        if self.numSpikes == 0:
            return

        doSpikeSelection = len(spikeSelection) > 0

        if doSpikeSelection:
            pass
        elif theMin is None or theMax is None:
            theMin = 0
            theMax = self.fileLoader.recordingDur  # self.sweepX[-1]

        # new interface, spike detect no longer auto generates these
        # need to do this every time because we get here when sweepNumber changes
        # if self.spikeClips is None:
        #    self._makeSpikeClips(spikeClipWidth_ms=spikeClipWidth_ms, sweepNumber=sweepNumber)
        # TODO: don't make all clips
        # self._makeSpikeClips(spikeClipWidth_ms=spikeClipWidth_ms, sweepNumber=sweepNumber)
        self._makeSpikeClips(
            preSpikeClipWidth_ms=preSpikeClipWidth_ms,
            postSpikeClipWidth_ms=postSpikeClipWidth_ms,
            sweepNumber=sweepNumber,
            epochNumber=epochNumber
        )

        # make a list of clips within start/stop (Seconds)
        theseClips = []
        theseClips_x = []
        tmpMeanClips = []  # for mean clip
        meanClip = []

        # spikeTimes are in pnts
        spikeTimes = self.getSpikeTimes(sweepNumber=sweepNumber, epochNumber=epochNumber)

        logger.info(f'spikeTimes:{len(spikeTimes)} sweepNumber:{sweepNumber} epochNumber:{epochNumber}')

        # if len(spikeTimes) != len(self.spikeClips):
        #    logger.error(f'len spikeTimes {len(spikeTimes)} !=  spikeClips {len(self.spikeClips)}')

        # self.spikeClips is a list of clips
        for idx, clip in enumerate(self.spikeClips):
            doThisSpike = False
            if doSpikeSelection:
                doThisSpike = idx in spikeSelection
            else:
                spikeTime = spikeTimes[idx]
                spikeTime = self.fileLoader.pnt2Sec_(spikeTime)
                if ignoreMinMax or (spikeTime >= theMin and spikeTime <= theMax):
                    doThisSpike = True
            if doThisSpike:
                theseClips.append(clip)
                theseClips_x.append(
                    self.spikeClips_x2[idx]
                )  # remember, all _x are the same
                if len(self.spikeClips_x) == len(clip):
                    tmpMeanClips.append(clip)  # for mean clip
        if len(tmpMeanClips):
            meanClip = np.mean(tmpMeanClips, axis=0)

        return theseClips, theseClips_x, meanClip

    # def numErrors(self):
    #     if self.dfError is None:
    #         return "N/A"
    #     else:
    #         return len(self.dfError)

    def getErrorReport(self):
        """Generate an error report, one row per error.

        Spikes can have more than one error.

        Returns:
            (pandas DataFrame): Pandas DataFrame, one row per error.
        """

        dictList = []

        # numError = 0
        # errorList = []

        # logger.info(f'Generating error report for {len(self.spikeDict)} spikes')

        #  20230422 spikeDict is not working as an iterable
        # use it as a list instead
        numSpikes = len(self.spikeDict)
        #for spike in self.spikeDict:
        for _spikeNumber in range(numSpikes):
            spike = self.spikeDict[_spikeNumber]
            # spike is sanpy.bAnalysisResults.analysisResult
            #print('spike:', spike)
            for error in spike["errors"]:
                # spike["errors"] is a list of dict
                # error is dict from _getErrorDict
                if error is None or error == np.nan or error == "nan":
                    continue

                # 20230422 add sweep and epoch to error dict
                #_spikeNumber = error['Spike']

                #print('  _spikeNumber:', _spikeNumber, type(_spikeNumber))

                # _sweep = self.getSpikeStat([_spikeNumber], 'sweep')
                # if len(_sweep)==0:
                #     logger.error(f"_spikeNumber:{_spikeNumber} sweep:{_sweep}")
                #     #print(self.getOneSpikeDict(_spikeNumber))

                error['Sweep'] = self.getSpikeStat([_spikeNumber], 'sweep')[0]
                error['Epoch'] = self.getSpikeStat([_spikeNumber], 'epoch')[0]

                dictList.append(error)

        if len(dictList) == 0:
            fakeErrorDict = self._getErrorDict(1, 1, "fake", "fake")
            dfError = pd.DataFrame(columns=fakeErrorDict.keys())
        else:
            dfError = pd.DataFrame(dictList)

        if self._detectionDict["verbose"]:
            logger.info(f"Found {len(dfError)} errors in spike detection")

        return dfError

    def _old_to_csv(self):
        """Save as a CSV text file with name <path>_analysis.csv'"""
        savefile = os.path.splitext(self._path)[0]
        savefile += "_analysis.csv"
        saveExcel = False
        alsoSaveTxt = True
        logger.info(f'Saving "{savefile}"')

        be = sanpy.bExport(self)
        be.saveReport(savefile, saveExcel=saveExcel, alsoSaveTxt=alsoSaveTxt)

    def _old__normalizeData(self, data):
        """Calculate normalized data for detection from Kymograph. Is NOT for df/d0."""
        return (data - np.min(data)) / (np.max(data) - np.min(data))

    def _not_used_loadAnalysis(self):
        """Not used."""
        saveBase = self._getSaveBase()

        # load detection parameters
        # self.detectionClass.load(saveBase)

        # load analysis
        # self.spikeDict.load(saveBase)

        saveBase = self._getSaveBase()
        savePath = saveBase + "-analysis.json"

        if not os.path.isfile(savePath):
            # logger.error(f'Did not find file: {savePath}')
            return

        logger.info(f"Loading from saved analysis: {savePath}")

        with open(savePath, "r") as f:
            # self._dDict = json.load(f)
            loadedDict = json.load(f)

        dDict = loadedDict["detection"]
        self.detectionClass._dDict = dDict

        analysisList = loadedDict["analysis"]
        self.spikeDict._myList = analysisList

        self._detectionDirty = False
        self._isAnalyzed = True

    def saveAnalysis_tocsv(self, path : str = None, verbose=False):
        """Save analysis to csv.

        CSV starts with one 
        Parameters
        ----------
        path : str
            Full path of file to save, if None will save as default.
        """

        if path is None:
            saveFolder = self._getSaveFolder()
            if not os.path.isdir(saveFolder):
                if verbose:
                    logger.info(f"making folder: {saveFolder}")
                os.mkdir(saveFolder)

            saveBase = self._getSaveBase()
            path = saveBase + "-analysis.csv"

        if verbose:
            logger.info(f'saving to: {path}')

        metaDataHeader = self.metaData.getHeader()

        with open(path, "w") as f:
            f.write(metaDataHeader)
            f.write("\n")

        df = self.asDataFrame()  # pd.DataFrame(self.spikeDict)
        if df is not None:
            df.to_csv(path, mode="a")
        # else:
            # happens when user sets metaDat but does not do analysis
            # logger.warning(f'asDataFrame() returned None')
            # logger.warning(f'  did not save: {self}')

    def saveAnalysis(self, forceSave=False):
        """Not used.

        Save detection parameters and analysis results as json.
        """
        if not self._detectionDirty and not forceSave:
            return

        saveFolder = self._getSaveFolder()
        if not os.path.isdir(saveFolder):
            logger.info(f"making folder: {saveFolder}")
            os.mkdir(saveFolder)

        saveBase = self._getSaveBase()
        savePath = saveBase + "-analysis.json"

        # save detection parameters
        # self.detectionClass.save(saveBase)
        dDict = self.detectionClass.getDict()

        saveDict = {}
        saveDict["detection"] = dDict

        # save list of dict
        # self.spikeDict = sanpy.bAnalysisResults.analysisResultList()
        # self.spikeDict.save(saveBase)
        analysisList = self.spikeDict.asList()

        saveDict["analysis"] = analysisList

        with open(savePath, "w") as f:
            json.dump(saveDict, f, cls=NumpyEncoder, indent=4)

        self._detectionDirty = False

        logger.info(f"Saved analysis to: {savePath}")

    def _getSaveFolder(self):
        """
        All analysis will be saved in folder 'sanpy_analysis'
        """
        filepath = self.fileLoader.filepath
        parentPath, fileName = os.path.split(filepath)
        saveFolder = os.path.join(parentPath, "sanpy_analysis")
        return saveFolder

    def _getSaveBase(self):
        """Get basename to append to to save

        This will always be in a subfolder named 'sanpy_analysis'

        For example, bDetection uses this to save <base>-detection.json
        """
        saveFolder = self._getSaveFolder()

        filepath = self.fileLoader.filepath
        parentPath, fileName = os.path.split(filepath)
        baseName = os.path.splitext(fileName)[0]
        savePath = os.path.join(saveFolder, baseName)

        return savePath

    @property
    def analysisDate(self):
        if self.spikeDict is not None:
            return self.spikeDict.analysisDate()

    @property
    def analysisTime(self):
        if self.spikeDict is not None:
            return self.spikeDict.analysisTime()

    def _api_getHeader(self):
        """Get header as a dict.

        TODO:
            - add info on abf file, like samples per ms

        Returns:
            dict: Dictionary of information about loaded file.
        """
        # recordingDir_sec = len(self.sweepX) / self.dataPointsPerMs / 1000
        recordingFrequency = self.dataPointsPerMs

        ret = {
            "myFileType": self.myFileType,  # ('abf', 'tif', 'bytestream', 'csv')
            "loadError": self.loadError,
            #'detectionDict': self.detectionClass,
            "path": self._path,
            "file": self.fileLoader.filename,
            "dateAnalyzed": self.dateAnalyzed,
            #'detectionType': self.detectionType,
            "acqDate": self.acqDate,
            "acqTime": self.acqTime,
            #
            "_recordingMode": self._recordingMode,
            "get_yUnits": self.get_yUnits(),
            #'currentSweep': self.currentSweep,
            "recording_kHz": recordingFrequency,
            "recordingDur_sec": self.recordingDur,
        }
        return ret

    def _api_getSpikeInfo(self, spikeNum=None):
        """Get info about each spike.

        Args:
            spikeNum (int): Get info for one spike, None for all spikes.

        Returns:
            list: List of dict with info for all (one) spike.
        """
        if spikeNum is not None:
            ret = [self.spikeDict[spikeNum]]
        else:
            ret = self.spikeDict
        return ret

    def _api_getSpikeStat(self, stat):
        """Get stat for each spike

        Args:
            stat (str): The name of the stat to get. Corresponds to key in self.spikeDict[i].

        Returns:
            list: List of values for 'stat'. Ech value is for one spike.
        """
        statList = self.getStat(statName1=stat, statName2=None)
        return statList

    def _api_getRecording(self):
        """Return primary recording

        Returns:
            dict: {'header', 'sweepX', 'sweepY'}

        TODO:
            Add param to only get every n'th point, to return a subset faster (for display)
        """
        # start = time.time()
        ret = {
            "header": self.api_getHeader(),
            "sweepX": self.sweepX2.tolist(),
            "sweepY": self.sweepY2.tolist(),
        }
        # stop = time.time()
        # print(stop-start)
        return ret

Attributes

dateAnalyzed: str = None instance-attribute

str: Date Time of analysis. TODO: make a property.

fileLoader: sanpy.fileLoader_base property

kymAnalysis property

Get the kymAnalysis object (if it exists).

loadError: bool = False instance-attribute

bool: True if error loading file/stream.

numErrors: int property

Get number of detection errors.

numSpikes property

Get the total number of detected spikes (all sweeps).

See getNumSpikes(sweep)

Functions

__init__(filepath=None, byteStream=None, loadData=True, fileLoaderDict=None, stimulusFileFolder=None, verbose=False)

Args: filepath (str): Path to either .abf or .csv with time/mV columns. byteStream (io.BytesIO): Binary stream for use in the cloud. loadData: If true, load raw data, otherwise just load header fileLoaderDict (dict) If None then fetch from sanpy.fileloaders.getFileLoaders() Do this if running in a script. If running an SanPy app, we pass the dict stimulusFileFolder:

Source code in sanpy/bAnalysis_.py

def __init__(
    self,
    filepath: str = None,
    byteStream=None,
    loadData: bool = True,
    fileLoaderDict: dict = None,
    stimulusFileFolder: str = None,
    verbose: bool = False,
):
    """
    Args:
        filepath (str): Path to either .abf or .csv with time/mV columns.
        byteStream (io.BytesIO): Binary stream for use in the cloud.
        loadData: If true, load raw data, otherwise just load header
        fileLoaderDict (dict)
            If None then fetch from sanpy.fileloaders.getFileLoaders()
            Do this if running in a script.
            If running an SanPy app, we pass the dict
        stimulusFileFolder:
    """

    """
    self._path = file  # todo: change this to filePath
    """
    """str: File path."""

    self._detectionDict: dict = None  # corresponds to an item in sanpy.bDetection

    # sept 9, moving this to file loader
    # fileloader holds meta data
    #self._metaData = MetaData(self)  #self.getMetaDataDict()

    self._isAnalyzed: bool = False

    self.loadError: bool = False
    """bool: True if error loading file/stream."""

    # self.detectionDict = None  # remember the parameters of our last detection
    """dict: Dictionary specifying detection parameters, see bDetection.getDefaultDetection."""

    # self._abf = None
    """pyAbf: If loaded from binary .abf file"""

    self.dateAnalyzed: str = None
    """str: Date Time of analysis. TODO: make a property."""

    # self.detectionType = None
    """str: From ('dvdt', 'mv')"""

    self.spikeDict: sanpy.bAnalysisResults.analysisResultList = (
        sanpy.bAnalysisResults.analysisResultList()
    )
    # class to store all analysis results

    # self._spikesPerSweep : int = None

    self.spikeClips = []  # created in self.spikeDetect()
    self.spikeClips_x = []  #
    self.spikeClips_x2 = []  #

    self.dfError = None  # dataframe with a list of detection errors
    self._dfReportForScatter = None  # dataframe to be used by scatterplotwidget

    self._detectionDirty = False

    # will be overwritten by existing uuid in self._loadFromDf()
    self.uuid = sanpy._util.getNewUuid()

    # self.tifData = None
    # when we have a tif kymograph

    # self.isBytesIO = False
    # when we are running in the cloud

    # TODO (cudmore) need to parse folder of file loaders in fileloders/ and determine
    # class to use to load file (using fileLoader.filetype
    self._fileLoader = None
    if filepath is not None and not os.path.isfile(filepath):
        logger.error(f'File does not exist: "{filepath}"')
        self.loadError = True
    else:

        if fileLoaderDict is None:
            fileLoaderDict = (
                sanpy.fileloaders.getFileLoaders()
            )  # EXPENSIVE, to do, pass in from app
        # print('2 fileLoaderDict:', fileLoaderDict)

        # print('1 fileLoaderDict:')
        # for _k,_v in fileLoaderDict.items():
        #     print(_k,_v)

        _ext = os.path.splitext(filepath)[1]
        # _ext = _ext[1:]
        try:
            if verbose:
                logger.info(f"Loading file with extension: {_ext}")
            constructorObject = fileLoaderDict[_ext]["constructor"]
            self._fileLoader = constructorObject(filepath)
            # may 2, 2023
            if self._fileLoader._loadError:
                logger.error(f'load error in file loader for ext: "{_ext}"')
                self.loadError = True

        except KeyError as e:
            logger.error(f'did not find a file loader for extension "{_ext}", available loaders are: {fileLoaderDict.keys()}')
            self.loadError = True

        self._kymAnalysis : sanpy.kymAnalysis = None
        if (self.fileLoader is not None) and (self.fileLoader.recordingMode == recordingModes.kymograph):
            if verbose:
                logger.info('creating kymAnalysis')
                logger.info(f'    self.fileLoader.filepath:{self.fileLoader.filepath}')
                logger.info(f'    self.fileLoader.tifData:{self.fileLoader.tifData.shape}')
                logger.info(f'    self.fileLoader.tifHeader:{self.fileLoader.tifHeader}')
            self._kymAnalysis = sanpy.kymAnalysis(self.fileLoader.filepath,
                                                  self.fileLoader.tifData,
                                                  self.fileLoader.tifHeader)

        if self._fileLoader is not None:
            # we need to so file loader meta data can set ba (Self) dirty when changed
            self.fileLoader.metadata._ba = self


    """
    if byteStream is not None:
        self._loadAbf(byteStream=byteStream,
                loadData=loadData,
                stimulusFileFolder=stimulusFileFolder)
    elif file is not None and file.endswith('.abf'):
        self._loadAbf(loadData=loadData)
    elif file is not None and file.endswith('.atf'):
        self._loadAtf(loadData=loadData)
    elif file is not None and file.endswith('.tif'):
        self._loadTif()
    elif file is not None and file.endswith('.csv'):
        self._loadCsv()
    else:
        pass
        #logger.error(f'Can only open abf/csv/tif/stream files: {file}')
        #self.loadError = True
    """

    # get default derivative
    if loadData and not self.loadError:
        self._rebuildFiltered()

    self._detectionDirty = False

    """
    self.setSweep()
    """

__str__()

Get a brief str representation. Usefull for print().

Source code in sanpy/bAnalysis_.py

def __str__(self):
    """Get a brief str representation. Usefull for print()."""
    # if self.isBytesIO:
    #      filename = '<BytesIO>'
    # else:
    #     filename = self.getFileName()
    fileLoadStr = self.fileLoader.__str__()
    txt = f"fileLoader: {fileLoadStr} spikes:{self.numSpikes}"
    return txt

asDataFrame(regenerateAnalysisDataFrame=False)

Return analysis as a Pandas DataFrame.

Important: This is a df copy of our self.spikeDict Do not modify and expect changes to stick

Source code in sanpy/bAnalysis_.py

def asDataFrame(self, regenerateAnalysisDataFrame=False):
    """Return analysis as a Pandas DataFrame.

    Important:
        This is a df copy of our self.spikeDict
        Do not modify and expect changes to stick
    """

    # re-add file metadata, it may have changed
    if regenerateAnalysisDataFrame:
        self.regenerateAnalysisDataFrame()

    return self._dfReportForScatter

getDetectionDict(asCopy=False)

Get the detection dictionary that was used for detect().

Source code in sanpy/bAnalysis_.py

def getDetectionDict(self, asCopy: bool = False):
    """Get the detection dictionary that was used for detect()."""
    if asCopy:
        return copy.deepcopy(self._detectionDict)
    else:
        return self._detectionDict

getErrorReport()

Generate an error report, one row per error.

Spikes can have more than one error.

Returns: (pandas DataFrame): Pandas DataFrame, one row per error.

Source code in sanpy/bAnalysis_.py

def getErrorReport(self):
    """Generate an error report, one row per error.

    Spikes can have more than one error.

    Returns:
        (pandas DataFrame): Pandas DataFrame, one row per error.
    """

    dictList = []

    # numError = 0
    # errorList = []

    # logger.info(f'Generating error report for {len(self.spikeDict)} spikes')

    #  20230422 spikeDict is not working as an iterable
    # use it as a list instead
    numSpikes = len(self.spikeDict)
    #for spike in self.spikeDict:
    for _spikeNumber in range(numSpikes):
        spike = self.spikeDict[_spikeNumber]
        # spike is sanpy.bAnalysisResults.analysisResult
        #print('spike:', spike)
        for error in spike["errors"]:
            # spike["errors"] is a list of dict
            # error is dict from _getErrorDict
            if error is None or error == np.nan or error == "nan":
                continue

            # 20230422 add sweep and epoch to error dict
            #_spikeNumber = error['Spike']

            #print('  _spikeNumber:', _spikeNumber, type(_spikeNumber))

            # _sweep = self.getSpikeStat([_spikeNumber], 'sweep')
            # if len(_sweep)==0:
            #     logger.error(f"_spikeNumber:{_spikeNumber} sweep:{_sweep}")
            #     #print(self.getOneSpikeDict(_spikeNumber))

            error['Sweep'] = self.getSpikeStat([_spikeNumber], 'sweep')[0]
            error['Epoch'] = self.getSpikeStat([_spikeNumber], 'epoch')[0]

            dictList.append(error)

    if len(dictList) == 0:
        fakeErrorDict = self._getErrorDict(1, 1, "fake", "fake")
        dfError = pd.DataFrame(columns=fakeErrorDict.keys())
    else:
        dfError = pd.DataFrame(dictList)

    if self._detectionDict["verbose"]:
        logger.info(f"Found {len(dfError)} errors in spike detection")

    return dfError

getNumSpikes(sweep=0)

Get number of spikes in a sweep.

See property numSpikes

Source code in sanpy/bAnalysis_.py

def getNumSpikes(self, sweep: int = 0):
    """Get number of spikes in a sweep.

    See property numSpikes
    """
    thresholdSec = self.getStat("thresholdSec", sweepNumber=sweep)
    return len(thresholdSec)

getSpikeClips(theMin, theMax, spikeSelection=[], preSpikeClipWidth_ms=None, postSpikeClipWidth_ms=None, sweepNumber=None, epochNumber='All', ignoreMinMax=False)

Get 2d list of spike clips, spike clips x, and 1d mean spike clip.

Args: theMin (float): Start seconds. theMax (float): Stop seconds. spikeSelection (list): List of spike numbers preSpikeClipWidth_ms (float): postSpikeClipWidth_ms (float):

Requires: self.spikeDetect() and self._makeSpikeClips()

Returns: theseClips (list): List of clip theseClips_x (list): ms meanClip (list)

Source code in sanpy/bAnalysis_.py

def getSpikeClips(
    self,
    theMin,
    theMax,
    spikeSelection=[],
    preSpikeClipWidth_ms=None,
    postSpikeClipWidth_ms=None,
    sweepNumber=None,
    epochNumber='All',
    ignoreMinMax=False  # added 20230418
):
    """Get 2d list of spike clips, spike clips x, and 1d mean spike clip.

    Args:
        theMin (float): Start seconds.
        theMax (float): Stop seconds.
        spikeSelection (list): List of spike numbers
        preSpikeClipWidth_ms (float):
        postSpikeClipWidth_ms (float):

    Requires: self.spikeDetect() and self._makeSpikeClips()

    Returns:
        theseClips (list): List of clip
        theseClips_x (list): ms
        meanClip (list)
    """

    if self.numSpikes == 0:
        return

    doSpikeSelection = len(spikeSelection) > 0

    if doSpikeSelection:
        pass
    elif theMin is None or theMax is None:
        theMin = 0
        theMax = self.fileLoader.recordingDur  # self.sweepX[-1]

    # new interface, spike detect no longer auto generates these
    # need to do this every time because we get here when sweepNumber changes
    # if self.spikeClips is None:
    #    self._makeSpikeClips(spikeClipWidth_ms=spikeClipWidth_ms, sweepNumber=sweepNumber)
    # TODO: don't make all clips
    # self._makeSpikeClips(spikeClipWidth_ms=spikeClipWidth_ms, sweepNumber=sweepNumber)
    self._makeSpikeClips(
        preSpikeClipWidth_ms=preSpikeClipWidth_ms,
        postSpikeClipWidth_ms=postSpikeClipWidth_ms,
        sweepNumber=sweepNumber,
        epochNumber=epochNumber
    )

    # make a list of clips within start/stop (Seconds)
    theseClips = []
    theseClips_x = []
    tmpMeanClips = []  # for mean clip
    meanClip = []

    # spikeTimes are in pnts
    spikeTimes = self.getSpikeTimes(sweepNumber=sweepNumber, epochNumber=epochNumber)

    logger.info(f'spikeTimes:{len(spikeTimes)} sweepNumber:{sweepNumber} epochNumber:{epochNumber}')

    # if len(spikeTimes) != len(self.spikeClips):
    #    logger.error(f'len spikeTimes {len(spikeTimes)} !=  spikeClips {len(self.spikeClips)}')

    # self.spikeClips is a list of clips
    for idx, clip in enumerate(self.spikeClips):
        doThisSpike = False
        if doSpikeSelection:
            doThisSpike = idx in spikeSelection
        else:
            spikeTime = spikeTimes[idx]
            spikeTime = self.fileLoader.pnt2Sec_(spikeTime)
            if ignoreMinMax or (spikeTime >= theMin and spikeTime <= theMax):
                doThisSpike = True
        if doThisSpike:
            theseClips.append(clip)
            theseClips_x.append(
                self.spikeClips_x2[idx]
            )  # remember, all _x are the same
            if len(self.spikeClips_x) == len(clip):
                tmpMeanClips.append(clip)  # for mean clip
    if len(tmpMeanClips):
        meanClip = np.mean(tmpMeanClips, axis=0)

    return theseClips, theseClips_x, meanClip

getSpikeDictionaries(sweepNumber=None)

Get spike dictionaries for current sweep

Source code in sanpy/bAnalysis_.py

def getSpikeDictionaries(self, sweepNumber=None):
    """Get spike dictionaries for current sweep
    """
    if sweepNumber is None:
        sweepNumber = "All"
    # logger.info(f'sweepNumber:{sweepNumber}')
    theRet = [
        spike
        for spike in self.spikeDict
        if sweepNumber == "All" or spike["sweep"] == sweepNumber
    ]
    return theRet

getSpikeSeconds(sweepNumber=None)

Get spike times (seconds) for current sweep

Source code in sanpy/bAnalysis_.py

def getSpikeSeconds(self, sweepNumber=None):
    """Get spike times (seconds) for current sweep"""
    # theRet = [spike['thresholdSec'] for spike in self.spikeDict if spike['sweep']==self.currentSweep]
    theRet = self.getStat("thresholdSec", sweepNumber=sweepNumber)
    return theRet

getSpikeStat(spikeList, stat)

Get one stat from a list of spikes

Parameters:

Name	Type	Description	Default
spikeList	List[int]		required
stat	str		required

Source code in sanpy/bAnalysis_.py

def getSpikeStat(self, spikeList : List[int], stat : str):
    """Get one stat from a list of spikes

    Parameters
    ----------
    spikeList : List[int]
    stat : str
    """

    # if isinstance(spikeList, int):
    #     spikeList = [spikeList]

    if len(spikeList) == 0:
        return None

    # logger.info(f'spikeList: {spikeList} stat:{stat}')

    retList = []
    # count = 0
    for idx, spike in enumerate(self.spikeDict):
        # logger.info(f'  idx:{idx}')
        if idx in spikeList:
            try:
                val = spike[stat]
                retList.append(val)
                # count += 1
            except KeyError as e:
                logger.error(e)
    # logger.info(f'  retList: {retList}')
    return retList

getSpikeTimes(sweepNumber=None, epochNumber='All')

Get spike times (points) for current sweep

Source code in sanpy/bAnalysis_.py

def getSpikeTimes(self, sweepNumber=None, epochNumber='All'):
    """Get spike times (points) for current sweep"""
    # theRet = [spike['thresholdPnt'] for spike in self.spikeDict if spike['sweep']==self.currentSweep]
    theRet = self.getStat("thresholdPnt", sweepNumber=sweepNumber, epochNumber=epochNumber)
    return theRet

getStat(statName1, statName2=None, sweepNumber=None, epochNumber=None, asArray=False, getFullList=False)

Get a list of values for one or two analysis results.

Parameters:

Name	Type	Description	Default
statName1	str	Name of the first analysis parameter to retreive.	required
statName2	str	Optional name of the second analysis parameter to retreive.	None
sweepNumber	int str or None	Optional sweep number, if None or 'All' then get all sweeps	None
epochNumber	int str or None	Optional epoch number, if None or 'All' then get all epochs	None
asArray	bool	If True then return as np.array(), otherwise return as a list	False

Notes

For a list of available analysis results, see bDetection.getDefaultDetection()

If the returned list of analysis results are in points, convert to seconds or ms using: pnt2Sec_(pnt) or pnt2Ms_(pnt).

Returns:

Type	Description
list or array	List of analysis parameter values, None if error. Returns a np.array is asArray is True

Source code in sanpy/bAnalysis_.py

def getStat(
    self,
    statName1,
    statName2: Optional[str] = None,
    sweepNumber: Optional[int] = None,
    epochNumber: Optional[int] = None,
    asArray: Optional[bool] = False,
    getFullList : Optional[bool] = False
):
    """Get a list of values for one or two analysis results.

    Parameters
    ----------
    statName1 : str
        Name of the first analysis parameter to retreive.
    statName2 : str
        Optional name of the second analysis parameter to retreive.
    sweepNumber : int str or None
        Optional sweep number, if None or 'All' then get all sweeps
    epochNumber : int str or None
        Optional epoch number, if None or 'All' then get all epochs
    asArray : bool
        If True then return as np.array(), otherwise return as a list

    Notes
    -----
    For a list of available analysis results,
        see [bDetection.getDefaultDetection()][sanpy.bDetection.bDetection]

    If the returned list of analysis results are in points,
        convert to seconds or ms using: pnt2Sec_(pnt) or pnt2Ms_(pnt).

    Returns
    -------
    list or np.array
        List of analysis parameter values, None if error.
        Returns a np.array is asArray is True
    """

    def clean(val):
        """Convert None to float('nan')"""
        if val is None:
            val = float("nan")
        return val

    x = []  # None
    y = []  # None
    error = False
    if len(self.spikeDict) == 0:
        # logger.error(f'Did not find any spikes in spikeDict')
        error = True
    elif statName1 not in self.spikeDict[0].keys():
        logger.error(f'Did not find statName1: "{statName1}" in spikeDict')
        # print('available stat names are:', self.spikeDict[0].keys())
        error = True
    elif statName2 is not None and statName2 not in self.spikeDict[0].keys():
        logger.error(f'Did not find statName2: "{statName2}" in spikeDict')
        error = True

    if sweepNumber is None:
        sweepNumber = "All"

    if epochNumber is None:
        epochNumber = "All"

    if not error:
        # original
        # x = [clean(spike[statName1]) for spike in self.spikeDict]

        if getFullList:
            # April 15, 2023, trying to fix bug in scatter plugin when we are
            # using sweep and epoch
            # strategy is to return all spikes, just nan out the ones we 
            # are not interested in
            x = []
            for spike in self.spikeDict:
                _include = \
                    (sweepNumber == "All" or spike["sweep"] == sweepNumber) \
                        and (epochNumber == "All" or spike["epoch"] == epochNumber)
                if _include:
                    x.append(clean(spike[statName1]))
                else:
                    x.append(float("nan"))

        else:
            # only current sweep and epoch
            # (1) was this
            # was causing errors with kym diam analysis
            x = [
                clean(spike[statName1])
                for spike in self.spikeDict
                if (sweepNumber == "All" or spike["sweep"] == sweepNumber)
                and (epochNumber == "All" or spike["epoch"] == epochNumber)
            ]
            # for _idx, spike in enumerate(self.spikeDict):
            #     if (sweepNumber == "All" or spike["sweep"] == sweepNumber) and (epochNumber == "All" or spike["epoch"] == epochNumber):
            #         try:
            #             val = spike[statName1]
            #         except (KeyError) as e:
            #             logger.error(f'did not find key "{statName1}" at spike {_idx}')
            #         clean(val)

        if statName2 is not None:
            # original
            # y = [clean(spike[statName2]) for spike in self.spikeDict]
            # only current spweek
            y = [
                clean(spike[statName2])
                for spike in self.spikeDict
                if sweepNumber == "All" or spike["sweep"] == sweepNumber
            ]

    if asArray:
        x = np.array(x)
        if statName2 is not None:
            y = np.array(y)

    if statName2 is not None:
        return x, y
    else:
        return x

getStatMean(statName, sweepNumber=None)

Get the mean of an analysis parameter.

Args: statName (str): Name of the statistic to retreive. For a list of available stats use bDetection.defaultDetection.

Source code in sanpy/bAnalysis_.py

def getStatMean(self, statName: str, sweepNumber: int = None):
    """
    Get the mean of an analysis parameter.

    Args:
        statName (str): Name of the statistic to retreive.
            For a list of available stats use bDetection.defaultDetection.
    """
    theMean = None
    x = self.getStat(statName, sweepNumber=sweepNumber)
    if x is not None and len(x) > 1:
        theMean = np.nanmean(x)
    return theMean

getSweepSpikeFromAbsolute(absSpikeIdx, sweep)

Get sweep spike from absolute spike.

See getAbsSpikeFromSweep()

Source code in sanpy/bAnalysis_.py

def getSweepSpikeFromAbsolute(self, absSpikeIdx: int, sweep: int) -> int:
    """Get sweep spike from absolute spike.

    See getAbsSpikeFromSweep()
    """
    sweepSpikeNum = self.spikeDict[absSpikeIdx]["sweepSpikeNumber"]
    return sweepSpikeNum

isAnalyzed()

Return True if this bAnalysis has been analyzed, False otherwise.

Source code in sanpy/bAnalysis_.py

def isAnalyzed(self):
    """Return True if this bAnalysis has been analyzed, False otherwise."""
    return self._isAnalyzed

isDirty()

Return True if analysis has been modified but not save.

Source code in sanpy/bAnalysis_.py

def isDirty(self):
    """Return True if analysis has been modified but not save."""
    return self._detectionDirty

printSpike(idx)

Print values in one spike analysis using self.spikeDict (sanpy.bAnalysisResults).

Source code in sanpy/bAnalysis_.py

def printSpike(self, idx):
    """
    Print values in one spike analysis using self.spikeDict (sanpy.bAnalysisResults).
    """
    spike = self.spikeDict[idx]
    for k, v in spike.items():
        if k == "widths":
            widths = v
            print(f"  spike:{idx} has {len(widths)} widths...")
            for wIdx, width in enumerate(widths):
                print(f"    spike:{idx} width:{wIdx}: {width}")
        elif k == "errors":
            errors = v
            print(f"  spike:{idx} has {len(errors)} errors...")
            for eIdx, error in enumerate(errors):
                print(f"    spike:{idx} error #:{eIdx}: {error}")
        else:
            print(f"{k}: {v}")

saveAnalysis(forceSave=False)

Not used.

Save detection parameters and analysis results as json.

Source code in sanpy/bAnalysis_.py

def saveAnalysis(self, forceSave=False):
    """Not used.

    Save detection parameters and analysis results as json.
    """
    if not self._detectionDirty and not forceSave:
        return

    saveFolder = self._getSaveFolder()
    if not os.path.isdir(saveFolder):
        logger.info(f"making folder: {saveFolder}")
        os.mkdir(saveFolder)

    saveBase = self._getSaveBase()
    savePath = saveBase + "-analysis.json"

    # save detection parameters
    # self.detectionClass.save(saveBase)
    dDict = self.detectionClass.getDict()

    saveDict = {}
    saveDict["detection"] = dDict

    # save list of dict
    # self.spikeDict = sanpy.bAnalysisResults.analysisResultList()
    # self.spikeDict.save(saveBase)
    analysisList = self.spikeDict.asList()

    saveDict["analysis"] = analysisList

    with open(savePath, "w") as f:
        json.dump(saveDict, f, cls=NumpyEncoder, indent=4)

    self._detectionDirty = False

    logger.info(f"Saved analysis to: {savePath}")

saveAnalysis_tocsv(path=None, verbose=False)

Save analysis to csv.

CSV starts with one

Parameters:

Name	Type	Description	Default
path	str	Full path of file to save, if None will save as default.	None

Source code in sanpy/bAnalysis_.py

def saveAnalysis_tocsv(self, path : str = None, verbose=False):
    """Save analysis to csv.

    CSV starts with one 
    Parameters
    ----------
    path : str
        Full path of file to save, if None will save as default.
    """

    if path is None:
        saveFolder = self._getSaveFolder()
        if not os.path.isdir(saveFolder):
            if verbose:
                logger.info(f"making folder: {saveFolder}")
            os.mkdir(saveFolder)

        saveBase = self._getSaveBase()
        path = saveBase + "-analysis.csv"

    if verbose:
        logger.info(f'saving to: {path}')

    metaDataHeader = self.metaData.getHeader()

    with open(path, "w") as f:
        f.write(metaDataHeader)
        f.write("\n")

    df = self.asDataFrame()  # pd.DataFrame(self.spikeDict)
    if df is not None:
        df.to_csv(path, mode="a")

setSpikeStat(spikeList, stat, value)

Set a spike stat for one spike or a list of spikes.

Used to set things like ('isBad', 'userType1', 'condition', ...)

Source code in sanpy/bAnalysis_.py

def setSpikeStat(self, spikeList: Union[list, int], stat: str, value):
    """Set a spike stat for one spike or a list of spikes.

    Used to set things like ('isBad', 'userType1', 'condition', ...)
    """
    if isinstance(spikeList, int):
        spikeList = [spikeList]
        # else:
        #     logger.error(f'Expecting list[int] or int but got spikeList type {type(spikeList)}')
        return

    if len(spikeList) == 0:
        return

    now = datetime.datetime.now()
    modDate = now.strftime("%Y%m%d")
    modTime = now.strftime("%H:%M:%S")

    for spike in spikeList:
        self.spikeDict[spike][stat] = value
        self.spikeDict[spike]["modDate"] = modDate
        self.spikeDict[spike]["modTime"] = modTime

    self._detectionDirty = True

    logger.info(f'set spikes {spikeList} stat "{stat}" to value "{value}"')

    """
    count = 0
    for idx, spike in enumerate(self.spikeDict):
        if idx in spikeList:
            try:
                spike[stat] = value
                count += 1
            except (KeyError) as e:
                logger.info(e)
    #
    logger.info(f'Given {len(spikeList)} and set {count}')
    """

setSpikeStat_time(startSec, stopSec, stat, value)

Set a spike stat for spikes in a range of time.

Source code in sanpy/bAnalysis_.py

def setSpikeStat_time(self, startSec: int, stopSec: int, stat: str, value):
    """Set a spike stat for spikes in a range of time."""

    # get spike list in range [startSec, stopSec]
    spikeSeconds = self.getSpikeSeconds()
    spikeList = [
        idx for idx, x in enumerate(spikeSeconds) if x >= startSec and x < stopSec
    ]
    self.setSpikeStat(spikeList, stat, value)

spikeDetect(detectionDict)

Run spike detection for all sweeps.

Each spike is a row and has 'sweep'

Args: detectionDict: From sanpy.bDetection

Source code in sanpy/bAnalysis_.py

def spikeDetect(self, detectionDict: dict):
    """Run spike detection for all sweeps.

    Each spike is a row and has 'sweep'

    Args:
        detectionDict: From sanpy.bDetection
    """

    rememberSweep = (
        self.fileLoader.currentSweep
    )  # This is BAD we are mixing analysis with interface !!!

    startTime = time.time()

    #
    # todo: ask user if they want to remove their settings for (isBad, userType)
    #

    self._detectionDict = detectionDict

    if detectionDict["verbose"]:
        logger.info("=== detectionDict is:")
        for k in detectionDict.keys():
            v = detectionDict[k]
            print(f'  {k} value:"{v}" is type {type(v)}')

    self._isAnalyzed = True

    self.spikeDict = sanpy.bAnalysisResults.analysisResultList()
    # we are filling this in, one dict for each spike
    # self.spikeDict = [] # we are filling this in, one dict for each spike

    # self._spikesPerSweep = [0] * self.fileLoader.numSweeps

    for sweepNumber in self.fileLoader.sweepList:
        # self.setSweep(sweep)
        self._spikeDetect2(sweepNumber)

    #
    self.fileLoader.setSweep(rememberSweep)

    stopTime = time.time()

    if detectionDict["verbose"]:
        logger.info(
            f"Detected {len(self.spikeDict)} spikes in {round(stopTime-startTime,3)} seconds"
        )