Common#

Extensions that are used for both MCorr (motion correction) and CNMF(E)

Accessor: caiman

DataFrame Extensions#

These can be called on the DataFrame

class mesmerize_core.CaimanDataFrameExtensions(df)#

Extensions for caiman related functions

Parameters:

df (DataFrame) –

uloc(u)#

Return the series corresponding to the passed UUID

Parameters:

u (str | UUID) –

Return type:

Series

add_item(algo, item_name, input_movie_path, params)#

Add an item to the DataFrame to organize parameters that can be used to run a CaImAn algorithm

Parameters:

  • algo (str) – Name of the algorithm to run, one of "mcorr", "cnmf" or "cnmfe"

  • item_name (str) – User set name for the batch item

  • input_movie_path (str) – Full path to the input movie

  • params (dict) – Parameters for running the algorithm with the input movie

save_to_disk(max_index_diff=0)#

Saves DataFrame to disk, copies to a backup before overwriting existing file.

Parameters:

max_index_diff (int) –

reload_from_disk()#

Returns the DataFrame on disk.

Example

df = df.caiman.reload_from_disk()
Return type:

pd.DataFrame

remove_item(index, remove_data=True, safe_removal=True)#

Remove a batch item from the DataFrame and delete all data associated to that batch item from disk if remove_data=True

Parameters:

  • index (int, str or UUID) – The index of the batch item to remove from the DataFrame as a numerical int index, str representing a UUID, or a UUID object.

  • remove_data (bool) – if True removes all output data associated to the batch item from disk. The input movie located at input_movie_path is not affected.

  • safe_removal (bool) –

    if True, this batch item is not removed and raises an exception if the output of this batch item is the input to another batch item. For example, if this is a motion correction batch item whose output is used as the input movie for a CNMF batch item.

    if False, this batch item is removed even if its output is the input to another batch item

get_params_diffs(algo, item_name)#

Get the parameters that differ for a given item_name run with a given algo

Parameters:

  • algo (str) – algorithm, one of “mcorr”, “cnmf”, or “cnmfe”

  • item_name (str) – The item name for which to get the parameter diffs

Returns:

pandas DataFrame) with dicts containing only the parameters that vary between batch items for the given item_name. The returned index corresponds to the index of the original DataFrame

Return type:

pd.DataFrame

get_children(index)#

For the motion correction batch item at the provided index, returns a list of UUIDs for CNMF(E) batch items that use the output of this motion correction batch item.

Note: Only Motion Correction items have children, CNMF(E) items do not have children.
Parameters:

index (int, str, or UUID) – the index of the mcorr item to get the children of, provided as a numerical int index, str representing a UUID, or a UUID object

Returns:

List of UUIDs of child CNMF items

Return type:

List[UUID]

get_parent(index)#

Get the UUID of the batch item whose output was used as the input for the batch item at the provided index.

If a parent exists, it is always an mcorr batch item
Parameters:

index (int, str, or UUID) – the index of the batch item to get the parent of, provided as a numerical int index, str representing a UUID, or a UUID object

Returns:

if UUID, this is the UUID of the batch item whose output was used for the input of the batch item at

the provided index

if None, the batch item at the provided index has no parent within the batch dataframe.

Return type:

UUID or None

Series Extensions#

These can be called on individual Series, or rows, of the DataFrame

class mesmerize_core.CaimanSeriesExtensions(s)#

Extensions for caiman stuff

Parameters:

s (Series) –

run(backend=None, wait=True, **kwargs)#

Run a CaImAn algorithm in an external process using the chosen backend

NoRMCorre, CNMF, or CNMFE will be run for this Series. Each Series (DataFrame row) has a input_movie_path and params for the algorithm

Parameters:

  • backend (str, optional) – One of the available backends, default on Linux & Mac is "subprocess". Default on Windows is "local" since Windows is inconsistent in the way it launches subprocesses

  • wait (bool, default True) – if using the "subprocess" backend, call wait() on the Popen instance before returning it

  • **kwargs – any kwargs to pass to the backend

get_input_movie_path()#
Returns:

full path to the input movie file

Return type:

Path

get_input_movie(reader=None, **kwargs)#

Get the input movie

Parameters:

  • reader (callable) – a function that take the input movie path and return an array-like

  • **kwargs – passed to reader function

Return type:

ndarray | Any

get_corr_image()#
Returns:

correlation image

Return type:

np.ndarray

get_pnr_image()#
Returns:

pnr image

Return type:

np.ndarray

get_projection(proj_type)#

Return the max, mean, or std (standard deviation) projection

Parameters:

proj_type (str) – one of "max", "mean", "std"

Returns:

max, mean, or std projection

Return type:

np.ndarray