From 4f302a13ee818196ebd03871484a29cef37dfa99 Mon Sep 17 00:00:00 2001
From: Swainlab <peter.swain@ed.ac.uk>
Date: Sun, 2 Jul 2023 11:48:19 +0100
Subject: [PATCH] docs for buddings
---
src/agora/io/signal.py | 4 +-
src/aliby/pipeline.py | 38 +------------
src/postprocessor/core/processor.py | 6 +-
src/postprocessor/core/reshapers/buddings.py | 59 ++++++++++----------
4 files changed, 38 insertions(+), 69 deletions(-)
diff --git a/src/agora/io/signal.py b/src/agora/io/signal.py
index cc852089..20b6d8ba 100644
--- a/src/agora/io/signal.py
+++ b/src/agora/io/signal.py
@@ -20,7 +20,9 @@ class Signal(BridgeH5):
"""
Fetch data from h5 files for post-processing.
- Signal assumes that the metadata and data are accessible to perform time-adjustments and apply previously recorded post-processes.
+ Signal assumes that the metadata and data are accessible to
+ perform time-adjustments and apply previously recorded
+ post-processes.
"""
def __init__(self, file: t.Union[str, Path]):
diff --git a/src/aliby/pipeline.py b/src/aliby/pipeline.py
index 6e37bac2..52638883 100644
--- a/src/aliby/pipeline.py
+++ b/src/aliby/pipeline.py
@@ -592,41 +592,6 @@ class Pipeline(ProcessABC):
)
return (traps_above_nthresh & traps_above_athresh).mean()
- # FIXME: Remove this functionality. It used to be for
- # older hdf5 file formats.
- def _load_config_from_file(
- self,
- filename: Path,
- process_from: t.Dict[str, int],
- trackers_state: t.List,
- overwrite: t.Dict[str, bool],
- ):
- with h5py.File(filename, "r") as f:
- for k in process_from.keys():
- if not overwrite[k]:
- process_from[k] = self.legacy_get_last_tp[k](f)
- process_from[k] += 1
- return process_from, trackers_state, overwrite
-
- # FIXME: Remove this functionality. It used to be for
- # older hdf5 file formats.
- @staticmethod
- def legacy_get_last_tp(step: str) -> t.Callable:
- """Get last time-point in different ways depending
- on which step we are using
-
- To support segmentation in aliby < v0.24
- TODO Deprecate and replace with State method
- """
- switch_case = {
- "tiler": lambda f: f["trap_info/drifts"].shape[0] - 1,
- "baby": lambda f: f["cell_info/timepoint"][-1],
- "extraction": lambda f: f[
- "extraction/general/None/area/timepoint"
- ][-1],
- }
- return switch_case[step]
-
def _setup_pipeline(
self, image_id: int
) -> t.Tuple[
@@ -682,7 +647,6 @@ class Pipeline(ProcessABC):
step: self.step_sequence.index(ow_id) < i
for i, step in enumerate(self.step_sequence, 1)
}
-
# set up
directory = config["general"]["directory"]
trackers_state: t.List[np.ndarray] = []
@@ -722,7 +686,7 @@ class Pipeline(ProcessABC):
)
config["tiler"] = steps["tiler"].parameters.to_dict()
except Exception:
- self._log(f"Overwriting tiling data")
+ self._log("Overwriting tiling data")
if config["general"]["use_explog"]:
meta.run()
diff --git a/src/postprocessor/core/processor.py b/src/postprocessor/core/processor.py
index 76be13cd..ec9bcd04 100644
--- a/src/postprocessor/core/processor.py
+++ b/src/postprocessor/core/processor.py
@@ -139,7 +139,7 @@ class PostProcessor(ProcessABC):
for k in dicted_params.keys():
if not isinstance(dicted_params[k], dict):
dicted_params[k] = dicted_params[k].to_dict()
- # merger and picker
+ # initialise merger and picker
self.merger = Merger(
MergerParameters.from_dict(dicted_params["merger"])
)
@@ -147,12 +147,12 @@ class PostProcessor(ProcessABC):
PickerParameters.from_dict(dicted_params["picker"]),
cells=Cells.from_source(filename),
)
- # processes, such as buddings
+ # get processes, such as buddings
self.classfun = {
process: get_process(process)
for process, _ in parameters["targets"]["processes"]
}
- # parameters for the process in classfun
+ # get parameters for the processes in classfun
self.parameters_classfun = {
process: get_parameters(process)
for process, _ in parameters["targets"]["processes"]
diff --git a/src/postprocessor/core/reshapers/buddings.py b/src/postprocessor/core/reshapers/buddings.py
index 4398b947..ba9fe2fc 100644
--- a/src/postprocessor/core/reshapers/buddings.py
+++ b/src/postprocessor/core/reshapers/buddings.py
@@ -13,25 +13,18 @@ from postprocessor.core.lineageprocess import (
class buddingsParameters(LineageProcessParameters):
- """
- Parameter class to obtain budding events.
-
- Define the location of lineage information in the h5 file.
-
- """
+ """Give the location of lineage information in the h5 file."""
_defaults = {"lineage_location": "postprocessing/lineage_merged"}
class buddings(LineageProcess):
"""
- Calculate buddings in a trap assuming one mother per trap.
+ Generate a dataframe of budding events.
- Return a pandas series with the buddings.
+ We assume one mother per trap.
- We define a budding event as when a bud is first identified.
-
- This bud may not be considered a bud until later in the experiment.
+ A bud may not be considered a bud until later in the experiment.
"""
def __init__(self, parameters: buddingsParameters):
@@ -41,38 +34,48 @@ class buddings(LineageProcess):
def run(
self, signal: pd.DataFrame, lineage: np.ndarray = None
) -> pd.DataFrame:
- """TODO."""
+ """
+ Generate dataframe of budding events.
+
+ Find daughters for mothers in a Signal for which we have lineage data.
+ Create a dataframe indicating the time each daughter first appears.
+
+ We use the data from Signal only to find when the daughters appear, by
+ their first non-NaN value.
+ """
+ # lineage is (trap, mother, daughter)
lineage = lineage or self.lineage
- # select traps and mother cells in a given signal
+ # select traps and mothers in the signal that have lineage data
traps_mothers: t.Dict[tuple, list] = {
- tuple(mo): [] for mo in lineage[:, :2] if tuple(mo) in signal.index
+ tuple(trap_mo): []
+ for trap_mo in lineage[:, :2]
+ if tuple(trap_mo) in signal.index
}
+ # find daughters for these traps and mothers
for trap, mother, daughter in lineage:
if (trap, mother) in traps_mothers.keys():
traps_mothers[(trap, mother)].append(daughter)
+ # sub dataframe of signal for the selected mothers
mothers = signal.loc[
set(signal.index).intersection(traps_mothers.keys())
]
- # create a new dataframe with dimensions (n_mother_cells * n_timepoints)
+ # a new dataframe with dimensions (n_mother_cells * n_tps)
buddings = pd.DataFrame(
np.zeros((mothers.shape[0], signal.shape[1])).astype(bool),
index=mothers.index,
columns=signal.columns,
)
buddings.columns.names = ["timepoint"]
- # get time of first appearance for every cell using Pandas
+ # get time of first non-NaN value of signal for every mother using Pandas
fvi = signal.apply(lambda x: x.first_valid_index(), axis=1)
# fill the budding events
- for mother_id, daughters in traps_mothers.items():
- daughters_idx = set(
- fvi.loc[
- fvi.index.intersection(
- list(product((mother_id[0],), daughters))
- )
- ].values
- ).difference({0})
- buddings.loc[
- mother_id,
- daughters_idx,
- ] = True
+ for trap_mother_id, daughters in traps_mothers.items():
+ times_of_bud_appearance = fvi.loc[
+ fvi.index.intersection(
+ list(product((trap_mother_id[0],), daughters))
+ )
+ ].values
+ # ignore zeros - ignore buds in first image
+ daughters_idx = set(times_of_bud_appearance).difference({0})
+ buddings.loc[trap_mother_id, daughters_idx] = True
return buddings
--
GitLab