tyssue.core package¶
Submodules¶
tyssue.core.history module¶

class
tyssue.core.history.
History
(sheet, save_every=None, dt=None, save_only=None, extra_cols=None, save_all=True)¶ Bases:
object
This class handles recording and retrieving time series of sheet objects.

property
cell_h
¶

property
edge_h
¶

property
face_h
¶

record
(time_stamp=None)¶ Appends a copy of the sheet datasets to the history instance.
time_stamp : float, save specific timestamp

retrieve
(time)¶ Return datasets at time time.
If a specific dataset was not recorded at time time, the closest record before that time is used.

property
time_stamps
¶

to_archive
(hf5file)¶ Saves the history to a HDF file
This file can later be accessed again with the HistoryHdf5.from_archive class method

property
vert_h
¶

property

class
tyssue.core.history.
HistoryHdf5
(sheet=None, save_every=None, dt=None, save_only=None, hf5file='', overwrite=False)¶ Bases:
tyssue.core.history.History
This class handles recording and retrieving time series of sheet objects.

classmethod
from_archive
(hf5file, columns=None, eptm_class=None)¶

record
(time_stamp=None, sheet=None)¶ Appends a copy of the sheet datasets to the history HDF file.
sheet: a
Sheet
object which we want to record. This argument can be used if the sheet object is different at each time point.

retrieve
(time)¶ Return datasets at time time.
If a specific dataset was not recorded at time time, the closest record before that time is used.

property
time_stamps
¶

classmethod
tyssue.core.monolayer module¶
Monolayer epithelium objects

class
tyssue.core.monolayer.
Monolayer
(name, datasets, specs=None, coords=None)¶ Bases:
tyssue.core.objects.Epithelium
3D monolayer epithelium

property
apical_edges
¶

property
apical_faces
¶

property
apical_verts
¶

property
basal_edges
¶

property
basal_faces
¶

property
basal_verts
¶

classmethod
from_flat_sheet
(name, apical_sheet, specs, thickness=1)¶

get_sub_sheet
(segment)¶ Returns a
Sheet
object of the corresponding segmentsegment: str, the corresponding segment, wether ‘apical’ or ‘basal’

guess_face_segment
(face)¶ Infers the face segment.

guess_vert_segment
(vert)¶ Infers the vertex segment from its surrounding edges.

property
lateral_edges
¶

property
lateral_faces
¶

property
lateral_verts
¶

segment_index
(segment, element)¶

property

class
tyssue.core.monolayer.
MonolayerWithLamina
(name, datasets, specs=None, coords=None)¶ Bases:
tyssue.core.monolayer.Monolayer
3D monolayer epithelium with a lamina meshing

property
lamina_edges
¶

property
tyssue.core.multisheet module¶
tyssue.core.objects module¶
Core definitions

class
tyssue.core.objects.
Epithelium
(identifier, datasets, specs=None, coords=None, maxbackup=5)¶ Bases:
object
Base class defining a connective tissue in 2D or 3D.

property
Nc
¶ The number of cells in the epithelium.

property
Ne
¶ The number of edges in the epithelium.

property
Nf
¶ The number of faces in the epithelium.

property
Nv
¶ The number of vertices in the epithelium.

backup
()¶ Creates a copy of self and keeps a reference to it in the self._backups deque.

property
cell_df
¶ The cell
pd.DataFrame
containing cell associated data e.g. the position of their center or their volume

copy
(deep_copy=True)¶ Returns a copy of the epithelium
 deep_copybool, default True
if True, use a copy of the original object’s datasets to create the new object. If False, datasets are not copied

cut_out
(bbox, coords=None)¶ Returns the index of edges with at least one vertex outside of the bounding box
 bboxsequence of shape (dim, 2)
the bounding box as (min, max) pairs for each coordinates.
 coordslist of str of len dim, default None
the coords corresponding to the bbox.

property
edge_df
¶ The edge
pd.DataFrame
containing edge associated data e.g. their length.This dataframe also contains the whole connexion of the epithelium through the “srce”, “trgt”, “face”, “cell” indices. In 2D, a halfedge is associated with a single (face, srce, trgt) positively oriented triangle. In 3D, the (cell, face, srce, trgt) positively oriented terahedron is also unique.

property
face_df
¶ The face
pd.DataFrame
containing face associated data e.g. the position of their center or their area

face_polygons
(coords=None)¶ Returns a pd.Series of arrays with the coordinates the face polygons
Each element of the Series is a (num_sides, num_dims) array of points ordered counterclockwise.
Vertices are assumed to be ordered in a face. If you are not sure it is the case, you can run sheet.reset_index(order=True) before calling this function.

get_invalid
()¶ Returns a mask over self.edge_df for invalid faces.

get_neighborhood
(elem_id, order, elem='cell')¶ Returns elem_id neighborhood up to a degree of order
For example, if order is 2, it wil return the adjacent cells (or faces) and theses cells neighbors.
 neighborspd.DataFrame with two colums, the index
of the neighboring cell (face), and it’s neighboring order

get_neighbors
(elem_id, elem='cell')¶ Returns the indexes of the adjacent elements (cells or faces) of the element of index elem_id.
 elem_idint
the index of the central element (a face or a cell)
element : {‘cell’  ‘face’}, default ‘cell’
 neghborsset
the cells (or faces) sharing an edge with the central cell (face)

get_opposite_faces
()¶ Populates the ‘opposite’ column of self.face_df with the index of the opposite face or 1 if the face has no opposite.

get_orbits
(center, periph)¶ Returns a dataframe with a (center, edge) MultiIndex with periph elements.
 centerstr,
the name of the center element for example ‘face’, ‘srce’
 periphstr,
the name of the periphery elements, for example ‘trgt’, ‘cell’
>>> cell_verts = sheet.get_orbits('face', 'srce') >>> cell_verts.loc[45] edge 218 75 219 78 220 76 221 81 222 90 223 87 Name: srce, dtype: int64

get_valid
()¶ Set the ‘is_valid’ column to true if the faces are all closed polygons, and the cells closed polyhedra.

idx_lookup
(elem_id, element)¶ returns the current index of the element with the “id” column equal to elem_id
 elem_idint
id of the element to retrieve
 element{“vert””edge””face””cell”}
the corresponding dataset.

remove
(edge_out, trim_borders=False, order_edges=False)¶ Remove the edges indexed by edge_out associated with all the cells and faces containing those edges
If trim_borders is True (defaults to False), there will be a single border edge per border face.

reset_index
(order=False)¶ Resets the datasets to have continuous indices
If order is True (the default), sorts the edges such that for each face, vertices are ordered clockwize

reset_topo
()¶ Recomputes the number of sides for the faces and the number of faces for the cells.

restore
()¶ Resets the eptithelium data to its last backed up state
A copy of the current state prior to restoring is kept in the _bad attribute for inspection. Calling this method multiple times (without calling backup) will go back in the epithelium backups.

sanitize
(trim_borders=False, order_edges=False)¶ Removes invalid faces and associated vertices
If trim_borders is True (defaults to False), there will be a single border edge per border face.

set_bbox
(margin=0.0)¶ Sets the attribute bbox with pairs of values bellow and above the min and max of the vert coords, with a margin.

property
settings
¶ Accesses the specs[‘settings’] dictionnary.

sum_cell
(df)¶ Sums the values of the edgeindexed dataframe df grouped by the values of self.edge_df[“cell”]
summed :
pd.DataFrame
the summed data, indexed by the source vertices.

sum_face
(df)¶ Sums the values of the edgeindexed dataframe df grouped by the values of self.edge_df[“face”]
summed :
pd.DataFrame
the summed data, indexed by the source vertices.

sum_srce
(df)¶ Sums the values of the edgeindexed dataframe df grouped by the values of self.edge_df[“srce”]
summed :
pd.DataFrame
the summed data, indexed by the source vertices.

sum_trgt
(df)¶ Sums the values of the edgeindexed dataframe df grouped by the values of self.edge_df[“trgt”]
summed :
pd.DataFrame
the summed data, indexed by the source vertices.

triangular_mesh
(coords=None, return_mask=False)¶ Return a triangulation of an epithelial sheet (2D in a 3D space), with added edges between face barycenters and junction vertices.
 coordslist of str:
pair of coordinates corresponding to column names for self.face_df and self.vert_df
 return_maskbool, optional, default True
if True, returns face_mask
 vertices(self.Nf+self.Nv, 3) ndarray
all the vertices’ coordinates
 triangles(self.Ne, 3) ndarray of ints
triple of the vertices’ indexes forming the triangular elements. For each junction edge, this is simply the index (srce, trgt, face). This is correctly oriented.
 face_mask: (self.Nf + self.Nv,) mask with 1 iff the vertex corresponds
to a face center

upcast_cell
(df)¶ Reindexes input data to self.edge_df.index by repeating the values for each cell entry
 df
pd.DataFrame
,pd.Series
np.ndarray
or string The data to be upcasted. If array like, should have self.Nc elements. If a string is passed it should be a column of self.cell_df
 upcast_df
pd.DataFrame
orpd.Series
The value repeated like the values of self.edge_df[“cell”]
 df

upcast_cols
(element, columns)¶ Syntactic sugar to upcast from the epithelium datasets.
 element: {‘srce’’trgt’’face’’cell’}
corresponding self.edge_df column over which to index if element is ‘srce’ or ‘trgt’, the upcast data will be taken form self.vert_df
 columns: index
the column(s) to be taken from the input dataset.

upcast_face
(df)¶ Reindexes input data to self.edge_df.index by repeating the values for each face entry
 df
pd.DataFrame
,pd.Series
np.ndarray
or string The data to be upcasted. If array like, should have self.Nf elements. If a string is passed it should be a column of self.face_df
 upcast_df
pd.DataFrame
orpd.Series
The value repeated like the values of self.edge_df[“face”]
 df

upcast_srce
(df)¶ Reindexes input data to self.edge_df.index by repeating the values for each source entry.
 df
pd.DataFrame
,pd.Series
np.ndarray
or string The data to be upcasted. If array like, should have self.Nv elements. If a string is passed it should be a column of self.vert_df
 upcast_df
pd.DataFrame
orpd.Series
The value repeated like the values of self.edge_df[“srce”]
 df

upcast_trgt
(df)¶ Reindexes input data to self.edge_df.index by repeating the values for each target entry
 df
pd.DataFrame
,pd.Series
np.ndarray
or string The data to be upcasted. If array like, should have self.Nv elements. If a string is passed it should be a column of self.vert_df
 upcast_df
pd.DataFrame
orpd.Series
The value repeated like the values of self.edge_df[“trgt”]
 df

update_num_faces
()¶ Updates the number of faces around the cells. The data is registered in the “num_faces” column of self.cell_df.

update_num_sides
()¶ Updates the number of halfedges around the faces. The data is registered in the “num_sides” column of self.face_df.

update_rank
()¶

update_specs
(new, reset=False)¶ Recursively updates the self.specs nested dictionnary, and set the new values to the corresponding columns in the datasets. If reset is True, existing values will be overwritten.

validate
()¶ returns True if the mesh is validated
e.g. has only closed polygons and polyhedra

validate_closed_cells
()¶ Returns True if all cells of the epithelium are closed.

property
vert_df
¶ The face
pd.DataFrame
containing vertex associated data e.g. the position of each vertex.

vertex_mesh
(coords, vertex_normals=True)¶ Returns the vertex coordinates and a list of vertex indices for each face of the tissue. If vertex_normals is True, also returns the normals of each vertex (set as the average of the vertex’ edges), suitable for .OBJ export
Vertices are assumed to be ordered in a face. If you are not sure it is the case, you can run sheet.reset_index() before calling this function.

property

tyssue.core.objects.
euler_characteristic
(edge_df)¶ Returns the Euler characteristic of the (non oriented) mesh represented by edge_df.
The Euler characteristic is the number of vertices minus the number of edges plus the number of faces
It is equal to 2 for a closedonitself mesh (topologicaly eq. to a sphere), 1 to a mesh with a border. It is not unique for monoloyers or bulk epithelia but provides a way to check wether a cell is closed.

tyssue.core.objects.
get_next_edges
(sheet)¶ returns a pd.Series with the index of the next edge for each edge

tyssue.core.objects.
get_opposite_faces
(eptm)¶

tyssue.core.objects.
get_prev_edges
(sheet)¶ returns a pd.Series with the index of the next edge for each edge

tyssue.core.objects.
get_simple_index
(edge_df)¶ returns a subset of the edge_df index corresponding to the non oriented edges (aka full edges).
This is faster than get_extra_indices and works also in 3D
tyssue.core.sheet module¶
An epithelial sheet, i.e a 2D mesh in a 2D or 3D space, akin to a HalfEdge data structure in CGAL.
 For purely 2D the geometric properties are defined in
tyssue.geometry.planar_geometry
A dynamical model derived from Fahradifar et al. 2007 is provided in tyssue.dynamics.planar_vertex_model
 For 2D in 3D, the geometric properties are defined in
tyssue.geometry.sheet_geometry
A dynamical model derived from Monier, Gettings et al. 2015 is provided in tyssue.dynamics.sheet_vertex_model

class
tyssue.core.sheet.
Sheet
(identifier, datasets, specs=None, coords=None)¶ Bases:
tyssue.core.objects.Epithelium
An epithelial sheet, i.e a 2D mesh in a 2D or 3D space, akin to a HalfEdge data structure in CGAL.
The geometric properties are defined in tyssue.geometry.sheet_geometry A dynamical model derived from Fahradifar et al. 2007 is provided in tyssue.dynamics.sheet_vertex_model

extract
(face_mask, coords=['x', 'y', 'z'])¶ Extract a new sheet from the sheet that correspond to a key word that define a face.
face_mask : column name in face composed by boolean value coords :
 sheet_fold_patch_extract :
subsheet corresponding to the fold patch area.

extract_bounding_box
(x_boundary=None, y_boundary=None, z_boundary=None, coords=['x', 'y', 'z'])¶ Extracts a new sheet from the embryo sheet
that correspond to boundary coordinate define by the user.
x_boundary : pair of floats y_boundary : pair of floats z_boundary : pair of floats coords : list of strings, default [‘x’, ‘y’, ‘z’]
coordinates over which to crop the sheet
subsheet : a new
Sheet
object

get_extra_indices
()¶ Computes extra indices:
self.free_edges: halfedges at the epithelium boundary
self.dble_edges: halfedges inside the epithelium, with an opposite
self.east_edges: half of the dble_edges, pointing east (figuratively)
 self.west_edges: half of the dble_edges, pointing west
(the order of the east and west edges is conserved, so that the ith west halfedge is the opposite of the ith east halfedge)
 self.sgle_edges: joint index over free and east edges, spanning
the entire graph without double edges
 self.wrpd_edges: joint index over free edges followed by the
 east edges twice, such that a vector over the whole halfedge
dataframe is wrapped over the single edges
 self.srtd_edges: index over the whole halfedge sorted such that
the free edges come first, then the east, then the west
Also computes:  self.Ni: the number of inside full edges
(i.e. len(self.east_edges))
self.No: the number of outside full edges (i.e. len(self.free_edges))
self.Nd: the number of double half edges (i.e. len(self.dble_edges))
self.anti_sym: pd.Series with shape (self.Ne,) with 1 at the free and east halfedges and 1 at the opposite halfedges.
East and west is resepctive to some orientation at the moment the indices are computed the partition stays valid as long as there are no changes in the topology, so due to vertex displacement, ‘east’ and ‘west’ might not stay valid. This is just a practical naming convention.
As the name suggest, this method is not working for edges in 3D pointing exactly north or south, ie iff edge[‘dx’] == edge[‘dy’] == 0. Until we need or find a better solution, we’ll just assert it worked.

get_force_inference
(coords=None, column=None, free_border_edges=False)¶ Measure force based on Brodland method.
g_gamma_matrix*tension_vector = 0 Equation is homogenous and to avoid tension_vectors = 0, Construction and solve the constrained leastsquares equation system [[g_gamma_matrix.T g_gamma_matrix, C^T_1],
[C_1, 0]] où C1 = {1….1}
shape of g_gamma_matrix = (Ne/2, Nv*len(coords))
Note
Results might not be consistens for highly curved epithelium
coords: coordinates column: None, specify a column name in edge_df to put tension value free_border_edges: bool, default False, take into account edges in the border of the tissue if True
edges_tensions: tension values array if column not define

get_neighborhood
(face, order, elem='face')¶ Returns face neighborhood up to a degree of order.
For example, if order is 2, it wil return the adjacent, faces and theses faces neighbors.
 neighborspd.DataFrame with two colums, the index
of the neighboring face, and it’s neighboring order

get_neighbors
(face, elem='face')¶ Returns the faces adjacent to face.

get_opposite
()¶

classmethod
planar_sheet_2d
(identifier, nx, ny, distx, disty, noise=None)¶ Creates a planar sheet from an hexagonal grid of cells.
identifier : string nx, ny : int
number of cells in the x and y axes
 distx, distyfloat,
the distances in x and y between the cells
 noisefloat, default None
position noise on the hexagonal grid
 planar_sheet: a 2D
Sheet
instance in the (x, y) plane

classmethod
planar_sheet_3d
(identifier, nx, ny, distx, disty, noise=None)¶ Creates a planar sheet from an hexagonal grid of cells.
identifier : string nx, ny : int
number of cells in the x and y axes
 distx, distyfloat,
the distances in x and y between the cells
 noisefloat, default None
position noise on the hexagonal grid
flat_sheet: a 2.5D
Sheet
instance

reset_topo
()¶ Recomputes the number of sides for the faces and the number of faces for the cells.

sort_edges_eastwest
()¶ reorder edges such the free edges are first, then the first half of the double edges, then the other half of the double edges, this way, each subset of the edges dataframe are contiguous.


tyssue.core.sheet.
get_opposite
(edge_df, raise_if_invalid=False)¶ Returns the indices opposite to the edges in edge_df

tyssue.core.sheet.
get_outer_sheet
(eptm)¶ Return a Sheet object formed by all the faces w/o an opposite face.