API reference

Public classes and functions in tglow-core v0.2.0, grouped by module, with their arguments and default values. Descriptions are the first paragraph of each docstring; entries without one are listed with their signature only. Names starting with an underscore are omitted, as are the tglow.qc modules, which are internal to the tglow-pipeline QC report.

Import from the full module path, for example:

from tglow.io.tglow_io import AICSImageReader
from tglow.utils.tglow_utils import rescale_stack

`tglow.io.compound_image_provider`

`memory_usage`

memory_usage()

class `CompoundImageProvider`

Method	Description
`__init__(path, nimg, channel, blacklist=None, plates=None, fields=None, planes=None, pseudoreplicates=0, merge_n=1, max_project=False, all_planes=False)`
`fetch_training_images()`
`fetch_compound_image_mean_max(max_project=False)`
`fetch_compound(sample_n)`
`fetch_image(q)`

`tglow.io.image_query`

Small utility for representing a plate/well/field/channel/plane query.

class `ImageQuery`

Container for plate/row/col/field/channel/plane identifiers.

Method	Description
`__init__(plate, row, col, field, channel=None, plane=None)`
`@classmethod from_plate_well(plate, well)`	Create an `ImageQuery` from a plate id and a well string (e.g. 'A01').
`get_well_id()`	Return a string well id, e.g. 'A01'.
`get_row_letter()`	Return the alphabetical row label for `row` (e.g. 'A').
`well_id_to_index(well_id) -> tuple`	Convert a well id string (e.g. 'A01') to a (row, col) tuple.
`to_relpath()`
`to_string()`	Return a compact string representation of the query.

`tglow.io.perkin_elmer_parser`

class `PerkinElmerParser` (object)

Parse PerkinElmer Index XML into a easy to access dictionary

Method	Description
`__init__(index_path, new_name=None)`
`estimate_pixel_sizes()`	Extracts the pixel sizes from a PE index xml in (z, y, x) and returns in in microns Returns none if could not be estimated
`parse_channels_from_images()`
`parse_channels()`	Get PE channels
`parse_planes()`	Get PE planes as a sorted list
`parse_images()`	Get all PE images
`parse_plate(new_name=None)`	Get all PE Plate
`parse_wells()`	Get all PE Wells
`parse_flatfields(use_background=False, channel=None)`
`reconstruct_flatfield_image(channel_data)`
`save(output_path)`
`write_manifest(output_path)`

`tglow.io.processed_image_provider`

Higher-level processed image provider.

class `ProcessedImageProvider`

Builds processed image stacks for a plate.

Method	Description
`__init__(path, plate, blacklist=None, plate_merge=None, registration_dir=None, flatfields=None, scaling_factors=None, mask_channels=None, mask_dir=None, mask_pattern=None, uint32=False, verbose=True, scaling_slope=0.001, scaling_bias=None)`
`build_channel_index()`	Construct a pandas DataFrame describing output channels.
`fetch_image(iq)`	Fetch and return a processed image stack for the provided `ImageQuery`.
`fetch_registration(iq) -> dict`	Load registration matrices for a well from the registration directory.
`get_wells()`	Return the set of wells indexed for the main plate.
`write_channel_index(output)`	Write `channel_index` to `output/<plate>/channel_indices.tsv`.

`tglow.io.tglow_io`

I/O helpers for tglow: readers, writers and index helpers.

class `NamedImageQuery` (ImageQuery)

Image query with an index that convers string names to indices

Method	Description
`__init__(index, plate, row, col, field, channel=None, plane=None)`

class `IndexedImageReader`

Read raw image data from a dictionary of /plate/row/col/field/channel/plane/path_to_file.tiff

Method	Description
`__init__(index, path, dtype=np.uint16, resolution=None) -> None`	Prefix path to the subpaths in index
`get_image(query) -> np.ndarray`	Get a single image
`read_stack(query) -> np.ndarray`	Read an image stack into a CZYX array
`get_filenames(query) -> list`	Get the filenames associated with an image query in the order they are read
`get_channel_order() -> list`
`get_plane_order() -> list`

class `PerkinElmerRawReader` (IndexedImageReader)

Read raw image data from a PerkinElmer export or from a re-formatted output as long as it has an index file

Method	Description
`__init__(index_xml, path, new_name=None, dtype=np.uint16) -> None`

class `AICSImageReader`

Reads image data from ome tiffs in a folder structure /plate/row/col/field.ome.tiff where field.ome.tiff is a CZYX array

Method	Description
`__init__(path, plates_filter=None, fields_filter=None, blacklist=None, dtype=np.uint16, resolution=None, pattern=None) -> None`
`get_wells(plate)`
`get_img(query)`
`get_fields(query)`
`read_image(query) -> np.ndarray`	Get a single image
`read_stack(query) -> np.ndarray`	Read an image stack into a CZYX array

class `BlacklistReader`

Method	Description
`__init__(path, sep='\t')`
`read_blacklist(separator=':')`	Read a simple blacklist file and return entries joined with `separator`.
`read_blacklist_as_prc(separator='/')`	Read the blacklist and return entries in 'plate/row/col' form.

class `ControlRecord`

Method	Description
`__init__(plate, well, channels, name)`
`get_row_col()`	Return (row, col) for the control well.
`get_rowchar()`	Return the row letter for the control (e.g. 'A').
`get_query(field)`	Return an `ImageQuery` for this control at the given `field`.

class `ControllistReader`

Method	Description
`__init__(path, plates_filter=None, blacklist=None, sep='\t')`
`read_controlist()`	Read a control list TSV and return `ControlRecord` objects.

class `AICSImageWriter`

Writes image data from ome tiffs in a folder structure /plate/row/col/field.ome.tiff where field.ome.tiff is a CZYX array

Method	Description
`__init__(path, channel_names=None, physical_pixel_sizes=None, skip_imagestats=False) -> None`
`write_stack(stack, query, channel_names=None, physical_pixel_sizes=None, image_names=None, stats_only=False)`	Write a CZYX array into the folder structure
`write_image_stats(query)`

`tglow.utils.tglow_plot`

`composite_images`

composite_images(imgs, equalize=False, aggregator=np.mean)

`plot_registration_imgs`

plot_registration_imgs(before, after, filename)

`plot_grey_as_magma`

plot_grey_as_magma(image, filename)

`plot_blobs`

plot_blobs(image, filename, blobs)

`plot_histogram`

plot_histogram(data, filename, bins=30, title='Histogram', xlabel='Value', ylabel='Frequency')

Create and save a histogram to a file.

`plot_histogram_df`

plot_histogram_df(data, filename, bins=30, title_prefix='Histogram', xlabel='Value', ylabel='Frequency', ncols=3)

Create and save histograms for each column in a pandas DataFrame with mean and median lines.

`tglow.utils.tglow_utils`

Utility helpers for tglow.

class `NpEncoder` (json.JSONEncoder)

JSON encoder that understands numpy types.

Method	Description
`default(obj)`

`get_channel_channel_info`

get_channel_channel_info(index_xml) -> dict

Extract channel metadata from a PerkinElmer Index.xml file.

`build_well_index`

build_well_index(upper=True, invert=False, rows_as_string=False) -> dict

`default_to_regular`

default_to_regular(d)

Convert nested defaultdicts into regular dicts recursively.

`float_to_16bit_unint_scaled`

float_to_16bit_unint_scaled(matrix, max_value) -> np.array

Scale and convert a float matrix into uint16 using a provided max.

`float_to_16bit_unint`

float_to_16bit_unint(matrix) -> np.array

`float_to_16bit_unint_inplace`

float_to_16bit_unint_inplace(matrix)

In-place conversion of floats to uint16 with rounding and clipping.

`float_to_32bit_unint`

float_to_32bit_unint(matrix) -> np.array

`dict_to_str`

dict_to_str(dict)

Return a compact string describing a dict's keys and value types.

`write_bin`

write_bin(matrix, file)

Write a numeric matrix to a simple binary format.

`apply_registration`

apply_registration(stack, alignment_matrix)

Apply a 2D affine transformation to a 2D/3D/4D image stack.

`apply_registration_cv`

apply_registration_cv(stack, alignment_matrix)

`sigmoid`

sigmoid(x, slope, bias)

`sigmoid_params`

sigmoid_params(x1, x2, tol=1e-06)

Compute the bias and slope for a logistic sigmoid given two x points and a tolerance.

`rescale_stack`

rescale_stack(stack, factors, slopes=None, biases=None, verbose=True)

Apply per-channel scale factors (optionally sigmoid-weighted) to a CZYX stack, returned as a new float32 array.

`rescale_stack_inplace`

rescale_stack_inplace(stack, factors, slopes=None, biases=None, verbose=True)

Apply per-channel scale factors (optionally sigmoid-weighted) to a CZYX stack in place, returned as float32.

`load_flatfield_profile`

load_flatfield_profile(model_dir)

Read the flatfield/darkfield/baseline arrays from a BaSiC.save_model directory.

`save_flatfield_profile`

save_flatfield_profile(model_dir, flatfield, darkfield, baseline=None, overwrite=False)

Write flatfield/darkfield/baseline arrays in the same on-disk format as BaSiC.save_model.

#API reference

#tglow.io.compound_image_provider

#memory_usage

#class CompoundImageProvider

#tglow.io.image_query

#class ImageQuery

#tglow.io.perkin_elmer_parser

#class PerkinElmerParser (object)

#tglow.io.processed_image_provider

#class ProcessedImageProvider

#tglow.io.tglow_io

#class NamedImageQuery (ImageQuery)

#class IndexedImageReader

#class PerkinElmerRawReader (IndexedImageReader)

#class AICSImageReader

#class BlacklistReader

#class ControlRecord

#class ControllistReader

#class AICSImageWriter

#tglow.utils.tglow_plot

#composite_images

#plot_registration_imgs

#plot_grey_as_magma

#plot_blobs

#plot_histogram

#plot_histogram_df

#tglow.utils.tglow_utils

#class NpEncoder (json.JSONEncoder)

#get_channel_channel_info

#build_well_index

#default_to_regular

#float_to_16bit_unint_scaled

#float_to_16bit_unint

#float_to_16bit_unint_inplace

#float_to_32bit_unint

#dict_to_str

#write_bin

#apply_registration

#apply_registration_cv

#sigmoid

#sigmoid_params

#rescale_stack

#rescale_stack_inplace

#load_flatfield_profile

#save_flatfield_profile