phenotypic.post package#

Post-measurement DataFrame transforms for measurement pipelines.

Provides operations that reshape or enrich measurement DataFrames after feature extraction. These run as the final stage of ImagePipeline.measure().

class phenotypic.post.AppendString(column: str = '', value: str = '')[source]#

Bases: PostMeasurement

Append a string to every value in a metadata column.

Converts each cell to a string and concatenates the given value to the end. Useful for adding suffixes like units or experimental labels.

Parameters:

  • column (str) – Name of the metadata column to modify. Metadata_ prefix is added automatically if missing.

  • value (str) – The string to append to each cell value.

Returns:

A copy of the input DataFrame with the modified column.

Return type:

pd.DataFrame

Raises:

KeyError – If the column does not exist in the DataFrame.

Examples

Append a unit suffix to a temperature column:

>>> import pandas as pd
>>> from phenotypic.post import AppendString
>>> df = pd.DataFrame({
...     "Metadata_Temp": ["30", "37"],
...     "ObjectLabel": [1, 2],
... })
>>> op = AppendString(column="Temp", value="C")
>>> result = op.apply(df)
>>> list(result["Metadata_Temp"])
['30C', '37C']
__del__()#

Automatically stop tracemalloc when the object is deleted.

apply(df: pandas.DataFrame) → pandas.DataFrame#

Apply the post-measurement transform.

Parameters:

df (pandas.DataFrame) – The merged measurement DataFrame.

Returns:

The transformed DataFrame.

Return type:

pandas.DataFrame

class phenotypic.post.ExpandMetadata(column: str = '', labels: List[str] | None = None, delimiter: str = '_', regex: bool = False)[source]#

Bases: PostMeasurement

Split a metadata column into multiple new metadata columns.

Takes a single delimited metadata column (e.g., a filename encoding experimental conditions) and expands it into separate labeled columns. Every row must produce exactly len(labels) parts or a ValueError is raised.

Parameters:

  • column (str) – Name of the metadata column to split. Metadata_ prefix is added automatically if missing.

  • labels (List[str] | None) – Names for the resulting columns, one per split part. Metadata_ prefix is added automatically if missing.

  • delimiter (str) – String or regex pattern to split on. Defaults to "_".

  • regex (bool) – If True, treat delimiter as a regex pattern. Defaults to False.

Returns:

The input DataFrame with new columns inserted

adjacent to the source column. The source column is always kept.

Return type:

pd.DataFrame

Raises:

  • ValueError – If labels is empty, or if any row produces a different number of parts than len(labels).

  • KeyError – If the source column does not exist in the DataFrame.

Examples

Split a filename into experimental conditions:

>>> import pandas as pd
>>> from phenotypic.post import ExpandMetadata
>>> df = pd.DataFrame({
...     "Metadata_ImageName": ["WT_30C_24h", "mut_37C_48h"],
...     "ObjectLabel": [1, 2],
... })
>>> expand = ExpandMetadata(
...     column="ImageName",
...     labels=["Strain", "Condition", "Time"],
...     delimiter="_",
... )
>>> result = expand.apply(df)
>>> list(result["Metadata_Strain"])
['WT', 'mut']
__del__()#

Automatically stop tracemalloc when the object is deleted.

apply(df: pandas.DataFrame) → pandas.DataFrame#

Apply the post-measurement transform.

Parameters:

df (pandas.DataFrame) – The merged measurement DataFrame.

Returns:

The transformed DataFrame.

Return type:

pandas.DataFrame

class phenotypic.post.MergeMetadata(columns: List[str] | None = None, label: str = '', delimiter: str = '_')[source]#

Bases: PostMeasurement

Merge multiple metadata columns into a single new metadata column.

Concatenates the string values of two or more metadata columns using a delimiter to produce a combined identifier. Useful for creating composite keys (e.g., combining Strain and Condition into SampleID).

Parameters:

  • columns (List[str] | None) – Names of the metadata columns to merge. Metadata_ prefix is added automatically if missing. Must contain at least 2 names.

  • label (str) – Name for the new merged column. Metadata_ prefix is added automatically if missing.

  • delimiter (str) – String used to join the column values. Defaults to "_".

Returns:

The input DataFrame with the new merged column

inserted after the last source column. All source columns are kept.

Return type:

pd.DataFrame

Raises:

  • ValueError – If columns contains fewer than 2 names.

  • KeyError – If any source column does not exist in the DataFrame.

Examples

Merge strain and condition into a sample ID:

>>> import pandas as pd
>>> from phenotypic.post import MergeMetadata
>>> df = pd.DataFrame({
...     "Metadata_Strain": ["WT", "mut"],
...     "Metadata_Condition": ["30C", "37C"],
...     "ObjectLabel": [1, 2],
... })
>>> merge = MergeMetadata(
...     columns=["Strain", "Condition"],
...     label="SampleID",
...     delimiter="_",
... )
>>> result = merge.apply(df)
>>> list(result["Metadata_SampleID"])
['WT_30C', 'mut_37C']
__del__()#

Automatically stop tracemalloc when the object is deleted.

apply(df: pandas.DataFrame) → pandas.DataFrame#

Apply the post-measurement transform.

Parameters:

df (pandas.DataFrame) – The merged measurement DataFrame.

Returns:

The transformed DataFrame.

Return type:

pandas.DataFrame

class phenotypic.post.PrependString(column: str = '', value: str = '')[source]#

Bases: PostMeasurement

Prepend a string to every value in a metadata column.

Converts each cell to a string and concatenates the given value to the beginning. Useful for adding prefixes like identifiers or labels.

Parameters:

  • column (str) – Name of the metadata column to modify. Metadata_ prefix is added automatically if missing.

  • value (str) – The string to prepend to each cell value.

Returns:

A copy of the input DataFrame with the modified column.

Return type:

pd.DataFrame

Raises:

KeyError – If the column does not exist in the DataFrame.

Examples

Prepend a strain prefix to an ID column:

>>> import pandas as pd
>>> from phenotypic.post import PrependString
>>> df = pd.DataFrame({
...     "Metadata_ID": ["001", "002"],
...     "ObjectLabel": [1, 2],
... })
>>> op = PrependString(column="ID", value="WT-")
>>> result = op.apply(df)
>>> list(result["Metadata_ID"])
['WT-001', 'WT-002']
__del__()#

Automatically stop tracemalloc when the object is deleted.

apply(df: pandas.DataFrame) → pandas.DataFrame#

Apply the post-measurement transform.

Parameters:

df (pandas.DataFrame) – The merged measurement DataFrame.

Returns:

The transformed DataFrame.

Return type:

pandas.DataFrame