Export (layer 4)¶

tables ¶

The results as tables (layer 4): one analysis, a whole study, CSV files.

One analysis is the output of layer 3 for one phase of a recording (TidalFeatures or PeepTrialFeatures). Its tables start every row with the keys patient_id, label (the recording, e.g. a time point) and phase (the part analysed):

recordings: one row per analysis: file, breaths, results of a PEEP trial;
breaths: one row per breath of the phase;
steps: one row per PEEP step (PEEP trials only);
settings: the parameters, the context and the version of fastEIT used;
log: the messages of every layer.

The objects of layers 2-3 store each value once; here they are joined, so that a breath or a step is one row with everything (times, quality, pressures, volumes, indices). Regional values are columns named {index}__{region}, for the halves of the lung (ventral, dorsal, right, left; the layers and the quadrants stay in the objects), with two underscores because both names can contain one: in R tidyr::pivot_longer(contains("__"), names_to = c("index", "region"), names_sep = "__") gives back one row per region.

recording_tables ¶

recording_tables(features: TidalFeatures | PeepTrialFeatures, label: str | None = None, phase: str | None = None) -> dict[str, pd.DataFrame]

Tables of one analysis (one phase of a recording).

Parameters:

Name	Type	Description	Default
`features`	`TidalFeatures \| PeepTrialFeatures`	Output of `tidal_features()` or `peep_trial_features()`.	required
`label`	`str \| None`	Name of the recording; `RecordingContext.label` if None.	`None`
`phase`	`str \| None`	Name of the phase, e.g. `"baseline"`; if None, the protocol (`"tidal"` or `"peep_trial"`). Needed when a recording has two phases of the same protocol.	`None`

Returns:

Type	Description
`dict[str, DataFrame]`	`recordings`, `breaths`, `steps` (PEEP trials only), `settings` and
`dict[str, DataFrame]`	`log`; every row starts with `patient_id`, `label`, `phase`.

combine_tables ¶

combine_tables(analyses: Sequence[dict[str, DataFrame]]) -> dict[str, pd.DataFrame]

The tables of several analyses (a study) as one set of tables.

A table missing in an analysis (steps of a tidal recording) is skipped; columns missing in an analysis (e.g. the results of a PEEP trial in a tidal row) are empty.

Parameters:

Name	Type	Description	Default
`analyses`	`Sequence[dict[str, DataFrame]]`	Outputs of `recording_tables()`.	required

Returns:

Type	Description
`dict[str, DataFrame]`	The same tables, with the rows of every analysis.

Raises:

Type	Description
`ValueError`	if there is no analysis, or two have the same `patient_id`, `label` and `phase`.

write_csv ¶

write_csv(tables: dict[str, DataFrame], folder: str | Path) -> dict[str, Path]

Write one CSV file per table, e.g. folder/breaths.csv.

UTF-8, comma separated, decimal point, empty cell for a missing value, booleans as 1 / 0: readable by R, SPSS, Excel and Python. Files with the same name are replaced.

Parameters:

Name	Type	Description	Default
`tables`	`dict[str, DataFrame]`	Output of `recording_tables()` or `combine_tables()`.	required
`folder`	`str \| Path`	Folder of the files, created if missing.	required

Returns:

Type	Description
`dict[str, Path]`	The path of each file, by table name.