dyco pwb-batch

Detect lags only, across files that are already split into averaging periods. Writes tlag_results.csv and stops — nothing is shifted and no corrected raw data is produced. Pair it with dyco apply-batch.

Also available as the standalone dyco-pwb-batch.

Important

Input files must already be wind-rotation corrected (double rotation or planar fit, e.g. EddyPro “Advanced” rotated output). This command does not rotate; a non-zero mean W corrupts the cross-correlation. dyco detect-remove handles rotation itself, and is the right choice unless your files are already split.

All options

Parallel PWB time-lag detection across EddyPro high-frequency files. Alias: uv run dyco-pwb-batch

usage: dyco pwb-batch [-h] --input-dir INPUT_DIR [--file-pattern FILE_PATTERN]
                      --output-dir OUTPUT_DIR --scalar LABEL:column
                      [--col-w COL_W] --col-tsonic COL_TSONIC
                      [--usecols USECOLS [USECOLS ...]]
                      [--col-names COL_NAMES [COL_NAMES ...]]
                      [--skiprows SKIPROWS]
                      [--na-values NA_VALUES [NA_VALUES ...]] [--hz HZ]
                      [--lag-max LAG_MAX] [--n-bootstrap N_BOOTSTRAP]
                      [--block-length BLOCK_LENGTH]
                      [--min-valid-frac MIN_VALID_FRAC]
                      [--hdi-thresh HDI_THRESH] [--dev-thresh DEV_THRESH]
                      [--hdi-prefilter HDI_PREFILTER] [--n-workers N_WORKERS]
                      [--save-plots] [--random-state RANDOM_STATE]
                      [--file-date-format FILE_DATE_FORMAT]

Named Arguments

--input-dir

Directory containing EddyPro rotated .txt files.

--file-pattern

Glob pattern for input files.

Default: '*.txt'

--output-dir

Directory for results CSV, checkpoint, and optional plots.

--scalar

Gas label and column name, e.g. CH4:ch4. Repeat for each gas.

--col-w

Column name for vertical wind W.

Default: 'w'

--col-tsonic

Column name for sonic temperature T_SONIC (required; enables 4-combination RFlux v3.2.0 mode).

--usecols

0-based column indices to read from each file.

Default: [0, 1, 2, 3, 4, 5]

--col-names

Column names to assign after selecting –usecols.

Default: ['u', 'v', 'w', 'ts', 'ch4', 'n2o']

--skiprows

Metadata rows before the column-name row.

Default: 9

--na-values

Strings to treat as NaN.

Default: ['-9999', '-9999.0', '-9999.0000000000000']

--hz

Sampling frequency in Hz.

Default: 20

--lag-max

CCF search half-width in seconds.

Default: 10.0

--n-bootstrap

Number of block-bootstrap replicates.

Default: 99

--block-length

Bootstrap block length in seconds.

Default: 20.0

--min-valid-frac

Minimum non-NaN fraction for a series to be processed.

Default: 0.3

--hdi-thresh

S1 HDI threshold in seconds.

Default: 0.5

--dev-thresh

S2 deviation threshold in seconds.

Default: 0.5

--hdi-prefilter

HDI pre-filter threshold in seconds (0 = disabled).

Default: 1.0

--n-workers

Parallel worker processes (default: os.cpu_count()).

--save-plots

Save one diagnostic PNG per period per scalar.

Default: False

--random-state

Base seed for reproducible bootstrap results.

--file-date-format

strptime pattern for the leading datetime in each filename, e.g. ‘%Y%m%d-%H%M’ for ‘20210820-0930_rotated.txt’. Adds a ‘timestamp’ column and uses dates as the summary-plot x-axis.