DroughtSpotter Utilities

Analysis pipeline for Phenospex DroughtSpotter data exported from TraitFinder. Produces weather and weight summary plots from a single zip file.

Setup

Create and activate the conda environment:

conda env create -f environment.yml
conda activate DroughtspotterUtilities

Folder layout

Place these files together in one folder before running:

my_experiment/
  Tabular-data_Experiment_20260701.zip   ← TraitFinder zip export
  experimentFile.csv                      ← experiment program CSV (target weights)

Running the pipeline

python run_pipeline.py "Tabular-data_Experiment_20260701.zip"

To change the number of days shown in the recent-data view (default is 5):

python run_pipeline.py "Tabular-data_Experiment_20260701.zip" --days 7

Always quote the zip path if it contains spaces.

Output structure

Each run creates a datetime-stamped folder inside outputs/ next to the zip:

my_experiment/
  outputs/
    2026-07-01_143210/
      pipeline_metadata.json
      most_recent/                         ← last N days
        *_weather_timeseries.png
        *_weather_daily.png
        *_weights_overview.png
        unit_plots/
          *_unit_*.png
      entire_dataset/                      ← full experiment duration
        *_weather_timeseries.png
        *_weather_daily.png
        *_weights_overview.png
        unit_plots/
          *_unit_*.png

Plot descriptions

File	Contents
*_weather_timeseries.png	Temperature, relative humidity, and PAR (flux density) at 10-min intervals
*_weather_daily.png	Daily mean ± min/max ribbon for each weather variable
*_weights_overview.png	All 120 unit weights on one chart, coloured by treatment
unit_plots/*_unit_*.png	Per-unit weight trace (blue) with target weight (red dashed) and air temperature overlay (orange, secondary axis)

Off-scale readings

When a balance is removed from the scale mid-experiment, the resulting near-zero weight reading is kept in the data but excluded from y-axis scaling — it will appear as a dip out of view rather than distorting the chart.

Per-unit plots — two panels

Each unit plot has two side-by-side panels:

• Local scale — y-axis fitted to this unit's own weight range, for pattern detail
• Experiment scale — y-axis shared across all units, for cross-unit comparison

experimentFile.csv format

The experiment CSV must contain a section with a header row that includes block:column:row. The first column is the unit ID and the second column is the target weight in grams. All rows above the header are ignored.

Example:

... (global settings) ...
unit (block:column:row),h,Irrigation_mode,...
1:1:1,25400,...
2:1:1,25800,...