Data & Output Files¶

Every procedure run writes a self-describing CSV file. This page explains where files go, how they're named, what's inside them, and how to aggregate them into a database.

Where data is saved¶

Files are written under the data directory, configured by Dir.data_dir (default ./data). With Filename.dated_folder: true, runs are grouped into per-day subfolders:

data/
└── 2026-06-21/
    ├── FakeProcedure2026-06-21_1.csv
    ├── It2026-06-21_1.csv
    └── It2026-06-21_2.csv

File naming¶

The filename is assembled from the Filename config section (parsed by PyMeasure):

Filename:
  prefix: ''            # empty → use the procedure class name
  suffix: ''
  ext: csv
  dated_folder: true    # data/<date>/
  index: true           # append _1, _2, … to avoid collisions
  datetimeformat: "%Y-%m-%d"

So the default pattern is:

{data_dir}/{date}/{ProcedureClass}{date}_{index}.{ext}

Set prefix to use a fixed name instead of the class name, or suffix to tag files (e.g. a run label).

Anatomy of a CSV¶

A PyMeasure results file has a header (procedure, parameters, metadata) and a data section:

#Procedure: <laser_setup.procedures.It.It>
#Parameters:
#   Chip group: Margarita
#   Chip number: 3
#   Sample: B
#   VDS: 0.075 V
#   VG: 15.0 V
#   ...
#Metadata:
#   Start time: 1718900000.0
#Data:
t (s),I (A),VL (V),Plate T (degC),Ambient T (degC),Clock (ms)
0.00,1.2e-06,0.0,25.0,23.4,1024
0.15,1.3e-06,0.0,25.1,23.4,1040
...

The header records every parameter value and metadata, so a run is fully reproducible — you can see exactly how it was taken.
The columns are the procedure's DATA_COLUMNS. Parameters listed in a procedure's EXCLUDE are omitted from the header.

Reading files programmatically

laser_setup.utils has helpers — read_pymeasure() splits header/data, and read_file_parameters() extracts just the parameter dict. PyMeasure's own Results.load() reconstructs a results object with a pandas DataFrame.

The Dirac point and `DP`¶

The IVg procedure estimates a device's Dirac point (the gate voltage of peak resistance) and other procedures can reference it. When you enter a gate voltage like DP + 15 V, utils.get_latest_DP() scans recent files for the selected chip/sample and substitutes the most recent measured Dirac point. If none is found, it falls back to 0 V. See Tutorial 2.

The parameters database¶

To analyze many runs at once, build a SQLite database from your CSV headers:

uv run laser_setup parameters_to_db

This walks data/*.csv, creates one table per procedure type, and inserts a row per file with all of its parameters (adding columns dynamically as new parameters appear). The database name comes from Dir.database (default database.db, relative to the data directory).

Browse it inside the app via View → Database, which opens the SQLiteWidget — a sortable, read-only table view. This makes it easy to find, say, every It run on a given chip at a given gate voltage.

Plots and styling¶

Live plots use pyqtgraph; exported/Matplotlib figures follow matplotlib_rcParams from the config (grid on, autolayout). The first two DATA_COLUMNS are the default x/y axes, configurable in the experiment window.

Logs¶

Two log streams are configured under Logging:

Console — colorized, level INFO, for live feedback.
File — log/laser_setup.log, level DEBUG, for post-mortem debugging.

Adjust levels in the Logging section of config.yaml. The live log is also visible in the GUI (experiment window Log tab, or View → Logs).