ScyIO UnB

Scientific Input / Output — graphical interface
Digital Signal Processing Group · University of Brasilia (UnB)

ScyIO UnB helps you read, inspect, convert, and merge scientific data files between common formats used in signal processing and machine-learning workflows. The interface has four working modes in the left column: Read, Convert, Batch, and Merge files. Below those modes, Configuration lets you customize colors and style.

Launch the application with python ScyIO_UnB_GUI.py (Spyder or a terminal), or run the Windows executable. Conversions and merges run in the background; watch the status bar and the log panel for progress.

1. Modes overview

Mode	Purpose
Read	Open one data file, list its variables, and preview values in the Inspector panel.
Convert	Convert a single file to another format and write it to a destination folder (same base name, new extension).
Batch	Convert all matching files in a folder to a chosen target format.
Merge files	Combine all matching files of one format in a folder into a single output file (`.mat`, `.parquet`, or `.h5`). Unlike Batch, Merge does not create one output per input file.

2. Supported formats

CSV — .csv tables (with or without header)
MATLAB — .mat
HDF5 — .h5 / .hdf5
Apache Parquet — .parquet / .pq
Raw binary signals — typically .sig / .SIG (requires sampling frequency and sample type)

3. Read mode

Select Read in the left column.
Drop a file onto the drop zone, or use Browse… / paste a path and click Open.
The detected format appears as a chip. Variables are listed in the Inspector on the right.
Click a variable name to see shape and a short numeric preview.

Read mode is for structured formats (CSV, MAT, HDF5, Parquet). Raw .sig files are handled in Convert / Batch, where you must supply fs and sample type.

4. Convert mode (single file)

Select Convert.
Choose the source file (drop zone or Browse).
Pick a Target format from the list of conversions available for that source type.
Set the Destination folder (default: d:/Dropbox/Python_Imports/HDF5/).
Click Convert file. The log shows the output path.

Available single-file conversions

Source	Targets
CSV	MATLAB (.mat), HDF5 (.h5), Parquet (.parquet)
MATLAB (.mat)	HDF5 (.h5), Parquet (.parquet)
HDF5 (.h5)	Parquet (.parquet)
Parquet	HDF5 (.h5)
Raw (.sig)	MATLAB (.mat), HDF5 (.h5), Parquet (.parquet)

Raw signal parameters (Convert)

When the source is a raw binary file (.sig), extra fields appear:

fs (Hz) — sampling frequency (example: 1024)
type — int → fixed-point int16, or float → floating-point float32

Output keeps the original stem and adds the new extension, e.g. DS01IS01.SIG → DS01IS01.mat.

5. Batch mode (folder)

Select Batch.
Choose the Source folder.
Set Source type (CSV, MATLAB, HDF5, Parquet, or Raw).
Choose Target format and Destination folder.
Optional toggles:
- Recursive — also convert files in subfolders
- Keep folder tree — preserve relative paths under the destination (useful for large MATLAB trees with duplicate stems)
- Skip invalid — continue if some files cannot be converted (recommended for mixed CSV folders)
Click Convert folder. The log lists written files.

Raw signal parameters (Batch)

When Source type is Raw (.sig), you must also set:

fs (Hz) — same for every file in the folder
type — int (int16) or float (float32)
extension — raw extension shared by all files (example: sig; case is ignored)

Example: convert all Isokinetic .SIG files with fs = 1024, type = int, extension = sig to .mat, .h5, or .parquet in the destination folder.

6. Merge files mode (folder → one file)

Use Merge files when you want every file of a given format in a folder written as one combined file. The button is in the left column, directly below Batch.

Batch vs Merge

	Batch	Merge files
Input	Many files in a folder	Many files in a folder
Output	One converted file per source file	A single combined file for the whole folder
Formats	CSV, MAT, HDF5, Parquet, Raw → chosen target	Same format in and out: MAT→MAT, Parquet→Parquet, HDF5→HDF5

How to merge

Select Merge files.
Set Source folder — directory that contains the files to combine.
Choose Source format (see table below). The Target name extension is updated automatically.
Edit Target name if you want a custom output filename (it must end with the correct extension).
Set Destination — folder where the merged file will be written. The final path is destination_directory / target_name.
Optional: enable Recursive to include matching files in subfolders.
Click Merge folder. Progress appears in the merge log and status bar.

Fields

Field	Meaning
Source folder	Folder that holds the source files (e.g. all `.mat` files).
Source format	Which file type to collect and merge.
Target name	Name of the single output file, including extension (`.mat`, `.parquet`, or `.h5`).
Destination	Folder where that output file is saved.
Recursive	If checked, also merge matching files found in subfolders.

Supported merge operations

Source format	What is merged	Default target name
MATLAB (.mat)	All `.mat` files → one `.mat`	`merge_files.mat`
Parquet (.parquet)	All `.parquet` files → one `.parquet`	`merge_files.parquet`
HDF5 (.h5)	All `.h5` / `.hdf5` tables → one `.h5`	`merge_files.h5`

Merged Parquet and HDF5 tables are written so they can be opened again in Read mode (and inspected in the GUI). Rows are stacked; columns such as source_name / source_path identify which original file each row came from (when available).

Examples

Example 1 — merge all MATLAB files

Source folder: D:/Dropbox/Python_Imports/TESTE/MAT
Source format: MATLAB (.mat)
Target name: merge_files.mat
Destination: D:/Dropbox/Python_Imports/TESTE
Result: D:/Dropbox/Python_Imports/TESTE/merge_files.mat

Example 2 — merge all Parquet files

Source folder: D:/Dropbox/Python_Imports/TESTE/PARQUET
Source format: Parquet (.parquet)
Target name: merge_files.parquet
Destination: D:/Dropbox/Python_Imports/TESTE
Result: D:/Dropbox/Python_Imports/TESTE/merge_files.parquet

Example 3 — merge all HDF5 files

Source folder: D:/Dropbox/Python_Imports/TESTE/HDF5
Source format: HDF5 (.h5)
Target name: merge_files.h5
Destination: D:/Dropbox/Python_Imports/TESTE
Result: D:/Dropbox/Python_Imports/TESTE/merge_files.h5

7. Configuration (colors and style)

Select Configuration (directly below Merge files) to customize the GUI appearance.

Colors — click a swatch to pick a new color for backgrounds, panels, accents, text, and highlights.
Style — font family, font size, and Qt style (e.g. Fusion).
Preview — apply the choices immediately without writing a file.
Save — apply and write ScyIO_UnB_GUI_config.json in the same folder as the application. On the next startup the GUI loads this file automatically.
Reset — restore the built-in default theme and delete the saved configuration file.

8. Inspector and status

Inspector (right) — summary, variable list, and value preview after a successful Read (including merged files opened in Read mode).
Status bar (bottom) — idle / working / done, plus a progress indicator while a job runs.
Only one job (convert, batch, or merge) runs at a time. Wait for it to finish before starting another.
Exit (left column) closes the application. If a job is still running, you will be asked to confirm.

9. Tips

Destination folders are created automatically if they do not exist.
Existing outputs are overwritten by default.
For MATLAB folders with nested experiments in Batch, enable Recursive and Keep folder tree.
For CSV folders that mix feature tables and report files, leave Skip invalid checked in Batch.
Raw conversion always needs the correct fs and sample type; wrong settings produce incorrect waveforms.
In Merge files, keep source files of one format in the source folder; the output extension must match the selected source format.
After a merge, use Read on the new file to list variables and preview values in the Inspector.

10. Related library

The GUI calls the Python library ScyIO_UnB.py. Scripted examples are in ScyIO_UnB_Test.py, including:

Example 10 — merge all .mat files into one .mat
Example 30 — merge all .h5 files into one .h5
Example 31 — merge all .parquet files into one .parquet