A desktop app for planning fiber links and reviewing OTDR results. It works fully offline and is built with Python and PySide6 (Qt).
FiberDesk calculates link budgets and PON splitter cascades, shows OTDR traces with their event tables, and compares a line's planned loss with what the OTDR measured. It also includes a TIA-598 colour code guide. It is meant for field technicians, link-planning engineers and trainers.
Demo data only. Every project, line, measurement, operator and serial number is fictional. A fixed seed generates them. Device profiles are modelled on the EXFO MaxTester series. They are demo profiles, and the values are illustrative. They are not manufacturer specifications.
| Screen | What it does |
|---|---|
| Overview | Counts of projects, lines, measurements and devices; PASS / WARN / FAIL distribution for OTDR results and saved link budgets; recent measurements; per-project health. |
| Projects & Lines | Projects (an FTTH area, a backbone…) with their A–B lines: length, fiber type (G.652.D / G.657.A1), connectors, splices, splitter cascade, optics. Add / edit / delete, search and filters. |
| Link Budget | Tx power, Rx sensitivity, length, wavelength, splices, connectors, splitter cascade, safety margin. Shows the power budget, total loss, remaining margin and required safety margin on separate rows, the verdict and a loss breakdown chart. Results can be saved to a line. |
| OTDR Results | Measurement list with filters; a zoomable trace (distance km vs level dB) with event markers. Other wavelengths of the same line (1310 / 1550 / 1625 nm) can be overlaid. The event table shows type, loss, reflectance and section attenuation, coloured by threshold. Line budget vs OTDR loss is shown side by side. One-page PDF report. |
| Color Code | TIA-598 lookup (e.g. fiber 37 of 48 → tube 4 Brown, fiber 1 Blue) and the full tube × fiber grid for 12–144 fiber cables. |
| PON / Splitter | Cascaded split (e.g. 1:4 + 1:8). Shows OLT power down to subscriber power, total loss and the GPON class B+ / C+ verdict, with a power-along-the-path chart. |
| Devices | OTDR profiles (wavelengths, dynamic range) and the device inventory with calibration dates and measurement counts. |
| Settings | Every default value and threshold below can be edited; verdicts are recalculated immediately. |
 |  |
 |  |
 |  |
 |  |
Requires Python 3.12+ on Linux (developed on Ubuntu).
git clone https://github.com/asunode/fiberdesk.git
cd fiberdesk
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt
.venv/bin/python -m fiberdeskOn first start the app creates data/fiberdesk.db and fills it with the demo data set.
Useful options:
python -m fiberdesk --reseed # rebuild the demo data, then start
python -m fiberdesk --db /path/to/my.db # use another database file
python -m fiberdesk.storage.seed # rebuild the demo data without starting the UI
python -m fiberdesk --screenshots out/ # save a PNG of every screen and exitYou can also rebuild the demo data from Settings → Demo data. FIBERDESK_DB overrides the database path. An installed package (pip install .) stores its database in ~/.local/share/fiberdesk/.
| Item | Value |
|---|---|
| Fiber attenuation 1310 / 1490 / 1550 / 1625 nm | 0.35 / 0.24 / 0.22 / 0.25 dB/km |
| Fusion splice | 0.10 dB (WARN above 0.30, FAIL above 0.50) |
| Connector pair | 0.50 dB (WARN above 0.50, FAIL above 0.75) |
| Splitter 1:2 / 1:4 / 1:8 / 1:16 / 1:32 / 1:64 | 3.7 / 7.3 / 10.5 / 13.7 / 17.1 / 20.5 dB |
| GPON class B+ / C+ budget | 28 / 32 dB |
| Safety margin | 3 dB (remaining margin below this is WARN) |
| Reflectance | WARN above −45 dB, FAIL above −35 dB |
| Macrobend | always at least WARN, FAIL above 0.5 dB |
| Splitter tolerance | WARN above nominal + 1 dB, FAIL above nominal + 2 dB |
power budget = Tx power − Rx sensitivity
total loss = length × attenuation(λ) + splices × splice loss
+ connectors × connector loss + Σ splitter loss
remaining = power budget − total loss
verdict = FAIL if remaining < 0
WARN if 0 ≤ remaining < safety margin
PASS if remaining ≥ safety margin
The safety margin is the headroom a link must keep. It is a threshold, not a loss: it is never subtracted from the budget, and the screen shows it on its own row next to the remaining margin.
The OLT output power minus the losses along the path gives the power at the ONU. For each GPON class the remaining margin is the class budget (B+ 28 dB, C+ 32 dB) minus the total loss, graded with the same rule as the link budget: FAIL below 0 dB, WARN below the safety margin, PASS otherwise.
Each measurement is stored as a list of events (connector, splice, macrobend, splitter, fiber end) with distance, loss, reflectance and the attenuation of the preceding section. The event table is that list, and the trace is synthesised from the same list with a fixed seed, so the two always agree. A measurement's result is its worst event.
Expected vs measured compares the line's budget at the measurement wavelength with the OTDR link loss from the A to the B connector, category by category. If the measurement exceeds the expected total by more than 0.5 dB it is WARN; by more than 1.5 dB it is FAIL.
Demo traces use a simple model: backscatter slope, step losses, reflective peaks, an end reflection with a drop to the noise floor, and noise that grows with distance. Macrobends lose much more at 1550/1625 nm than at 1310 nm, as they do in real fiber.
.venv/bin/python -m pytestGitHub Actions runs the suite on every push and pull request (.github/workflows/tests.yml).
The suite covers the calculation core (budget, splitter / PON, colour code, OTDR events and traces), the database layer and demo seed, and an offscreen smoke test of every screen, including PDF export.
fiberdesk/
core/ calculation core — pure Python + numpy, no Qt
storage/ SQLite repository and the demo data generator
ui/ PySide6 window, pages, widgets, theme, PDF report
i18n.py every user-facing string (add a language here)
tests/ pytest suite
screenshots/ real screenshots of the app
FiberDesk does not connect to instruments and does not parse .sor / .trc files. There is no cloud, user account, web or mobile version.
MIT. EXFO and MaxTester are trademarks of their respective owner. This project is not affiliated with or endorsed by EXFO.