Windows launchers: run the project by double-clicking¶
These files let a member run the project on a Windows computer without git or typed commands. They work from a ZIP download of the repository. The full walkthrough is recipe 7 in docs/NO_GIT_GUIDE.md.
Before the first run¶
- Install Python 3.11 or newer from python.org/downloads. On the first installer screen, tick Add python.exe to PATH.
- Download the project: on the repository page, green Code button, Download ZIP. Right-click the ZIP, Extract All, and extract to a folder outside OneDrive, for example
C:\dev\. - Open the extracted folder, then this
windowsfolder.
If Windows shows "Windows protected your PC", click More info, then Run anyway. The files are plain text; you can open any of them in Notepad to read exactly what it does.
The files, in order¶
| File | When to use it | What it does | Changes |
|---|---|---|---|
1-set-up.bat |
Once, before anything else; again after downloading a newer ZIP | Finds Python, creates the private .venv folder, installs the project's packages, downloads the GitSkills sample (about 90 MB) |
Adds .venv and data/samples inside the project folder only |
2-run-pipeline.bat |
Fresh-run check for a sprint review (S1-10, gemba walk) | Reruns the full analysis, compares every result file with the committed one (line endings ignored), and opens a report in Notepad | Writes build/fresh_run/ only; results/ is not changed |
3-open-labelling-page.bat |
Second raters, at the start of each labelling session | Asks your rater id and kind, builds your label sheet and reading packets, and opens your labelling page in the browser | Writes your sheet in data/annotations/ if it does not exist yet, and pages in data/annotations/work/ |
4-save-my-labels.bat |
Second raters, at the end of each session | Checks the CSV you downloaded from the labelling page and saves it as your label file. Drag the CSV onto this file; a name your browser changed, like drift_la_r1 (1).csv, is fine. If the download changes labels you saved before, it asks before replacing them |
Updates your own label file in data/annotations/ |
Every window says what it is doing, and stops with a message if something fails. If one fails, take a screenshot and post it as a comment on your task on GitHub.
Common problems¶
| What you see | Why | What to do |
|---|---|---|
| "Windows protected your PC" | The file came from the internet | More info, then Run anyway |
| "Python 3.11 or newer was not found" | Python is missing, or was installed without the PATH option | Install it again from python.org and tick Add python.exe to PATH |
| "No such file or directory" during install, with a hint about long paths | Windows refuses file paths over 260 characters, and some packages have deep folders | Move the project folder to C:\dev\, delete the .venv folder inside it, and run 1-set-up.bat again |
| The window closes at once | It was started from inside the ZIP | Extract the ZIP first (right-click, Extract All) |
What the fresh-run report means¶
2-run-pipeline.bat exits with one of three results:
- Every file matches. Your computer produced the same numbers as the committed results. That is the measurement system check passing.
- Some files differ. The report lists which ones. Post the report on your task; a difference is a finding, not your mistake.
- Something went wrong. Usually the sample is missing (run
1-set-up.bat) or the computer ran out of memory (close other programs and try again).
The same check from a terminal: python scripts/fresh_run_check.py.