Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/elenacarino-max/Pildora4_ext_StarWars/llms.txt

Use this file to discover all available pages before exploring further.

The practice notebooks in notebooks/02_practica/ provide structured exercises that go beyond the live demonstration and give students the opportunity to apply MLflow tracking independently. Where the live coding notebooks follow an instructor’s narration, the practice notebooks are designed for self-directed work: students encounter a problem, attempt a solution, check against the reference, and finally tackle an open-ended challenge that mirrors the hardest mission in the Streamlit app. The three notebooks form a deliberate progression — guided exercises, worked solution, and autonomous challenge — matched to the same escalating difficulty as the Tatooine → Coruscant → Mustafar mission arc.

Available notebooks

01_Tracking_Ejercicios.ipynb

Initial tracking exercises. Students practice each MLflow call in isolation and in combination: registering an experiment, opening a run with a context manager, logging a parameters dictionary, logging a metrics dictionary, saving artifact files, and logging the trained sklearn model. Each exercise includes a brief prompt and a blank cell for the student’s solution.

02_Archivos_Jedi_SOLUCION.ipynb

Complete solution reference. Every exercise from 01_Tracking_Ejercicios.ipynb is solved with working, annotated code. Instructors use this to check student work, provide targeted hints, or display on screen when a class is stuck. Students can consult it after a genuine attempt — using it as a first resort defeats the purpose of the exercises.

03_Orden_66_RETO.ipynb

The final challenge — Orden 66. Students apply all six MLflow concepts autonomously in a single, unsupported notebook. The challenge mirrors the Mustafar autonomous mission in the app: no scaffolding, no hints, and the full 5–4–2–1 tracking pattern must be produced from memory. Completing it successfully is the clearest evidence that a student has internalised the complete tracking workflow.

Progression

The three notebooks follow the same escalating difficulty arc as the app missions:
NotebookEquivalent app stageSupport level
01_Tracking_Ejercicios.ipynbChambers 1–6 + TatooineGuided — prompts provided
02_Archivos_Jedi_SOLUCION.ipynbInstructor referenceFull solution visible
03_Orden_66_RETO.ipynbMustafar (autonomous)None — blank canvas
Students who work through all three notebooks in order will have seen each concept at least twice in a notebook context and once in the gamified app context before attempting the final challenge.

Using notebooks alongside the app

The practice notebooks use the same Digits dataset, the same Random Forest algorithm, and the same five-parameter dictionary (dataset, n_estimators, max_depth, test_size, random_state) as the Streamlit app. As a result, the metric values produced in the notebooks are directly comparable to the orientative results shown in the app’s Comparador and Consejo screens. Students who complete both the app missions and the practice notebooks will notice the numbers aligning — reinforcing the idea that MLflow is a reproducible record of a real, deterministic experiment. The shared parameter schema also means a student can copy the parametros dictionary from a notebook cell directly into the app’s mission editor (or vice versa) without modification.

Running the notebooks

Install the required packages and launch JupyterLab from the app-v2 directory:
cd app-v2
pip install -r requirements.txt
jupyter lab  # or jupyter notebook
The notebooks are located at notebooks/02_practica/ relative to the repository root. Navigate there in the JupyterLab file browser after launch. No additional configuration is needed. MLflow will create a local SQLite tracking database and an mlruns/ artifact folder automatically on the first run, in whichever directory the notebook kernel is started from.
03_Orden_66_RETO.ipynb is the notebook equivalent of the Mustafar mission (n_estimators=300, max_depth=None). If a student completes it successfully — producing a valid run with five logged parameters, four logged metrics, two artifact files, and a serialised model — they have demonstrated mastery of the full 5–4–2–1 Jedi Code tracking pattern. This notebook makes an effective exit-ticket or take-home assessment for the session.

Build docs developers (and LLMs) love