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 teacher panel provides a set of real-time interventions that let you keep the activity flowing without interrupting the class narrative. All controls are scoped to a single selected team and take effect immediately — no page reload is required. Alongside these interventions, the panel exposes every team’s live scoring data and their submitted defenses, and gives you one-click export to JSON or CSV when the session ends.
Team monitoring table
The Seguimiento de equipos table is built from list_teams() in database.py. It joins the teams, mission_runs, and attempts tables so that every row is fully up to date. The columns visible to the teacher are:
| Column | Source field | Description |
|---|
| Equipo | name | Team name chosen at join time |
| Estado | phase | Current phase: chambers, missions, comparator, defense, or completed |
| Cámara | chamber_index | Chambers unlocked, displayed as x/6 (capped at 6) |
| Misiones | missions_completed | Distinct missions submitted, displayed as x/3 |
| Intentos | attempts | Total code submission attempts across all challenges |
| Pistas | hints_used | Cumulative hints requested (each costs 5 points) |
| Puntos | score | Current score |
| Holocrón | holocron_id | Title of the assigned holocron challenge, or — if not yet assigned |
st.dataframe() and hides the index. It refreshes automatically whenever the teacher triggers any intervention.
Available interventions
All interventions are grouped in the Intervenciones del Consejo section. Select a team from the dropdown, then use the action buttons.
Unlock next chamber
Function: unlock_next_chamber(team_id)
Advances the selected team to the next chamber without requiring a correct code submission. Internally this calls complete_chamber(), which adds +20 points to the team’s score and increments chamber_index. If chamber_index reaches 6, the team’s phase transitions to missions.
Use this when a team is genuinely stuck and you want to preserve the session’s pace rather than leaving them blocked at a single challenge.
# database.py — unlock_next_chamber
def unlock_next_chamber(team_id: str) -> None:
team = get_team(team_id)
if team and team["phase"] == "chambers":
complete_chamber(team_id, int(team["chamber_index"]))
The automatic five-failure path (triggered by the student portal, not the teacher) uses force_advance_chamber() instead, which deducts 10 points as a penalty. The teacher-side unlock via complete_chamber() does not apply that penalty.
Reset team
Function: reset_team(team_id)
Returns the selected team to a completely clean state. The function deletes all records associated with the team and resets the team row itself:
- Deletes all rows from
attempts, hint_usage, mission_runs, and defenses
- Sets
chamber_index = 0, phase = 'chambers', score = 0, hints_used = 0, holocron_id = NULL
A confirmation checkbox (Confirmar reinicio) must be checked before the Reiniciar equipo button becomes active.
# database.py — reset_team
def reset_team(team_id: str) -> None:
with connect() as db:
db.execute("DELETE FROM attempts WHERE team_id=?", (team_id,))
db.execute("DELETE FROM hint_usage WHERE team_id=?", (team_id,))
db.execute("DELETE FROM mission_runs WHERE team_id=?", (team_id,))
db.execute("DELETE FROM defenses WHERE team_id=?", (team_id,))
db.execute(
"UPDATE teams SET chamber_index=0, phase='chambers', score=0, "
"hints_used=0, holocron_id=NULL WHERE id=?",
(team_id,),
)
reset_team is irreversible. The team loses all progress, scores, submitted mission runs, and sealed defenses. There is no undo. Use only when a team genuinely needs to start over (e.g. wrong team name was used at join time).
Reassign holocron
Function: reassign_holocron(team_id, holocron_id)
Swaps the defense challenge assigned to a team. Seven holocron challenges are available, each with a different analytical focus:
| Holocron ID | Title | Focus |
|---|
consejo-desconfiado | Consejo desconfiado | Multi-metric evidence |
recursos-limitados | Recursos limitados | Complexity vs. performance |
auditoria-imperial | Auditoría imperial | Traceability |
senado-galactico | Senado galáctico | Non-technical communication |
lado-oscuro-metrica | El lado oscuro de la métrica | Per-class recall (digit 8) |
rescate-prioritario | Rescate prioritario | Per-class recall (digit 1) |
mision-produccion | Misión en producción | Production readiness |
Global hint
Function: set_global_hint(session_id, hint)
Broadcasts a plain-text message (up to 500 characters) to all teams in the session. The message appears in every team’s sidebar immediately. Use this for time announcements, conceptual nudges, or clarifications that apply to the whole class — for example:
“Remember: data known before training goes in Parameters, not Metrics.”
The hint is stored at the session level and overwrites the previous message. See Sessions for full details.
Exporting results
Export controls appear in the Copias de seguridad y exportación section at the bottom of the panel. Two formats are available, both generated by export_session() in database.py.
JSON export
Downloads a complete backup of the session as a UTF-8 encoded JSON file named {session_code}-resultados.json.
The JSON array contains one object per team. Each object includes all fields from list_teams() plus two nested keys:
runs — all mission_runs rows for the team, with metrics_json decoded to a metrics dictdefense — the team’s defenses row, or null if no defense was submitted
Use JSON export for archiving, post-session analysis, or feeding results into other tools.
CSV export
Downloads a UTF-8-BOM encoded summary table named {session_code}-resumen.csv with the following columns:
| Column | Description |
|---|
name | Team name |
phase | Final phase reached |
chamber_index | Number of chambers completed |
missions_completed | Number of missions completed |
score | Final score |
hints_used | Total hints used |
holocron_id | Assigned holocron challenge ID |
# database.py — export_session (CSV path)
fields = [
"name", "phase", "chamber_index", "missions_completed",
"score", "hints_used", "holocron_id"
]
writer = csv.DictWriter(output, fieldnames=fields)
writer.writeheader()
writer.writerows({key: team.get(key) for key in fields} for team in payload)
Export both JSON and CSV at the end of every session before closing it. The JSON gives you a full audit trail including all mission run metrics and defense text; the CSV is ready to paste directly into a gradebook.
Viewing defenses
The Defensas selladas section lists all completed defenses directly in the panel. Each defense expands into a detailed view showing every field the team submitted:
| Field | Description |
|---|
selected_run | The MLflow run ID the team chose as their recommendation |
evidence | Free-text justification for the selected run |
artifact | Name of the artifact the team consulted (e.g. confusion matrix) |
discarded_run | The run ID the team chose to reject |
discard_reason | Free-text explanation for the rejection |
limitation | An acknowledged limitation of the team’s decision |
recommendation | Final recommendation displayed as an info callout |