Bug: Command Center excludes zero-second completions from average solve time

[DevaanshPathak]

Component

• finbot/apps/cc/routes/challenges.py::_challenge_list_with_stats
• finbot/apps/cc/templates/pages/challenges.html

Problem

The Command Center challenge list excludes completed progress rows when completion_time_seconds is 0:

completed_rows = [
    p
    for p in progress_rows
    if p.status == "completed" and p.completion_time_seconds
]

Zero is a valid completion time. Both challenge-completion paths store the elapsed duration using int(total_seconds()), so any solve completed in under one second is recorded as 0. Because 0 is falsy, these solves are omitted from the average.

The template also uses a truthiness check for avg_solve_seconds, so a valid average of zero is rendered as -.

Steps to reproduce

1. Pass _challenge_list_with_stats a challenge with one UserChallengeProgress row where:
   • status == "completed"
   • completion_time_seconds == 0
2. Inspect the returned challenge statistics.

Actual result:

{
    "completions": 1,
    "players": 1,
    "avg_solve_seconds": None,
}

Expected result:

{
    "completions": 1,
    "players": 1,
    "avg_solve_seconds": 0,
}

Impact

Fast first-attempt solves are excluded from the average, biasing solve-time statistics upward. Challenges containing only zero-second completions show recorded completions but no average solve time in the Command Center.

Proposed fix

• Include completed rows when completion_time_seconds is not None rather than testing its truthiness.
• In the template, distinguish None from a valid zero value when deciding whether to render the metric.
• Add regression coverage for zero-second, non-zero, and missing completion times.

Acceptance criteria

• A completed row with completion_time_seconds=0 contributes to the average.
• Rows with completion_time_seconds=None remain excluded.
• A mix of 0 and 10 seconds produces an average of 5 seconds.
• The challenge list does not render a missing-value marker for a valid zero-second average.

[DevaanshPathak]

I reproduced this on current main and am working on a focused fix for zero-second average calculation and rendering, with regression coverage for zero, non-zero, and missing completion times.