Skill: Strava for Freja.io

Purpose

Provide Strava-specific capabilities in a reusable skill module instead of hardcoded flow logic.

Setup

1. •Configure STRAVA_CLIENT_ID, STRAVA_CLIENT_SECRET, and STRAVA_REDIRECT_URI.
2. •Start Freja backend.
3. •In Telegram, run /strava connect.
4. •Complete OAuth in browser.
5. •Optional manual fallback: /strava connect <code>.

Commands

• •/strava status → show connection status and token metadata.
• •/strava connect → generate auth URL.
• •/strava disconnect → remove stored tokens and cache.
• •analysera min strava ... → run analytics report.

Analysis options

Text parser supports period and type filters:

• •Period: 7, 30, 90 days.
• •Type filter keywords: run, ride, walk.

Output contract

Telegram responses are always structured as:

1. •Svar:
2. •Sammanfattning
3. •Trender
4. •Highlights
5. •Rekommendationer

Mock mode

Use deterministic fixtures for local verification:

• •STRAVA_MOCK=1
• •STRAVA_MOCK_FIXTURE=mixed_run_ride or run_only_hr

Internal modules

• •storage.py: token/cache SQLite layer.
• •strava_auth.py: OAuth URL + code exchange + token refresh.
• •strava_client.py: endpoint client + retry + normalization.
• •strava_analytics.py: deterministic coaching analytics.
• •strava_commands.py: Telegram parsing + formatting.