Offline Map Field Logger

Project roadmap

1. 01

   Design the app

   ~2.5h

   Write DESIGN.md: a track has an id (UUID), start and end, and points (latitude, longitude, accuracy, altitude and time); a note has an id, position, text, optional photos and time; each item has an upload status. Describe the upload API (POST /batches with an idempotency key, returning the ids accepted), the batch size, retry rules, and how recording keeps going when the app is in the background. Done when: DESIGN.md covers data, recording, uploading and failure cases.

2. 02

   Record GPS tracks

   ~3h

   Use the geolocator package to request permission (explaining why first) and stream positions with a distance filter and accuracy setting chosen for battery life. Save each point to Drift immediately, dropping points with poor accuracy or impossible jumps. Keep recording in the background (with a foreground service notification on Android and background location on iOS). Done when: a walk is recorded completely, even with the screen off, and points are in the database.

3. 03

   Add notes with photos

   ~2h

   Let the user add a note at the current position with text and photos (copied into the app's storage and shrunk). Show tracks and notes on a map (for example with flutter_map and OpenStreetMap tiles, cached for offline use where allowed). Done when: notes appear on the map at the right place, with their photos.

4. 04

   Upload in batches

   ~3h

   Write an upload service that collects unuploaded items, splits them into batches (points in chunks, photos separately), sends each batch with an idempotency key so a retry never creates duplicates on the server, and marks items uploaded only after the server confirms. Retry failures with increasing waits, and upload only on Wi-Fi if the user prefers. Write a small server to receive batches. Done when: a day of offline recordings uploads fully once the phone is online.

5. 05

   Recover from interruptions

   ~2h

   Handle the app being killed while recording (resume or close the track cleanly on restart), uploads interrupted halfway (resent safely thanks to the idempotency keys), a full disk, and GPS turned off (tell the user). Show the pending upload count and last upload time. Done when: killing the app during recording or uploading loses nothing and duplicates nothing.

6. 06

   Manage state with Riverpod

   ~1.5h

   Organize the app with Riverpod providers for the location stream, the recording state, the database, and the upload status, so screens only read state. Add a track list with distance, duration and upload status, and a way to export a track as a GPX file. Done when: every screen reflects the current state, and GPX export works.

7. 07

   Test it

   ~2.5h

   Write tests with a fake position stream (including noisy and jumping points), an in-memory Drift database, and a fake server that fails, times out or accepts: filtering bad points, track distance, batch splitting, retries and idempotency, and recovery after a simulated crash. Test recording outdoors and note battery use in README.md. Done when: all tests pass, and README.md describes real-world results.

Resources

• Docsgeolocator package
• DocsDrift documentation
• DocsRiverpod documentation
• Docsflutter_map documentation
• ArticleGPX format (Wikipedia)