Client Lambda API

The browser client calls the Lambda (lambda-resize) HTTP API for operations that require full frames or video: get-frame, get-movie-data (playback URL), tracking, rotate-and-zip, start-processing, new-frame. The VM (Flask/gunicorn) does not run tracking, serve frames, or serve movie playback URLs.

See ArchitectureDesign.md for the three principles: frames/video → lambda-resize, HTML → flask_app, metadata → flask_api.

Lambda API base URL

The server injects the Lambda API base URL into every page via base.html so it is available to all JavaScript as the global LAMBDA_API_BASE:

• Set in the template from get_lambda_api_base() (see apikey.page_dict()).

• Example value: https://your-lambda-api.example.com/

• If not configured, the variable is empty and client code should skip Lambda calls (e.g. show “Lambda URL not configured”).

All client calls to the Lambda API are authorized: the client sends api_key (and other parameters) in the request; the Lambda validates api_key (e.g. via DynamoDB) before performing the operation.

Endpoints used by the client

Operation	Method	Path / body	Purpose
Get frame	GET	/api/v1/frame?api_key=...&movie_id=...&frame_number=...&size=...	Single frame as JPEG
Get movie data	GET	/api/v1/movie-data?api_key=...&movie_id=...&format=zip or format=json (optional)	Default: 302 redirect to signed S3 URL (movie). format=zip: 302 to zip. format=json: 200 JSON { "movie_id", "url" (MP4), "zip_url" (if present) } for playback or download links.
Status	GET	/status	Health check
Track movie	POST	/api/v1	Body: { "action": "track-movie", "api_key", "movie_id", "frame_start" }
New frame	POST	/api/v1	Body: { "action": "new-frame", "api_key", "movie_id", "frame_number", "frame_base64_data"?(optional) }
Rotate and zip	POST	/api/v1	Body: { "action": "rotate-and-zip", ... }
Start processing	POST	/api/v1	Body: { "action": "start-processing", ... }

Tracking

Tracking is requested by POSTing to LAMBDA_API_BASE + 'api/v1' with JSON body:

• action: "track-movie"

• api_key: from the logged-in user (same as other API calls)

• movie_id: movie to track

• frame_start: the edited source frame (0-based). PlantTracer preserves this frame, deletes stored annotations for frame_start + 1 ... end, and resumes tracing at the following frame.

The Lambda validates api_key, resolves user_id, then runs tracking (writes trackpoints and zip to DynamoDB/S3). The client then polls for completion (e.g. via Flask get-movie-metadata with get_all_if_tracking_completed) and updates the UI.

Where the variable is set

• Template: src/app/templates/base.html — declares const LAMBDA_API_BASE = "{{ lambda_api_base }}";

• Server: src/app/apikey.py — page_dict() sets lambda_api_base from get_lambda_api_base() so every page that uses the base template gets the same global.