ryanbaumann / athletedataviz Goto Github PK

Call activity API on each act download in stravaparse.py. store activity metadata in the database for additional styling and filtering.

In particular, store the Google polyline of the Activity. Change the query for v_stream_linestring to use the Google polyline summary data instead of the Stream data in order to increase query speed.

Multiplex Strava api calls via threading when user searches for new segments.

Branch name "Segments++ multiplex"

Partition the Stream table by ath_ID. Should allow "one indexed table" per athlete in order to greatly reduce the effective index size of the table when querying for heatmap points, reducing maintenence and increasing the ability to auto-archive old data.

Add drop down options for heat cad/power elevtion

Expose additional stream metadata for heatmapping.

Optimize query for segments++ data by returning all data as normal json in encoded polyline format. Then convert to geojson and pass to map on client side using mapbox-polyline library.

This should allow the entire segment layer to be returned in one query. This the map.moveend() and map.zoomed() query binds can be removed. Will greatly improve user experience on mobile and reduce number of requests served a ton.

In process of this, evaluate how to make "get new segments" a client side API call, further un-burdening the server from repeated API requests.

Create client side API call to get current leaderboard when user clicks on a segment. Style the leaderboard results into the pop-up.

Change PK Indexes to the Ath_ID, Act_ID, Seg_ID, and Stream_Id respectively. Drop the auto incrementing PK column. Should allow for lower disk space usage and increased memory efficient when making large queries.

Question: can we achieve this for the Stream table? Or should act_ID + Timestamp be the PK?

Image file upload to Server fails on Safari/IE due to File API support in those browsers.

Convert HTTP logic to use a manual file header parameter so all browsers can generate and save high res images.