All notable changes to this project will be documented in this file. Dates are displayed in UTC.
- Switch to PC tiler
#1
- turn off gpu placeholder for production
#616 - upgrade cert manager to try and solve ssl issues
#619 - Hotfix for SSL renewal
#621 - Patch User
#635 - available instances
#641 - Add instance type filtering
#640 - Generic Std
#637
- Fix position of model upload docs
#604 - Patch AOI request wasn't updating its classes in the database
#603 - tell tf to ignore db changes
#602 - Avoid div by 0
#601 - Fix tile 5xx
#600 - Make DELETE model endpoint actually delete it
#611 - Allow to update the active field of Model with PATCH
#610 - Model Total
#613 - Model Count Fix
#612 - Model Pagination
#608 - Admin List All Models
#606
- AOI Classes
#599
- Refactor OSMTag
#585 - Schemas
#582 - OSM Retrain Fix
#580 - Update Schemas
#577 - Add deeplabv3plus
#576 - read auth0 url from github secrets and pass into deployment
#572
- TagMap In ModelGet
#569 - Terminate instance when abort is True
#565 - gpu pod ttl 2h and run janitor every 1 hour
#567 - apidoc@0.50
#564 - Fix patch batch url in docs
#563 - Bug in Model TagMap
#562 - LibSchema
#559 - Make staging auth0 tenant default
#554 - upgrade pip in deploy Dockerfile
#556
- OSM Data Processing
#543 - GPU OSM
#550 - QA Tiles Proxy
#553 - Tile Geometry Type Filtering
#558 - Tile Parser
#555
- Housekeeping
#551
- AOI Patch BugFix
#548 - Change default Auth0 tenant endpoint
#546 - remove old azure docs. update deployment docs with new terraform setup
#544
- Fix PX_Stats
#517 - Small fix to patch list
#515 - Use base url from flight.request and fix type in docstrings
#510 - set max body size on API ingress to 300mb, refs #511
#513 - Use 403 error instead of 401 when the user doesn't fit the permission
#507 - Fix permissions in /share/:shareuuid/download/* endpoints
#505 - start count at 1 instead of 0
#482
- Objectify CheckPoint Class
#518 - Objectify Instance Class
#516 - Objectification of Model Class
#509 - Standard Azure Blob Storage
#508 - AOI
#502 - Ingress nginx
#503 - Remove model#abort
#501 - Add bookmarked_at field to AOI
#496 - Add Share Downloads
#500 - Fix Res Error
#499 - Update API documentation
#498 - Return 0 when there isn't active gpus running
#497 - give 10 gpus for staging
#494 - Global Retrain Test
#488 - Global Tests
#481 - switch to pytorch 1.9 + cuda 11.1
#415 - Lock ws to v7
#480 - Batch Error
#479 - Reduce Progress Updates
#478 - Lint All The Things
#477 - Lint Python
#476
- fix incorrect coalesce for active.false condition
#463 - Terminate after batch
#460 - Checkpoint Bug
#454
- Refactor Project Class
#459 - Routify Remaining Routes
#457 - make max inference 200km2
#458 - Node-PG => Slonik
#449 - Never delete
#444 - Dynamic Route Loading
#414
- set aoi maxzoom to 20
#439 - Add url_params & check aoi existance
#437 - sometimes stats aren't what's expected
#431
- Standard Dependency Maintenance
#413 - Retrain Fix
#416 - Embedding Fix
#418 - Ensure px stats exist before using
#423 - Use categorical endpoint
#429 - add statistics endpoint
#428 - Use sum of histo instead of px x,y
#427
- Pre-Commit + Linting
#398 - retrain metrics improvements
#403 - k8s upgrade 1.9
#417 - Filter by AOI
#424 - create nginx container to redirect to www
#419 - Persist user submitted retrain samples
#422
- Retrain Geom Fix
#386 - gpu shared volume memory
#367 - make tile caching longer and remove debug for production
#376
- Remove colormap that isn't needed for pre-colored Tiffs
#371 - AOI Share
#348 - Shares List
#361 - Batch inference
#346 - AOI Loading
#328 - Abort Retrain
#333
- Retrain Improvements: Decrease point sample from polygon
#357 - less memory intensive retrain
#364 - update TiTiler version to 0.3.1
#351 - Add the ability to load AOIs
#345 - return gpu availability with the config endpoint
#342 - make 15 gpus before using cpus on staging for testing
#332 - Use big placeholders
#331 - Nodeselectors + TTL updates
#325
- Fix Class Brush
#365 - Fix share call
#362 - Actually generate aoi share tile url
#366 - checkpoint edge case fix
#350 - Model Prediction Count up not down
#356 - remove duplicate getPodStatus method
#353 - dont make username a primary key
#341 - import mapping from shapely. possible merge conflict bug
#336 - Instance#Terminate
#334
- k8s instance state
#326 - Ping Pong
#319 - Fallback CPUs
#305 - sort and filter
#290 - Share aois
#281 - Cogify
#285 - AOI Patch Tiles
#287 - Patch List in AOI API
#286 - Class Brush
#282 - Clip Predictions
#257 - Debug & Abort
#283 - Add brush API
#262 - AOI Patch API
#256 - Tiff Colour
#255 - Abort & Status
#251 - Thread Spawn
#250 - Parent ID
#249 - Retrain + Checkpoint polygons
#245 - Bookmark AOIs
#238 - Dynamic Chkpt Loading
#232 - access seed data for retraining
#213 - Delete Instance
#226 - AOI Delete
#221 - Delete project
#216 - AOI ColourMap
#212 - Seed Data
#210 - Checkpoint Metrics
#146 - Terminate GPU pods
#129 - AOI TileJSON
#101 - Kube Config
#113 - Checkpoint Class Patch
#99 - Checkpoint Bookmarks
#98 - Retraining & Checkpoints
#91 - Add Knex for database migrations
#93 - AOI Name & Checkpoint ID
#97 - Project Tests
#96 - create GPU pods
#70 - Instance Timeout
#84 - Tiler Scaling
#81 - Future Models
#76 - Projects
#74 - CheckPoints
#66 - SSL Setup with Lets Encrypt
#62 - AOI Simplify
#65 - Model APIs
#63 - Ingress + App Gateway + Metrics
#43 - PNG Inference
#55 - Project
#54 - Enable Auth0
#38 - Staging & test deploys
#51 - add COG endpoints
#50 - k8s api
#26 - Auth0 Main
#44 - Prediction MVP
#40 - Model Server
#37 - Router
#34 - Shoe in Machine
#28 - Get model API
#20 - Enable CORS
#25 - Class Brush Scoping
86fb6c7