Colouring cities kick off meeting 5th January 2021

# Colouring cities kick off meeting 5th January 2021 ## Proposed REG tasks - Some more specific details than the below in [this issue](https://github.com/colouring-london/colouring-london/issues/680) - Server VM setup/management - Review what has already been set up on - Improve the automation of (re)provisioning a VM for London and prepare whole VM template for other cities. - Automate CI/deployment pipelines: - Build and release from GitHub to staging/production through GitHub Actions (currently only a CI task is set up) - Think about some automation of database schema updates and better db migrations setup (currently all manual / git-based) - Improve testing setup (currently only a few unit tests) - Help set up an integration testing environment/pipeline (would help greatly for automated dependency updates) - Help set up front-end testing - Tile server improvements - the tile server is a custom implementation in Node+Mapnik due to having to render building data visualisation as raster tiles because of OS licensing. - Review if possible to use a more off-the-shelf solution that would help with the below points, or - Improve tile cache control, build utilities for easier clearing of cache for selected layers, regions, etc - Add some request priority control, which would help with keeping the app smooth (as even things like building selection highlight is sent as raster, again due to OS limitations) - Build a way for less technical users to easily edit data map styles. Currently works by editing Mapnik XML files - sometimes hard even for programmers. E.g. look at whether it's possible to combine the raster tiles requirement with a style specification like the Mapbox JSON, which can be edited with GUI tools like Maputnik ## Notes Generating data computationally through inference (AI) - not front-end - but back-end management or data merging/inference? - We are trying to infer things that the tax office has on buildings based on age, then get citizens (locals) to check these things (crowdsourcing) - The site is a tool for the crowdsourcing, like with openstreetmap, the dataset is the product - relation to the historial maps project - layer one or two historical maps into the system - proving a point that open data is worth it - Meet with Tom on technical side of things - Working practices, github, slack etc - Mateusz on front-end updates for new data requirements - Is all data crowdsourced apart from the Ordnance Survey polygons? # Meeting 11th January ## Questions - What is Tom/Mateusz role for the project duration? - Minimal, can giver overview - Mateusz (still working on, not regular) and Matcheck, worked on dev work more recently - **EC:** Get Tom Russel or Mateusz Konieczny to add EC as contributor to [repo](https://github.com/colouring-london/colouring-london) - **EC:** Also get one of them to show EC VM setup/management (Azure?) - **EC:** Should I use an existing project board? - If not I will create GH issues for above tasks and set up project board (follow [template](https://github.com/colouring-london/colouring-london/projects) in the colouring-london repo) - Prioritise and have milestones (dates) signed off by team - I should probably set up my own project board and break down Polly's - Might be worth sorting through the open issues in a session - Milestones aren't useful and could probably be deleted - Think about GH branches, contributor guidelines - Set up blocks to mergeing - Document this and so we can have release branch, especially if there is a Terraform ## Notes - Room for improvement around integration and deployment, so currently, there is VM with the repo cloned and setup carried out - Terraform and Ansible setup scripts would be super valuable but not urgent - Would be useful to help me understand the setup - I should have access to the VM in Turing - get access to colouring London subscription at Turing Azure - setup of website: https://github.com/colouring-london/colouring-london/blob/master/docs/setup-production-environment.md - deploy template: https://github.com/colouring-london/colouring-london/issues/new?assignees=&labels=&template=deploy.md&title= - ^ template when you make a new issue - VMs each have IP addresses, might need to get Azure to manage the DNS record to associate IP with domain (test with staging) - Big priority task: updating the polygons, with the latest OS (2021), currently 2016 - first task after familiarisation with codebase - Include method to auto-update/build the colouring london dataset with new OS releases - Take the OS polygons as buildings (ignore the philosophical implications) ## EC TODOs week 10th Jan: 1. [ ] Run colouring London locally 2. [x] Request to be in colouring London Azure sub with IT 3. [ ] Log into the VM and see what the commit of the repo the setup was done at 4. [ ] Propose in the Friday meeting to have a mega sort-through session of issues in order for me (Ed) to have a project board for the 8 months (any timelined milestones, but ideally just project board unless there are specific deadlines for things) 5. [ ] Perhaps try a fresh setup on another VM? (this will help with the Terraform-ing) 6. [ ] Should we set up github discussions ## Meeting 14th Jan - Only Matchek as access to the Azure, Polly should email instroducing us, he is no longer funded on the project - doodle for the 2hr session - Look at Mateusz PR? - using the OSM for testing - Mateusz has docs for the VMs setup - The terraform stuff can be based on stuff in here: https://github.com/colouring-london/colouring-london-config - there's also a vagrant script: https://github.com/colouring-london/colouring-london/blob/master/provision/vm_provision.sh - Documentation could be a big thing to work on, but not the readthedocs as no docstrings - There is a wiki (no info) that could be for entire projects - gitbook maybe? Have a look at the options, needs to be not too hard for Polly to edit - Maybe look at the issues and add the below if not already there: Main priorities: - Documentation in the README and docs - Find everything from readme - Docstrings wherever possible - *gibook for a later date* - Updating the setup (terraform) - Updating the polygons