App running at https://mparismap.herokuapp.com/ .
- Streamlit app folder
- 00.PreparingData.ipynb (pre processing from opendatasets)
- 01.Stats.ipynb: initial insights (to be completed)
- 02.PreparingRouteFinder.ipynb: draft router finder
- 03.CastingShadows.ipynb: running tests with shadows mapping
- Zoning on arrondissements 18, 19, 20, 10, 11
- Been getting all routes from within OSM
- Preliminary mapping of best route based on trees
- Trees and lighting are put in
- Add shadows
- Use distance to fountains, parcs, green spaces
- What other layers to use?
- Develop something from building footprint (in data/BoI_simple.pkl)
- Example of the network
- Merging all buildings in one
- Details of streets
- Find
- Should get: X,Y start, X,Y end + set of preferences ("weights")
- Should receive, waypoints (nodes) (ordered list of XYs), and layers of interest close to the path (XYs of trees, lights, .. )
- Up: tells if the API is up
- Discover: sends end start users, settings - cache available
- Define
- Defining them: So to have a couple of from-to couples, and personas (ie prepared sets of weights)
- Store results in dataframe
- Run route finders and cache results
- Start
- End
- Weights
- Returns: json with waypoints, trees, lights, parcs, shadows
-
Streamlit:
- inputs
- map
- description of the map on the right
-
Personas + "you"
- Luc: home->work (day), then work->home night (shadows, light)
- Anna: montmartre->b&b in 19th : trees and lights
- Laure: South of 12th -> la villette
- Been preparing the code as in a python standalone module.
- Code as usual is in the "module" folder, here "pypama" (python paris maps)
- Version is generated from the module itself, as in the pypama/version.py file
- Adding binaries (the data pickles as part of the module in MANIFEST.in)
- Added a setup.py file to generate the module
- gitignore file prepared to exclude all "useless" bits from the repo
- Added some support scripts (
build
andinstall
to generate the module)
- Discovered how to generate a requirements.txt file for projects. Useful, and added in the
[createReqs.sh](createReqs.sh)
script. - From there, preparing a "TestingModule" notebook to try out changes in the module as the same time it's tested. The
%load_ext autoreload %autoreload 2
magic comes in handy to avoid reloading everything at the same time.
- Release e74e19ff
- Prepared the
99.app.py
launched bystreamlit run 99.app.py
. Opens up an app window to play around with weigths. - Text inputs ready for the search
- Formula in
calculatePath
still needs to be finetuned. - No API yet, just to play around with some of the layers. Caching still to be optimized =)
- v0.0.1 with workable version is released!
- Prepared the
-
FastAPI-based API branch pushed here
-
API-dependant streamlit app branch pushed here
-
@todo: pull the branches on mm80843, connecting two free servers from heroku
- Ongoing
- First workable version is released!
- Contains building footprints, shadows, lights, trees, green spaces, water points..
- Route finding is done!