It's kinda nice to see a non-AI project on here.
Nice. Very light-weight compared to proper local routers like Graphhopper, OSRM, etc., which can be overkill for simple tasks. Although the 'routing' here is nx.shortest_path, which is just Dijkstra, so pretty slow compared to other easy to implement routing algorithms (even just bi-directional Dijkstra or A*... although contraction hierarchies would be huge gain here since edge weights are fixed). Also not sure why readme describes it as an approximation? Dijkstra is guaranteed to return lowest cost path. Maybe approximation because assuming free-flow, or if the NAR dataset is incomplete?
Thx for the heads up on optimizations available. The “Approximations” comment does not apply to the shortest path calculation, but rather to the distances and upper bound times estimations. This is the consequence of enabling routing for points that dont exist as nodes (closest node approximation).
From poking around in the source code I found this 282M SQLite database:
wget https://services.arcgis.com/xOi1kZaI0eWDREZv/arcgis/rest/services/NTAD_North_American_Roads/FeatureServer/replicafilescache/NTAD_North_American_Roads_3862439624850511818.geodatabase
Connection to NTAD_North_American_Roads_3862439624850511818.geodatabase failed check: no such module: VSRS
"This NTAD dataset is a work of the United States government as defined in 17 U.S.C. § 101 and as such are not protected by any U.S. copyrights. This work is available for unrestricted public use."
I based my work on this, maybe the link is out, thx for testing. The dataset has already been consumed and collapsed into a smaller graph representation.
if you follow the link in the github readme, there are a few other download options for the dataset, like CSV or shapefile
I’d love to see if I could assist in adding road type filters such as avoid multi lane highways for example
Contributions are welcomee!!
Came here to complain about US Telcos being willing to do anything other than enabling dynamic routing.
Glad to see this is for roads.
Edit: thanks very much for the suggestions, especially adding the Python version to the uv command line. I totally missed that, and that totally fixed it. Apologies for the OT tech support derailment.
--------------
Question for those familiar with uv. US Routing apparently requires a very specific Python version (3.11 and nothing else), but my system has Python 3.10.9 installed at the moment and I'd rather not upgrade the global version just now. My understanding from reading a lot of uv evangelism on HN and elsewhere is that uv fixes this type of dilemma. But, having just tried to use it to install this package, it's just giving me the same old Python version errors:
C:\devel\us-routing-master\us_routing>uv venv
Using CPython 3.10.9 interpreter at: c:\WinPython-31090
\python-3.10.9.amd64\python.exe
Creating virtual environment at: .venv
Activate with: .venv\Scripts\activate
C:\devel\us-routing-master\us_routing>.venv\Scripts\activate
(us_routing) C:\devel\us-routing-master\us_routing>uv pip
install us-routing
x No solution found when resolving dependencies:
`-> Because the current Python version (3.10.9) does not
satisfy Python>=3.11,<3.12 and us-routing==0.1.0
depends on Python>=3.11,<3.12, we can conclude that us-
routing==0.1.0 cannot be used.
And because only us-routing==0.1.0 is available and you
require us-routing, we can conclude that your
requirements are unsatisfiable.
By the looks of things, uv is telling you it’s creating a venv with Python 3.10.9 still. Since you didn’t specify you wanted another version, it probably defaulted to the first available system version.
What you want is `uv venv —python 3.11` to create a virtual environment with Python 3.11 without messing up your global system env. This should also install a portable version of Python 3.11 if needed (which it will be since you don’t have it).
https://docs.astral.sh/uv/pip/environments/#creating-a-virtu...
This library can use any version of Python 3.11, which you can install alongside your existing 3.10.9 without changing your global python version. I don't typically work in Windows so the codeblock below is AI generated, but follows the path I would normally take - manage installed python versions using pyenv, changing the python version only for this directory via .python_version, and then creating uv environment using that.
``` :: Step 1: Install pyenv-win (make sure Git is installed) git clone https://github.com/pyenv-win/pyenv-win.git "$env:USERPROFILE\.pyenv"
:: Step 2: Add pyenv to PATH (run or add to your profile) $env:PYENV = "$env:USERPROFILE\.pyenv\pyenv-win" $env:PATH += ";$env:PYENV\bin;$env:PYENV\shims"
:: Step 3: Restart your terminal or reload environment if needed :: (you can paste the above $env:... lines again after restart)
:: Step 4: Install Python 3.11 pyenv install 3.11.9
:: Step 5: Set the local Python version for your project folder cd C:\devel\us-routing-master\us_routing pyenv local 3.11.9
:: Step 6: Verify correct Python is selected pyenv which python # should point to 3.11.x
:: Step 7: Create uv environment using Python 3.11 uv venv .venv --python "$(pyenv which python)"
:: Step 8: Activate the environment .venv\Scripts\activate
:: Step 9: Install your package uv pip install us-routing ```
pyenv is a great way to have many versions of Python installed, whether or not your global is mapped to the latest. You don't even need to set the local .python_version.. you could just do `uv venv .venv --python=python3.11`
This is a project I did not maintain much for the past few months, but recently I migrated many other codebases to be uv managed, and I find this optimal for many reasons. Happy to receive contributions and fix the quite strict requirements I set. That would probably be the faster way.
You need to request a specific python version compatible with this project. Give `uv venv --python 3.11` a try.
You just created a venv but didn't change the Python version.
I imagine you'd want:
uv venv --python 3.11
Does it work for shorter distances? within a city from one business to another address?
it can be extended arbitrary given the proper dataset. For the current default settings, road class covers everything below 3. This translates to
FREEWAY = 1 # Freeway (Multi-lane, controlled access)
PP_TH = 2 # Primary Provincial/Territorial highway
SP_TH_MA = 3 # Secondary Provincial/territorial highway/ municipal arterial
MC_USP_TH = 4 # Municipal collector/Unpaved secondary provincial/territorial highway
LS_WR = 5 # Local street/ winter road
3 was the sweetspot. The dataset can be explored here in case you want to get an intuition on detail level. https://geodata.bts.gov/datasets/usdot::north-american-roads...
Routing library, having nothing to do with Google or Google maps.
does that account for 1% of gmaps functionality?
Graph traversal is a classical AI problem. See Ch. 3 from AIMA: https://aima.cs.berkeley.edu/index.html
I assume you mean non-LLM.
Graph transformers, using the same technology as in LLMs, have been a topic of research for many years, also for routing problems. So yes, it is non-LLM, but they could still have a lot in common.