streetlevel

streetlevel is a library for downloading panoramas and metadata from street-level imagery services such as Google Street View, Apple Look Around, and several others. It provides a simple abstraction over the internal APIs of the supported services – this means that no API keys are required, but the library may break unexpectedly.

(Nearly) all functions are available as either a sync function using requests or an async function using aiohttp, requiring a ClientSession.

Installation

streetlevel is available on PyPI:

pip install streetlevel

macOS users may need to install additional dependencies:

brew install gettext
brew install inih

Example

Downloading the closest Google Street View panorama to a specific location, sync:

from streetlevel import streetview

pano = streetview.find_panorama(46.883958, 12.169002)
streetview.download_panorama(pano, f"{pano.id}.jpg")

Or async:

from streetlevel import streetview
from aiohttp import ClientSession

async with ClientSession() as session:
    pano = await streetview.find_panorama_async(46.883958, 12.169002, session)
    await streetview.download_panorama_async(pano, f"{pano.id}.jpg", session)

Documentation

Documentation is available at streetlevel.readthedocs.io.

Functionality overview

Services covering multiple countries are on the left; services covering one specific country are on the right.

✔ implemented / available; 🟡 partially implemented; ❌ not implemented; ⚫ not available / not applicable

1: Returns closest only
2: Tile, z=17
3: Bounding box
4: Unstitched
5: Appears to be a synthetic depth map created from elevation data and building footprints
6: Month and year only for official coverage, full date for inofficial coverage
7: Only heading; pitch/roll do not appear to be available
8: Previous and next image in sequence
9: Month and year only
10: There is a has_depth field in the raw metadata, but I've yet to find a panorama that actually has depth
11: Pitch/roll are only available for the new 3D imagery
12: Camera altitude is available, however
13: 3D imagery panos have both an equirectangular and a cubemap version; everything else is only available as cubemap
14: Non-3D imagery appears to use a synthetic depthmap created from elevation data and building footprints