nheri-simcenter / brails Goto Github PK

Hi. I just installed the brails package and tried the example workflow:

      `from brails.InventoryGenerator import InventoryGenerator
      
      invGenerator = InventoryGenerator(location='Berkeley, CA',
                                        fpSource='usastr',
                                        baselineInv='nsi',
                                        lengthUnit='m',
                                        outputFile='BaselineInvBerkeleyCA.geojson')
      
      invGenerator.enabled_attributes()
      
      
      invGenerator.generate(attributes='all',
                            GoogleAPIKey='mykey',
                            nbldgs=100,
                            outputFile='BldgInvBerkeleyCA.geojson')
      
      invGenerator.inventory`

I got a GoogleAPIKey and ran the code. However, I get this error:
AttributeError: 'FootprintHandler' object has no attribute 'enabledBaselineInvInputTypes'

Hello,

What part of the code does planar rectification?

Thank you.

I encountered an HTTP 404 error while using the InventoryGenerator module in BRAILS. The error occurs when the script tries to download a model file from the URL https://zenodo.org/record/10448047/files/facadeParserv2.pth.
The URL causing the error is located in line 199 of 'FacadeParser.py'

The following warning is raised when running CityBuilder::build:

BRAILS/brails/CityBuilder.py:185: UserWarning: Geometry is in a geographic CRS. Results from 'centroid'
are likely incorrect. Use 'GeoSeries.to_crs()' to re-project geometries to a projected CRS before this operation.

If this is accounted for and the result is known to be correct, the following page can be useful for singling out and silencing the warning:
https://docs.python.org/3/library/warnings.html

When I initialize the ConstructionTypeClassifier mode, it tried to download the checkpoint from this url: https://zenodo.org/records/7271554/files/trained_model_constype.pth
And it shows that the page is not found

Would you mind sharing the checkpoint?
Or how many images do I need to prepare for each category to train a good model?

Several functions are defined with default values for required parameters (eg CityBuilder constructor).

This is leading to hard-to-find bugs and greatly impacts usability.

premRefline calculations occasionally results in NaN or Inf slopes or data vertices. Current patch is just a temporary fix.

for cno,cluster in enumerate(principalDirSegments):
---> xp = np.zeros((2len(cluster)))
yp = np.zeros((2len(cluster)))
for idx, segment in enumerate(cluster):

TypeError: object of type 'numpy.int64' has no len()

Lack of API key validation is leading to cryptic errors (eg. PIL.UnidentifiedImageError).

in Line 107, for the variable indNot calculation, the denominator has slopeSeg instead of slopeSeg[0], as in the calculation of ind. Is this a mistake?