Switch to pyproject.toml and python 3.10+

We change the project dependency manager from requiresments.txt to uv generated pyproject.toml, a manager 100+ times faster than pip.

Breaking updates

• We switch the python version requirements to python 3.10+

Enhancement

• Support OSS download for China Mainland

Support Metashape multi-spectral projects

Enhancement

• Support the metashape projects of multi-spectral cameras
• idp.shp.show_shp_fields add line number
• Support read ROI from GeoJSON
• Tolerate non-consecutive camera IDs in Metashape doc.xml

Bug fixs

• handel metashape one sensor with multiple calibration tags
• Support read_shp / read_geojson with name_field=-1 or '#' to set index
• pyproj 3.6.1 changed the path to get unit
• fix xy reverse from CRS definition
• handle Metashape duplicate the same sensor label with different sensor attributes
• Fix numpy deprecates np.float data type
• Metashape project unable to load photos in multiple subfolders
• Bug with mode="point" for ROI.get_z_from_dsm()

Testing

• Add GitHub actions and the readme badge for running test when pull request
• Remove the Lotus Dataset dependicy for pytest, use TestData only

Provide fully API documentation

Bug fixs

• cvtools.imarray_crop turn 0-1 float rgb to blank uint8
• cvtool.imarray_crop cannot handle roi with float
• Fix roi[a:b] slice lose inner functions
• Better notice if ROI has no z values
• ROI duplicate read z values
• Better notice when DSM did not cover all ROI, return Nan values

Enhancement

• Support GeoTiff object pixel2geo & geo2pixel
• Container Class support slicing funtion
• Add progress bar when get_z_from_dsm
• Filter the order of back2raw by plots to camera center
• Metashape chunk_id should also work by chunk.label.
• Cache get_photo_position results for next use
• Beautiful print Container objects
• Provide visualize function to preview back2raw results
• Remove pix4d project image suffix
• save back2raw results to json file
• Save back2raw images to png file with ROI as alpha layer

API modify

• easyidp.PointCloud
  • set_offset_value() -> update_offset_value()
  • crop() -> crop_rois()
• easyidp.GeoTiff
  • crop() -> crop_rois()
• easyidp.shp
  • convert_proj() -> easyidp.geotools.convert_proj()
• easyidp.metashape
  • convert_proj3d() -> easyidp.geotools.convert_proj3d()
  • _is_single_point() -> easyidp.geotools.is_single_point()

Documentation

• Provide full API documentation for current functions
• Provide small demo docs for each function

Support update dataset and fix usage bugs

Enhancements

• re-arrange test data
• force to redownload files for dataset

Bug fixes

• Better error notify when ROI is too small than one pixel
• Backward Projection with Metashape crs changed according to roi's crs
• Support obtain cam_position on Geotiff.crs directly

Rebuild almost functions from v1.0

Now support the following functions in the readme demo:

• 1. Read ROI
• 2. Crop by ROI
• 3. Read Reconstruction projects
  • read single reconstruction project by idp.Pix4D() and idp.Metashape()
  • Batch loading
• 4. Backward Projection

Enhancements

• Point cloud IO
• shapefile IO
• Geotiff IO
• Support ROI.back2raw
• Support ROI.crop

Testing publish on pypi

Now the easyidp v2.0 package support install from pypi

pip install easyidp

Final support for V1.0

We are going to reconstruct the full package to make it easier to use, the package structure will be totally rewrite. If you need previous version, please check the archive here