When serializing datasets with sample names the sample names are stored directly in project.json. All of these sample names are stored multiple times for all the data that was in the HDF5 file. For datasets with e.g. 2.5 million points and fairly long sample names, this results in a project.json of multiple gigabytes, which results in the JSON loader being unable to load it properly.
When serializing datasets with sample names the sample names are stored directly in project.json. All of these sample names are stored multiple times for all the data that was in the HDF5 file. For datasets with e.g. 2.5 million points and fairly long sample names, this results in a project.json of multiple gigabytes, which results in the JSON loader being unable to load it properly.