Block-wise connected components not working properly

[constantinpape]

Not everything is merged across block boundaries properly.
In order to reproduce with data from platy browser:

import numpy as np
import pandas as pd
import elf.parallel
from elf.io import open_file

# leaving this here to reproduce issues with elf.parallel.label
def make_nephridia_segmentation():
    table_path = '../../data/0.6.5/tables/sbem-6dpf-1-whole-segmented-cilia/cell_mapping.csv'
    seg_path = '../../data/0.6.5/segmentations/sbem-6dpf-1-whole-segmented-cells.h5'
    out_path = '/g/kreshuk/pape/Work/nephr_tmp.n5'
    out_key = 'test'

    table = pd.read_csv(table_path, sep='\t')
    cell_ids = table['cell_id'].values
    cell_ids = np.unique(cell_ids)
    if cell_ids[0] == 0:
        cell_ids = cell_ids[1:]
    print(cell_ids)

    scale = 4 
    key = 't00000/s00/%i/cells' % scale
    with open_file(seg_path, 'r') as f, open_file(out_path) as f_out:
        ds = f[key]
        print(ds.shape)

        out = f_out.require_dataset(out_key, shape=ds.shape, chunks=ds.chunks, compression='gzip',
                                    dtype='uint8')

        print("Isin ...")
        out = elf.parallel.isin(ds, cell_ids, out=out, n_threads=16, verbose=True)
        print("Label ...")
        # FIXME label is still not doing a completely accurate job of merging across block boundaries         
        out = elf.parallel.label(out, out, n_threads=16, verbose=True)
        print("Compute max ...")
        max_id = elf.parallel.max(out, n_threads=16, verbose=True)
        print("Max component id:", max_id)