From 07e9d6270cbe168c78340419156ba2bcedc45ad5 Mon Sep 17 00:00:00 2001
From: Constantin Pape <constantin.pape@iwr.uni-heidelberg.de>
Date: Fri, 27 Sep 2019 10:34:45 +0200
Subject: [PATCH] Update registration mechanism and fix edu names
---
scripts/attributes/genes.py | 13 +++---------
update_registration.py | 40 +++++++++++++++++++++++--------------
2 files changed, 28 insertions(+), 25 deletions(-)
diff --git a/scripts/attributes/genes.py b/scripts/attributes/genes.py
index 988445c..2f007ce 100755
--- a/scripts/attributes/genes.py
+++ b/scripts/attributes/genes.py
@@ -47,26 +47,19 @@ def create_auxiliary_gene_file(meds_path, out_file, return_result=False):
names_dset = 'gene_names'
dset = 't00000/s00/0/cells'
- # NOTE special treatment of edu will not be necessary if we add MED postfix to them again
- # get all the meds in the folder
- # 1.) meds
+ # get all the med files in the image folder
med_files = glob(os.path.join(meds_path, "*MED.h5"))
gene_names = [os.path.splitext(os.path.basename(f))[0][:-4] for f in med_files]
- # 2.) edu
- edu_files = glob(os.path.join(meds_path, "*edu*"))
- gene_names += [os.path.splitext(os.path.basename(f))[0] for f in edu_files]
-
- all_files = med_files + edu_files
# find out where the gene name actually starts (the 5th dash-separated word)
name_start = find_nth(gene_names[0], '-', 4) + 1
# cut all the preceeding prospr-... part
gene_names = [name[name_start:] for name in gene_names]
- num_genes = len(all_files)
+ num_genes = len(med_files)
assert len(gene_names) == num_genes, "%i, %i" % (len(gene_names), len(num_genes))
- with h5py.File(all_files[0], 'r') as f:
+ with h5py.File(med_files[0], 'r') as f:
spatial_shape = f[dset].shape
shape = (num_genes,) + spatial_shape
diff --git a/update_registration.py b/update_registration.py
index 3cd0870..56ca9c8 100755
--- a/update_registration.py
+++ b/update_registration.py
@@ -12,7 +12,7 @@ import numpy as np
import pandas as pd
from pybdv import make_bdv
-from scripts.files import copy_release_folder, make_folder_structure, make_bdv_server_file
+from scripts.files import copy_release_folder, make_folder_structure, make_bdv_server_file, get_source_names
from scripts.files.xml_utils import get_h5_path_from_xml, write_simple_xml
from scripts.release_helper import add_version
from scripts.extension.registration import ApplyRegistrationLocal, ApplyRegistrationSlurm
@@ -41,7 +41,7 @@ def get_tags(new_tag):
return tag, new_tag
-def get_out_name(prefix, name):
+def parse_prospr(prefix, name):
name = os.path.split(name)[1]
name = os.path.splitext(name)[0]
name = name.split('--')[0]
@@ -50,9 +50,6 @@ def get_out_name(prefix, name):
# name = name.split('-')[0]
if name in REGION_NAMES:
name = '-'.join([prefix, 'segmented', name])
- # TODO edus shouls still have MED postfix !
- elif name.startswith('edu'): # edus are no meds
- name = '-'.join([prefix, name])
else:
name = '-'.join([prefix, name, 'MED'])
return name
@@ -79,7 +76,7 @@ def copy_to_h5(inputs, output_folder):
def apply_registration(input_folder, new_folder,
transformation_file, source_prefix,
- target, max_jobs):
+ target, max_jobs, name_parser):
task = ApplyRegistrationSlurm if target == 'slurm' else ApplyRegistrationLocal
tmp_folder = './tmp_registration'
os.makedirs(tmp_folder, exist_ok=True)
@@ -98,7 +95,7 @@ def apply_registration(input_folder, new_folder,
# output_folder = os.path.join(new_folder, 'images')
output_folder = os.path.join(tmp_folder, 'outputs')
os.makedirs(output_folder, exist_ok=True)
- output_names = [get_out_name(source_prefix, name) for name in inputs]
+ output_names = [name_parser(source_prefix, name) for name in inputs]
outputs = [os.path.join(output_folder, name) for name in output_names]
# update the task config
@@ -142,7 +139,7 @@ def apply_registration(input_folder, new_folder,
copy_to_h5(outputs, output_folder)
-def update_gene_attributes(new_folder, target):
+def update_prospr(new_folder, target):
# update the auxiliaty gene volume
image_folder = os.path.join(new_folder, 'images')
@@ -173,8 +170,12 @@ def update_gene_attributes(new_folder, target):
write_genes_table(seg_path, aux_out_path, out_path,
labels, tmp_folder, target)
+ # TODO we should register the virtual cells here as well
-def update_regestration(transformation_file, input_folder, source_prefix, target, max_jobs,
+
+# we should encode the source prefix and the transformation file to be used
+# in a config file in the transformation folder
+def update_registration(transformation_file, input_folder, source_prefix, target, max_jobs,
new_tag=None):
""" Update the prospr segmentation.
This is a special case of 'update_patch', that applies a new prospr registration.
@@ -187,6 +188,10 @@ def update_regestration(transformation_file, input_folder, source_prefix, target
max_jobs [int] - max number of jobs for computation
new_tag [str] - new version tag (default: None)
"""
+ prefixes = get_source_names()
+ if source_prefix not in prefixes:
+ raise ValueError("Invalid source name %s" % source_prefix)
+
tag, new_tag = get_tags(new_tag)
print("Updating platy browser from", tag, "to", new_tag)
@@ -198,16 +203,21 @@ def update_regestration(transformation_file, input_folder, source_prefix, target
# copy the release folder
copy_release_folder(folder, new_folder, exclude_prefixes=[source_prefix])
+ if source_prefix == "prospr-6dpf-1-whole":
+ name_parser = parse_prospr
+ else:
+ raise NotImplementedError
+
# apply new registration to all files of the source prefix
transformation_file = os.path.abspath(transformation_file)
apply_registration(input_folder, new_folder,
transformation_file, source_prefix,
- target, max_jobs)
-
- # TODO we should register virtual cells here too
+ target, max_jobs, name_parser)
- # update gene table
- update_gene_attributes(new_folder, target)
+ if source_prefix == "prospr-6dpf-1-whole":
+ update_prospr(new_folder, target)
+ else:
+ raise NotImplementedError
add_version(new_tag)
make_bdv_server_file(new_folder, os.path.join(new_folder, 'misc', 'bdv_server.txt'),
@@ -236,5 +246,5 @@ if __name__ == '__main__':
new_tag = args.new_tag
new_tag = None if new_tag == '' else new_tag
- update_regestration(args.transformation_file, args.input_folder, args.source_prefix,
+ update_registration(args.transformation_file, args.input_folder, args.source_prefix,
args.target, args.max_jobs, new_tag)
--
GitLab