run_pipeline.py

import os
import tempfile
from collections import defaultdict
import string

import htsomeropy
import pandas as pd
from tqdm import tqdm

import cellprofiler_core.preferences as cpprefs
cpprefs.set_headless()
import cellprofiler
import cellprofiler.modules as cpm
import cellprofiler_core.pipeline as cpp
from cellprofiler_core.modules.injectimage import InjectImage


HERE = os.path.dirname(os.path.abspath(__file__))
PIPELINE_PATH = os.path.join(HERE, "pipeline.cppipe")
N_WELLS = 20

# tempfile
new_output_directory = os.path.normcase(tempfile.mkdtemp())
cpprefs.set_default_output_directory(new_output_directory)


def load_pipeline(pipeline_path=PIPELINE_PATH):
    pipeline = cpp.Pipeline()
    pipeline.load(pipeline_path)
    return pipeline


def prep_pipeline(pipeline):
    """remove first 4 modules which aren't needed"""
    for i in range(4):
        pipeline.remove_module(1)
    return pipeline


try:
    omero = htsomeropy.gateway.Omero()
    omero.interactive_login()
    print(" ** connected to omero ** ")

    plate = omero.plate(plate_id=52)

    orig_pipeline = load_pipeline(PIPELINE_PATH)
    pipeline = prep_pipeline(orig_pipeline)

    wells = plate.wells
    wells = wells[:N_WELLS]
    well_count = len(wells)

    files = defaultdict(list)

    for well in tqdm(wells):
        for field_n, field in enumerate(tqdm(well.fields, leave=False), 1):
            pipeline_copy = pipeline.copy()
            for img, channel_name in field.channels:
                inject_image_module = InjectImage(channel_name, img)
                inject_image_module.set_module_num(1)
                pipeline_copy.add_module(inject_image_module)

            #print(f"analysing plate:{plate.name}, well:{well.name} field:{field_n}")
            m = pipeline_copy.run()

            # collect results
            for obj in ["nuclei", "cells"]:
                path = os.path.join(new_output_directory, f"MyExpt_{obj}.csv")
                f = pd.read_csv(path, index_col=None, header=0)
                f["Metadata_well"] = well.name
                f["Metadata_site"] = field_n
                f["CellCount"] = f.shape[0]
                files[obj].append(f)

    for obj in ["nuclei", "cells"]:
        df = pd.concat(files[obj], ignore_index=True)
        print(df)
        df.to_csv(f"output_{obj}.csv", index=False)



finally:
    omero.disconnect()