diff --git a/src/find_interaction_cluster/clip_figures/clip_analyser.py b/src/find_interaction_cluster/clip_figures/clip_analyser.py
index b30447e4a4151f26a398e7386e2db55aa578e06c..4346a2f2de4ce5267c92f51329db7a729bcf2e4f 100644
--- a/src/find_interaction_cluster/clip_figures/clip_analyser.py
+++ b/src/find_interaction_cluster/clip_figures/clip_analyser.py
@@ -17,6 +17,8 @@ from ..community_finder import multiple_community_launcher, get_projects
from ..community_figures.fig_functions import create_community_fig
from ...logging_conf import logging_def
import logging
+from typing import Tuple
+import multiprocessing as mp
def bedtools_intersect(gene_bed: Path, clip_bed: Path,
@@ -197,10 +199,46 @@ def create_table(feature: str, clip_file: Path,
return merge_dataframes(df_clip, df_com, feature)
+def select_community_file(project: str, weight: int, global_weight: int,
+ same_gene: bool, feature: str,
+ community_file: str = "") -> Tuple[Path, Path]:
+ """
+ Return the community file and output folder that will be used.
+
+ :param project: The name of the project of interest
+ :param weight: The minimum weight of interaction to consider
+ :param global_weight: The global weight to consider. if \
+ the global weight is equal to 0 then then density figure are calculated \
+ by project, else all projet are merge together and the interaction \
+ seen in `global_weight` project are taken into account
+ :param same_gene: Say if we consider as co-localised, exons within the \
+ same gene (True) or not (False) (default False)
+ :param feature: The feature we want to analyse (default 'exon')
+ :param community_file: A file containing custom communities. If \
+ it equals to '' then weight, global weight and same genes parameter are \
+ used to find the community files computed with ChIA-PET data.
+ :return: The community file used and the output folder used.
+
+ """
+ if community_file == "":
+ com_file = find_or_create_community(project, weight, global_weight,
+ same_gene, feature)
+ output = com_file.parent / f"CLIP_community_figures_{feature}"
+ else:
+ com_file = Path(community_file)
+ if not com_file.is_file():
+ raise FileNotFoundError(f"File {com_file} was not found !")
+ tmp_name = com_file.name.replace(".txt", "")
+ output = ConfigClip.output_folder / \
+ f"CLIP_community_figures-{feature}-{tmp_name}"
+ return com_file, output
+
+
def create_figure(project: str, weight: int, global_weight: int,
same_gene: bool, feature: str, clip_file: Path,
feature_bed: Path, test_type: str = "permutation",
- iteration: int = 10000) -> None:
+ iteration: int = 10000, display_size: bool=False,
+ community_file: str = "") -> None:
"""
Create the final figure
:param project: The name of the project of interest
@@ -218,21 +256,30 @@ def create_figure(project: str, weight: int, global_weight: int,
:param test_type: The king of test to perform for frequency analysis. \
(default 'lm') (choose from 'lm', 'permutation')
:param iteration: The number of iteration to make
+ :param community_file: A file containing custom communities. If \
+ it equals to '' then weight, global weight and same genes parameter are \
+ used to find the community files computed with ChIA-PET data.
+ :param display_size: True to display the size of the community. \
+ False to display nothing. (default False)
+ :param ps: The number of processes to create (default 1)
"""
- com_file = find_or_create_community(project, weight, global_weight,
- same_gene, feature)
- output = com_file.parent / f"CLIP_community_figures_{feature}"
- output.mkdir(exist_ok=True)
+ logging.info(f"Working on {clip_file}")
+ com_file, output = select_community_file(project, weight, global_weight,
+ same_gene, feature,
+ community_file)
+ output.mkdir(exist_ok=True, parents=True)
outfile = output / f"{clip_file.name.split('.')[0]}.pdf"
final_table = create_table(feature, clip_file, feature_bed, com_file)
create_community_fig(final_table, feature, "peak_density", outfile,
- test_type, iteration=iteration)
+ test_type, iteration=iteration,
+ display_size=display_size)
def clip_folder_analysis(clip_folder: Path, project: str, weight: int,
global_weight: int, same_gene: bool, feature: str,
test_type: str = "permutation",
- iteration: int = 10000,
+ iteration: int = 10000, display_size: bool=False,
+ community_file: str = "", ps: int = 1,
logging_level: str = "DEBUG") -> None:
"""
Create the final figure
@@ -249,6 +296,12 @@ def clip_folder_analysis(clip_folder: Path, project: str, weight: int,
:param test_type: The king of test to perform for frequency analysis. \
(default 'lm') (choose from 'lm', 'permutation')
:param iteration: The number of iteration to make
+ :param display_size: True to display the size of the community. \
+ False to display nothing. (default False)
+ :param community_file: A file containing custom communities. If \
+ it equals to '' then weight, global weight and same genes parameter are \
+ used to find the community files computed with ChIA-PET data.
+ :param ps: The number of processes to create (default 1)
:param logging_level: The level of data to display (default 'DISABLE')
"""
logging_def(ConfigGraph.community_folder, __file__, logging_level)
@@ -256,7 +309,11 @@ def clip_folder_analysis(clip_folder: Path, project: str, weight: int,
else ConfigClip.bed_exon
files = list(clip_folder.glob("*.bed")) + \
list(clip_folder.glob("*.bed.gz"))
+ pool = mp.Pool(processes=min(len(files), ps))
+ processes = []
for mfile in files:
- logging.info(f"Working on {mfile}")
- create_figure(project, weight, global_weight, same_gene, feature,
- mfile, feature_bed, test_type, iteration)
+ args = [project, weight, global_weight, same_gene, feature,
+ mfile, feature_bed, test_type, iteration, display_size,
+ community_file]
+ processes.append(pool.apply_async(create_figure, args))
+ [p.get(timeout=None) for p in processes]