Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Rf parallel #20

Merged
merged 20 commits into from
May 19, 2021
Merged

Rf parallel #20

Show file tree
Hide file tree
Changes from 16 commits
Commits
Show all changes
20 commits
Select commit Hold shift + click to select a range
216147d
fix the color heatmap at random forest
yael1994 Nov 26, 2020
b65ff16
fix the commit from the pull request
yael1994 Nov 29, 2020
ec5944d
add a flag for stop before run random forest
yael1994 Dec 16, 2020
68aaa18
strat to build run parallel random forest
yael1994 Jan 12, 2021
0200272
random forest parallel with new file train_random forest
yael1994 Jan 13, 2021
2fe6ba7
spaces
yael1994 Feb 23, 2021
c89503d
save results in new file at each model
yael1994 Feb 23, 2021
73ef454
Arranging the code
yael1994 Feb 24, 2021
8a4c6e8
fix prints to summary file
yael1994 Mar 10, 2021
94dd7ec
add flages to module wraper
yael1994 Mar 11, 2021
f552698
change the position of break
yael1994 Mar 29, 2021
375ac90
change the imports
yael1994 Apr 27, 2021
766a435
add seed for train random forest
yael1994 Apr 28, 2021
a72f396
Merge branch 'reads_filtration_change_seq_length' into RF_parallel
yael1994 Apr 28, 2021
409d658
move new params out to all pipeline
yael1994 Apr 28, 2021
952bd61
move new params out to all pipeline
yael1994 Apr 28, 2021
c31ce03
resolve the comments
yael1994 May 4, 2021
f803d87
Merge branch 'reads_filtration_change_seq_length' into RF_parallel
yael1994 May 4, 2021
63112a8
Fixed typeos
shacharmo May 19, 2021
9525e37
Merge branch 'reads_filtration_change_seq_length' into RF_parallel
yael1994 May 19, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 17 additions & 7 deletions IgOmeProfiling_pipeline.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,13 +10,15 @@
sys.path.insert(0, src_dir)

from auxiliaries.pipeline_auxiliaries import *


def run_pipeline(fastq_path, barcode2samplename_path, samplename2biologicalcondition_path, analysis_dir, logs_dir,
left_construct, right_construct, max_mismatches_allowed, min_sequencing_quality, minimal_length_required, gz,
max_msas_per_sample, max_msas_per_bc,
max_number_of_cluster_members_per_sample, max_number_of_cluster_members_per_bc,
allowed_gap_frequency, concurrent_cutoffs, meme_split_size, use_mapitope, number_of_random_pssms,
rank_method, tfidf_method, tfidf_factor, shuffles, shuffles_percent, shuffles_digits,
allowed_gap_frequency, concurrent_cutoffs, meme_split_size, use_mapitope, stop_random_forest, number_of_random_pssms,
number_parallel_random_forest, min_value_error_random_forest,
rank_method, tfidf_method, tfidf_factor, shuffles, shuffles_percent, shuffles_digits, seed_random_forest_classifier,
run_summary_path, error_path, queue, verbose, argv):

os.makedirs(os.path.split(run_summary_path)[0], exist_ok=True)
Expand Down Expand Up @@ -86,8 +88,11 @@ def run_pipeline(fastq_path, barcode2samplename_path, samplename2biologicalcondi

module_parameters = [first_phase_output_path, second_phase_output_path, third_phase_output_path,
third_phase_logs_path, samplename2biologicalcondition_path, number_of_random_pssms,
third_phase_done_path, f'--shuffles_percent {shuffles_percent}', f'--shuffles_digits {shuffles_digits}' ,f'--rank_method {rank_method}', f'--error_path {error_path}',
'-v' if verbose else '', f'-q {queue}','-m' if use_mapitope else '']
third_phase_done_path, '--stop_random_forest' if stop_random_forest else '',
f'--number_parallel_random_forest' {number_parallel_random_forest}, f'--min_value_error_random_forest' {min_value_error_random_forest},
f'--shuffles_percent {shuffles_percent}', f'--shuffles_digits {shuffles_digits}'
f'--rank_method {rank_method}', f'--seed_random_forest_classifier {seed_random_forest_classifier}', f'--error_path {error_path}', '-v' if verbose else '',
f'-q {queue}','-m' if use_mapitope else '']
if rank_method == 'tfidf':
if tfidf_method:
module_parameters += ['--tfidf_method', tfidf_method]
Expand Down Expand Up @@ -161,14 +166,18 @@ def run_pipeline(fastq_path, barcode2samplename_path, samplename2biologicalcondi
parser.add_argument('-m', '--mapitope', action='store_true', help='use mapitope encoding')

# optional parameters for the modelling step
parser.add_argument('--stop_random_forest', action='store_true', help='A boolean flag for mark if we need to run the random forest')
yael1994 marked this conversation as resolved.
Show resolved Hide resolved
parser.add_argument('--number_of_random_pssms', default=100, type=int, help='Number of pssm permutations')
parser.add_argument('--number_parallel_random_forest', default=20, type=int, help='How many random forest to run in parallel')
yael1994 marked this conversation as resolved.
Show resolved Hide resolved
parser.add_argument('--min_value_error_random_forest', default=0, type=float, help='A min value for error that the run can stop in the random forest')
yael1994 marked this conversation as resolved.
Show resolved Hide resolved
parser.add_argument('--rank_method', choices=['pval', 'tfidf', 'shuffles'], default='pval', help='Motifs ranking method')
parser.add_argument('--tfidf_method', choices=['boolean', 'terms', 'log', 'augmented'], default='boolean', help='TF-IDF method')
parser.add_argument('--tfidf_factor', type=float, default=0.5, help='TF-IDF augmented method factor (0-1)')
parser.add_argument('--shuffles', default=5, type=int, help='Number of controlled shuffles permutations')
parser.add_argument('--shuffles_percent', default=0.2, type=float, help='Percent from shuffle with greatest number of hits (0-1)')
parser.add_argument('--shuffles_digits', default=2, type=int, help='Number of digits after the point to print in scanning files.')

parser.add_argument('--seed_random_forest_classifier', defual=123 , type=int, help='A number for create the random forest stable when run the same configuration')
yael1994 marked this conversation as resolved.
Show resolved Hide resolved

# general optional parameters
parser.add_argument('--run_summary_path', type=str,
help='a file in which the running configuration and timing will be written to')
Expand All @@ -195,6 +204,7 @@ def run_pipeline(fastq_path, barcode2samplename_path, samplename2biologicalcondi
args.left_construct, args.right_construct, args.max_mismatches_allowed, args.min_sequencing_quality, args.minimal_length_required, True if args.gz else False,
args.max_msas_per_sample, args.max_msas_per_bc,
args.max_number_of_cluster_members_per_sample, args.max_number_of_cluster_members_per_bc,
args.allowed_gap_frequency, concurrent_cutoffs, args.meme_split_size, args.mapitope, args.number_of_random_pssms,
args.allowed_gap_frequency, concurrent_cutoffs, args.meme_split_size, args.mapitope,True if args.stop_random_forest else False,
yael1994 marked this conversation as resolved.
Show resolved Hide resolved
args.number_of_random_pssms, args.number_parallel_random_forest, args.min_value_error_random_forest,
args.rank_method, args.tfidf_method, args.tfidf_factor, args.shuffles, args.shuffles_percent, args.shuffles_digits,
run_summary_path, error_path, args.queue, True if args.verbose else False, sys.argv)
args.seed_random_forest_classifier, run_summary_path, error_path, args.queue, True if args.verbose else False, sys.argv)
yael1994 marked this conversation as resolved.
Show resolved Hide resolved
109 changes: 61 additions & 48 deletions model_fitting/module_wraper.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,9 +21,11 @@ def repeat_items(list):


def build_classifier(first_phase_output_path, motif_inference_output_path,
classification_output_path, logs_dir, samplename2biologicalcondition_path,
fitting_done_path, number_of_random_pssms, rank_method, tfidf_method, tfidf_factor,
shuffles, shuffles_percent, shuffles_digits, queue_name, verbose, error_path, use_mapitope, argv):
classification_output_path, logs_dir, samplename2biologicalcondition_path, number_of_random_pssms,
fitting_done_path, stop_random_forest, num_of_configurations_to_sample,
number_parallel_random_forest, min_value_error_random_forest, rank_method, tfidf_method, tfidf_factor,
shuffles, shuffles_percent, shuffles_digits, seed_random_forest_classifier,queue_name, verbose, seed_random_forest_classifier, error_path, use_mapitop, argv):

is_pval = rank_method == 'pval'
os.makedirs(classification_output_path, exist_ok=True)
os.makedirs(logs_dir, exist_ok=True)
Expand All @@ -42,7 +44,6 @@ def build_classifier(first_phase_output_path, motif_inference_output_path,
scanning_dir_path = os.path.join(bc_dir_path, 'scanning')
os.makedirs(scanning_dir_path, exist_ok=True)


# compute scanning scores (hits and values)
logger.info('_'*100)
logger.info(f'{datetime.datetime.now()}: scanning peptides vs motifs (hits and values)')
Expand Down Expand Up @@ -145,49 +146,54 @@ def build_classifier(first_phase_output_path, motif_inference_output_path,


# fitting a random forest model (hits and values)
logger.info('_'*100)
logger.info(f'{datetime.datetime.now()}: fitting model')
script_name = 'random_forest.py'
num_of_expected_results = 0
all_cmds_params = [] # a list of lists. Each sublist contain different parameters set for the same script to reduce the total number of jobs
for bc in biological_conditions:
aggregated_values_path = os.path.join(classification_output_path, bc, f'{bc}_values.csv')
pvalues_done_path = os.path.join(logs_dir, f'{bc}_values_done_fitting.txt')
aggregated_hits_path = os.path.join(classification_output_path, bc, f'{bc}_hits.csv')
hits_done_path = os.path.join(logs_dir, f'{bc}_hits_done_fitting.txt')

value_cmd = [aggregated_values_path, pvalues_done_path]
hits_cmd = [aggregated_hits_path, hits_done_path]
if rank_method == 'tfidf' or rank_method == 'shuffles':
value_cmd.append('--tfidf')
hits_cmd.append('--tfidf')

if not os.path.exists(pvalues_done_path):
all_cmds_params.append(value_cmd)
else:
logger.debug(f'Skipping fitting as {pvalues_done_path} found')
num_of_expected_results += 1

if not os.path.exists(hits_done_path):
all_cmds_params.append(hits_cmd)
if not stop_random_forest:
print('not stop')
logger.info('_'*100)
logger.info(f'{datetime.datetime.now()}: fitting model')
script_name = 'random_forest.py'
num_of_expected_results = 0
all_cmds_params = [] # a list of lists. Each sublist contain different parameters set for the same script to reduce the total number of jobs
for bc in biological_conditions:
aggregated_values_path = os.path.join(classification_output_path, bc, f'{bc}_values.csv')
pvalues_done_path = os.path.join(logs_dir, f'{bc}_values_done_fitting.txt')
aggregated_hits_path = os.path.join(classification_output_path, bc, f'{bc}_hits.csv')
hits_done_path = os.path.join(logs_dir, f'{bc}_hits_done_fitting.txt')

value_cmd = [aggregated_values_path, pvalues_done_path, logs_dir, error_path, '--num_of_configurations_to_sample', num_of_configurations_to_sample,
'--number_parallel_random_forest', number_parallel_random_forest, '--min_value_error_random_forest', min_value_error_random_forest, '--seed_random_forest_classifier', seed_random_forest_classifier]
hits_cmd = [aggregated_hits_path, hits_done_path, logs_dir, error_path, '--num_of_configurations_to_sample', num_of_configurations_to_sample,
'--number_parallel_random_forest', number_parallel_random_forest, '--min_value_error_random_forest', min_value_error_random_forest, '--seed_random_forest_classifier', seed_random_forest_classifier]
if rank_method == 'tfidf' or rank_method == 'shuffles':
value_cmd.append('--tfidf')
hits_cmd.append('--tfidf')
if not os.path.exists(pvalues_done_path):
all_cmds_params.append(value_cmd)
else:
logger.debug(f'Skipping fitting as {pvalues_done_path} found')
num_of_expected_results += 1

if not os.path.exists(hits_done_path):
all_cmds_params.append(hits_cmd)
else:
logger.debug(f'Skipping fitting as {hits_done_path} found')
num_of_expected_results += 1

if len(all_cmds_params) > 0:
doubled_bc = repeat_items(biological_conditions)
for cmds_params, bc in zip(all_cmds_params, doubled_bc):
cmd = submit_pipeline_step(f'{src_dir}/model_fitting/{script_name}',
[cmds_params],
logs_dir, f'{bc}_model',
queue_name, verbose)
num_of_expected_results += 1 # a single job for each biological condition

wait_for_results(script_name, logs_dir, num_of_expected_results, example_cmd=cmd,
error_file_path=error_path, suffix='_done_fitting.txt')
else:
logger.debug(f'Skipping fitting as {hits_done_path} found')
num_of_expected_results += 1

if len(all_cmds_params) > 0:
doubled_bc = repeat_items(biological_conditions)
for cmds_params, bc in zip(all_cmds_params, doubled_bc):
cmd = submit_pipeline_step(f'{src_dir}/model_fitting/{script_name}',
[cmds_params],
logs_dir, f'{bc}_model',
queue_name, verbose)
num_of_expected_results += 1 # a single job for each biological condition

wait_for_results(script_name, logs_dir, num_of_expected_results, example_cmd=cmd,
error_file_path=error_path, suffix='_done_fitting.txt')
logger.info(f'Skipping fitting, all found')
else:
logger.info(f'Skipping fitting, all found')

print('stop before')
logger.info(f'Stop before random forest')

# TODO: fix this bug with a GENERAL WRAPPER done_path
# wait_for_results(script_name, num_of_expected_results)
Expand Down Expand Up @@ -217,12 +223,17 @@ def get_faa_file_name_from_path(path, use_mapitope):
parser.add_argument('number_of_random_pssms', default=100, type=int, help='Number of pssm permutations')
parser.add_argument('done_file_path', help='A path to a file that signals that the module finished running successfully.')

parser.add_argument('--stop_random_forest', action='store_true',help='A boolean flag for mark if we need to run the random forest')
parser.add_argument('--num_of_configurations_to_sample', default=100, type=int, help='How many random configurations of hyperparameters should be sampled?')
parser.add_argument('--number_parallel_random_forest', default=20, type=int, help='How many random forest to run in parallel')
parser.add_argument('--min_value_error_random_forest', default=0, type=float, help='A min value for error that the run can stop in the random forest')
parser.add_argument('--rank_method', choices=['pval', 'tfidf', 'shuffles'], default='pval', help='Motifs ranking method')
parser.add_argument('--tfidf_method', choices=['boolean', 'terms', 'log', 'augmented'], default='boolean', help='TF-IDF method')
parser.add_argument('--tfidf_factor', type=float, default=0.5, help='TF-IDF augmented method factor (0-1)')
parser.add_argument('--shuffles', default=5, type=int, help='Number of controlled shuffles permutations')
parser.add_argument('--shuffles_percent', default=0.2, type=float, help='Percent from shuffle with greatest number of hits (0-1)')
parser.add_argument('--shuffles_digits', default=2, type=int, help='Number of digits after the point to print in scanning files.')
parser.add_argument('--seed_random_forest_classifier', defual=123 , type=int, help='A number for create the random forest stable when run the same configuration')
parser.add_argument('--error_path', type=str, help='a file in which errors will be written to')
parser.add_argument('-q', '--queue', default='pupkoweb', type=str, help='a queue to which the jobs will be submitted')
parser.add_argument('-v', '--verbose', action='store_true', help='Increase output verbosity')
Expand All @@ -239,6 +250,8 @@ def get_faa_file_name_from_path(path, use_mapitope):
error_path = args.error_path if args.error_path else os.path.join(args.parsed_fastq_results, 'error.txt')

build_classifier(args.parsed_fastq_results, args.motif_inference_results, args.classification_output_path,
args.logs_dir, args.samplename2biologicalcondition_path, args.done_file_path,
args.number_of_random_pssms, args.rank_method, args.tfidf_method, args.tfidf_factor,
args.shuffles, args.shuffles_percent, args.shuffles_digits, args.queue, True if args.verbose else False, error_path, args.mapitope, sys.argv)
args.logs_dir, args.samplename2biologicalcondition_path, args.number_of_random_pssms, args.done_file_path,
True if args.stop_random_forest else False, args.num_of_configurations_to_sample,
args.number_parallel_random_forest, args.min_value_error_random_forest, args.rank_method,
args.tfidf_method, args.tfidf_factor, args.shuffles, args.shuffles_percent, args.shuffles_digits, args.seed_random_forest_classifier,
args.queue, True if args.verbose else False, error_path, args.mapitope, sys.argv)
Loading