Skip to content

Commit 145196c

Browse files
pbashyal-nmdppbashyal-nmdp
committed
Run pre-commit
1 parent e8e389a commit 145196c

File tree

9 files changed

+142
-74
lines changed

9 files changed

+142
-74
lines changed

‎.github/PULL_REQUEST_TEMPLATE/pull_request_template.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -3,6 +3,6 @@ Description of Pull Request..
33
## Checklist
44

55
- [ ] Unit tests
6-
- [ ] Documentation
6+
- [ ] Documentation
77

88
Fixes #

‎.gitignore‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -139,4 +139,5 @@ allure_report/
139139
# cython temp files
140140
grim/**/*.c
141141

142+
.DS_Store
142143
output/

‎graph_generation/generate_neo4j_multi_hpf.py‎

Lines changed: 5 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -206,7 +206,10 @@ def loci_order(loc_values):
206206

207207

208208
def generate_graph(
209-
config_file="../conf/minimal-configuration-script.json", em_pop=None, em=False, use_default_path = False
209+
config_file="../conf/minimal-configuration-script.json",
210+
em_pop=None,
211+
em=False,
212+
use_default_path=False,
210213
):
211214
##############################################################################
212215
# Configure
@@ -218,7 +221,7 @@ def generate_graph(
218221
# freq_file = path + freq_file
219222
path = ""
220223
if use_default_path:
221-
path = os.path.dirname(os.path.realpath(__file__)) + "/"
224+
path = os.path.dirname(os.path.realpath(__file__)) + "/"
222225
parser = argparse.ArgumentParser()
223226
parser.add_argument(
224227
"-c",

‎grim/grim.py‎

Lines changed: 8 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -23,24 +23,20 @@
2323
#
2424

2525

26-
27-
28-
2926
from .imputation.impute import Imputation
3027
from .imputation.networkx_graph import Graph
3128

3229
import sys
3330
import os
3431

3532
# adding Folder_2 to the system path
36-
sys.path.insert(0, os.path.dirname(os.path.realpath(__file__)).replace("/grim", ""))
33+
sys.path.insert(0, os.path.dirname(os.path.realpath(__file__)).replace("/grim", ""))
3734

3835

3936
from graph_generation import generate_neo4j_multi_hpf
4037
from grim.run_impute_def import run_impute
4138

4239

43-
4440
def graph_freqs(conf_file="", for_em=False, em_pop=None):
4541
use_default_path = False
4642
if conf_file == "":
@@ -51,7 +47,10 @@ def graph_freqs(conf_file="", for_em=False, em_pop=None):
5147
)
5248

5349
generate_neo4j_multi_hpf.generate_graph(
54-
config_file=conf_file, em_pop=em_pop, em=for_em, use_default_path = use_default_path
50+
config_file=conf_file,
51+
em_pop=em_pop,
52+
em=for_em,
53+
use_default_path=use_default_path,
5554
)
5655

5756

@@ -68,7 +67,9 @@ def impute(conf_file=""):
6867
os.path.dirname(os.path.realpath(__file__)).replace("/grim", "")
6968
+ "/graph_generation/"
7069
)
71-
project_dir_in_file = os.path.dirname(os.path.realpath(__file__)).replace("/grim", "") + "/"
70+
project_dir_in_file = (
71+
os.path.dirname(os.path.realpath(__file__)).replace("/grim", "") + "/"
72+
)
7273
run_impute(conf_file, project_dir_graph, project_dir_in_file)
7374

7475

‎grim/run_impute_def.py‎

Lines changed: 107 additions & 56 deletions
Original file line numberDiff line numberDiff line change
@@ -5,75 +5,103 @@
55
import sys
66
import os
77

8-
sys.path.insert(0, os.path.dirname(os.path.realpath(__file__)))
8+
sys.path.insert(0, os.path.dirname(os.path.realpath(__file__)))
99

1010
from .imputation.impute import Imputation
1111
from .imputation.networkx_graph import Graph
1212

1313
# Profiler start
14-
#pr = cProfile.Profile()
15-
#pr.enable()
14+
# pr = cProfile.Profile()
15+
# pr.enable()
1616

17-
def run_impute(conf_file = "../conf/minimal-configuration.json", project_dir_graph = "", project_dir_in_file = ""):
1817

19-
configuration_file = conf_file
18+
def run_impute(
19+
conf_file="../conf/minimal-configuration.json",
20+
project_dir_graph="",
21+
project_dir_in_file="",
22+
):
2023

21-
#project_dir = ""# "../"
22-
#output_dir = "output/"
24+
configuration_file = conf_file
2325

26+
# project_dir = ""# "../"
27+
# output_dir = "output/"
2428

2529
# Read configuration file and load properties
2630
with open(configuration_file) as f:
2731
json_conf = json.load(f)
2832

2933
graph_files_path = json_conf.get("graph_files_path")
30-
if graph_files_path[-1] != '/':
31-
graph_files_path += '/'
34+
if graph_files_path[-1] != "/":
35+
graph_files_path += "/"
3236
output_dir = json_conf.get("imuptation_out_path", "output")
33-
if output_dir[-1] != '/':
34-
output_dir += '/'
37+
if output_dir[-1] != "/":
38+
output_dir += "/"
3539
config = {
36-
"planb": json_conf.get('planb', True),
37-
"pops": json_conf.get('populations'),
38-
"priority": json_conf.get('priority'),
39-
"epsilon": json_conf.get('epsilon', 1e-3),
40-
"number_of_results": json_conf.get('number_of_results', 1000),
41-
"number_of_pop_results": json_conf.get('number_of_pop_results', 100),
40+
"planb": json_conf.get("planb", True),
41+
"pops": json_conf.get("populations"),
42+
"priority": json_conf.get("priority"),
43+
"epsilon": json_conf.get("epsilon", 1e-3),
44+
"number_of_results": json_conf.get("number_of_results", 1000),
45+
"number_of_pop_results": json_conf.get("number_of_pop_results", 100),
4246
"output_MUUG": json_conf.get("output_MUUG", True),
4347
"output_haplotypes": json_conf.get("output_haplotypes", False),
44-
"node_file": project_dir_graph + graph_files_path + json_conf.get("node_csv_file"),
45-
"top_links_file": project_dir_graph + graph_files_path + json_conf.get("top_links_csv_file"),
46-
"edges_file": project_dir_graph + graph_files_path +json_conf.get("edges_csv_file"),
47-
"imputation_input_file": project_dir_in_file + json_conf.get("imputation_in_file"),
48-
"imputation_out_umug_freq_file": output_dir + json_conf.get("imputation_out_umug_freq_filename"),
49-
"imputation_out_umug_pops_file": output_dir + json_conf.get("imputation_out_umug_pops_filename"),
50-
"imputation_out_hap_freq_file": output_dir + json_conf.get("imputation_out_hap_freq_filename"),
51-
"imputation_out_hap_pops_file": output_dir + json_conf.get("imputation_out_hap_pops_filename"),
52-
"imputation_out_miss_file": output_dir + json_conf.get("imputation_out_miss_filename"),
53-
"imputation_out_problem_file": output_dir + json_conf.get("imputation_out_problem_filename"),
48+
"node_file": project_dir_graph
49+
+ graph_files_path
50+
+ json_conf.get("node_csv_file"),
51+
"top_links_file": project_dir_graph
52+
+ graph_files_path
53+
+ json_conf.get("top_links_csv_file"),
54+
"edges_file": project_dir_graph
55+
+ graph_files_path
56+
+ json_conf.get("edges_csv_file"),
57+
"imputation_input_file": project_dir_in_file
58+
+ json_conf.get("imputation_in_file"),
59+
"imputation_out_umug_freq_file": output_dir
60+
+ json_conf.get("imputation_out_umug_freq_filename"),
61+
"imputation_out_umug_pops_file": output_dir
62+
+ json_conf.get("imputation_out_umug_pops_filename"),
63+
"imputation_out_hap_freq_file": output_dir
64+
+ json_conf.get("imputation_out_hap_freq_filename"),
65+
"imputation_out_hap_pops_file": output_dir
66+
+ json_conf.get("imputation_out_hap_pops_filename"),
67+
"imputation_out_miss_file": output_dir
68+
+ json_conf.get("imputation_out_miss_filename"),
69+
"imputation_out_problem_file": output_dir
70+
+ json_conf.get("imputation_out_problem_filename"),
5471
"factor_missing_data": json_conf.get("factor_missing_data", 0.01),
55-
"loci_map": json_conf.get("loci_map", {"A": 1, "B":3, "C": 2, "DQB1": 4, "DRB1": 5} ),
56-
"matrix_planb": json_conf.get("Plan_B_Matrix", [
57-
[[1, 2, 3, 4, 5]],
58-
[[1, 2, 3], [4, 5]],
59-
[[1], [2, 3], [4, 5]],
60-
[[1, 2, 3], [4], [5]],
61-
[[1], [2, 3], [4], [5]],
62-
[[1], [2], [3], [4], [5]]
63-
]),
64-
"pops_count_file": project_dir_graph + json_conf.get("pops_count_file",'' ),
65-
"use_pops_count_file": json_conf.get("pops_count_file",False),
66-
"number_of_options_threshold": json_conf.get("number_of_options_threshold", 100000),
67-
"max_haplotypes_number_in_phase": json_conf.get("max_haplotypes_number_in_phase",100 ),
68-
"bin_imputation_input_file": project_dir_in_file + json_conf.get("bin_imputation_in_file", "None"),
72+
"loci_map": json_conf.get(
73+
"loci_map", {"A": 1, "B": 3, "C": 2, "DQB1": 4, "DRB1": 5}
74+
),
75+
"matrix_planb": json_conf.get(
76+
"Plan_B_Matrix",
77+
[
78+
[[1, 2, 3, 4, 5]],
79+
[[1, 2, 3], [4, 5]],
80+
[[1], [2, 3], [4, 5]],
81+
[[1, 2, 3], [4], [5]],
82+
[[1], [2, 3], [4], [5]],
83+
[[1], [2], [3], [4], [5]],
84+
],
85+
),
86+
"pops_count_file": project_dir_graph + json_conf.get("pops_count_file", ""),
87+
"use_pops_count_file": json_conf.get("pops_count_file", False),
88+
"number_of_options_threshold": json_conf.get(
89+
"number_of_options_threshold", 100000
90+
),
91+
"max_haplotypes_number_in_phase": json_conf.get(
92+
"max_haplotypes_number_in_phase", 100
93+
),
94+
"bin_imputation_input_file": project_dir_in_file
95+
+ json_conf.get("bin_imputation_in_file", "None"),
6996
"nodes_for_plan_A": json_conf.get("Plan_A_Matrix", []),
7097
"save_mode": json_conf.get("save_space_mode", False),
71-
"UNK_priors" : json_conf.get("UNK_priors", "MR")
72-
98+
"UNK_priors": json_conf.get("UNK_priors", "MR"),
7399
}
74100

75101
# Display the configurations we are using
76-
print('****************************************************************************************************')
102+
print(
103+
"****************************************************************************************************"
104+
)
77105
print("Performing imputation based on:")
78106
print("\tPopulation: {}".format(config["pops"]))
79107
print("\tPriority: {}".format(config["priority"]))
@@ -86,34 +114,57 @@ def run_impute(conf_file = "../conf/minimal-configuration.json", project_dir_gra
86114
print("\tTop Links File: {}".format(config["edges_file"]))
87115
print("\tInput File: {}".format(config["imputation_input_file"]))
88116
print("\tOutput UMUG Format: {}".format(config["output_MUUG"]))
89-
print("\tOutput UMUG Freq Filename: {}".format(config["imputation_out_umug_freq_file"]))
90-
print("\tOutput UMUG Pops Filename: {}".format(config["imputation_out_umug_pops_file"]))
117+
print(
118+
"\tOutput UMUG Freq Filename: {}".format(
119+
config["imputation_out_umug_freq_file"]
120+
)
121+
)
122+
print(
123+
"\tOutput UMUG Pops Filename: {}".format(
124+
config["imputation_out_umug_pops_file"]
125+
)
126+
)
91127
print("\tOutput Haplotype Format: {}".format(config["output_haplotypes"]))
92-
print("\tOutput HAP Freq Filename: {}".format(config["imputation_out_hap_freq_file"]))
93-
print("\tOutput HAP Pops Filename: {}".format(config["imputation_out_hap_pops_file"]))
128+
print(
129+
"\tOutput HAP Freq Filename: {}".format(config["imputation_out_hap_freq_file"])
130+
)
131+
print(
132+
"\tOutput HAP Pops Filename: {}".format(config["imputation_out_hap_pops_file"])
133+
)
94134
print("\tOutput Miss Filename: {}".format(config["imputation_out_miss_file"]))
95135
print("\tOutput Problem Filename: {}".format(config["imputation_out_problem_file"]))
96136
print("\tFactor Missing Data: {}".format(config["factor_missing_data"]))
97137
print("\tLoci Map: {}".format(config["loci_map"]))
98138
print("\tPlan B Matrix: {}".format(config["matrix_planb"]))
99139
print("\tPops Count File: {}".format(config["pops_count_file"]))
100140
print("\tUse Pops Count File: {}".format(config["use_pops_count_file"]))
101-
print("\tNumber of Options Threshold: {}".format(config["number_of_options_threshold"]))
102-
print("\tMax Number of haplotypes in phase: {}".format(config["max_haplotypes_number_in_phase"]))
141+
print(
142+
"\tNumber of Options Threshold: {}".format(
143+
config["number_of_options_threshold"]
144+
)
145+
)
146+
print(
147+
"\tMax Number of haplotypes in phase: {}".format(
148+
config["max_haplotypes_number_in_phase"]
149+
)
150+
)
103151
if config["nodes_for_plan_A"]:
104152
print("\tNodes in plan A: {}".format(config["nodes_for_plan_A"]))
105153
print("\tSave space mode: {}".format(config["save_mode"]))
106-
print('****************************************************************************************************')
107-
154+
print(
155+
"****************************************************************************************************"
156+
)
108157

109158
all_loci_set = set()
110159
for _, val in config["loci_map"].items():
111160
all_loci_set.add(str(val))
112161

113-
config["full_loci"] = ''.join(sorted(all_loci_set))
162+
config["full_loci"] = "".join(sorted(all_loci_set))
114163
# Perform imputation
115164
graph = Graph(config)
116-
graph.build_graph(config["node_file"], config["top_links_file"], config["edges_file"])
165+
graph.build_graph(
166+
config["node_file"], config["top_links_file"], config["edges_file"]
167+
)
117168
imputation = Imputation(graph, config)
118169

119170
# Create output directory if it doesn't exist
@@ -123,5 +174,5 @@ def run_impute(conf_file = "../conf/minimal-configuration.json", project_dir_gra
123174
imputation.impute_file(config)
124175

125176
# Profiler end
126-
#pr.disable()
127-
#pr.print_stats(sort="time")
177+
# pr.disable()
178+
# pr.print_stats(sort="time")

‎requirements.txt‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
cython==0.29.32
22
numpy>=1.20.2
3-
networkx==2.5.1
3+
networkx==2.5.1

‎scripts/parallel-imputation.py‎

Lines changed: 9 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -56,9 +56,15 @@
5656
"number_of_pop_results": json_conf.get("number_of_pop_results", 100),
5757
"output_MUUG": json_conf.get("output_MUUG", True),
5858
"output_haplotypes": json_conf.get("output_haplotypes", False),
59-
"node_file": project_dir + json_conf.get("graph_files_path") + json_conf.get("node_csv_file"),
60-
"top_links_file": project_dir + json_conf.get("graph_files_path") + json_conf.get("top_links_csv_file"),
61-
"edges_file": project_dir + json_conf.get("graph_files_path") + json_conf.get("edges_csv_file"),
59+
"node_file": project_dir
60+
+ json_conf.get("graph_files_path")
61+
+ json_conf.get("node_csv_file"),
62+
"top_links_file": project_dir
63+
+ json_conf.get("graph_files_path")
64+
+ json_conf.get("top_links_csv_file"),
65+
"edges_file": project_dir
66+
+ json_conf.get("graph_files_path")
67+
+ json_conf.get("edges_csv_file"),
6268
"imputation_input_file": project_dir + json_conf.get("imputation_in_file"),
6369
"imputation_out_umug_freq_file": output_dir
6470
+ json_conf.get("imputation_out_umug_freq_filename"),

‎scripts/runfile.py‎

Lines changed: 9 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -46,9 +46,15 @@
4646
"number_of_pop_results": json_conf.get("number_of_pop_results", 100),
4747
"output_MUUG": json_conf.get("output_MUUG", True),
4848
"output_haplotypes": json_conf.get("output_haplotypes", False),
49-
"node_file": project_dir + json_conf.get("graph_files_path") + json_conf.get("node_csv_file"),
50-
"top_links_file": project_dir + json_conf.get("graph_files_path") + json_conf.get("top_links_csv_file"),
51-
"edges_file": project_dir + json_conf.get("graph_files_path") + json_conf.get("edges_csv_file"),
49+
"node_file": project_dir
50+
+ json_conf.get("graph_files_path")
51+
+ json_conf.get("node_csv_file"),
52+
"top_links_file": project_dir
53+
+ json_conf.get("graph_files_path")
54+
+ json_conf.get("top_links_csv_file"),
55+
"edges_file": project_dir
56+
+ json_conf.get("graph_files_path")
57+
+ json_conf.get("edges_csv_file"),
5258
"imputation_input_file": project_dir + json_conf.get("imputation_in_file"),
5359
"imputation_out_umug_freq_file": output_dir
5460
+ json_conf.get("imputation_out_umug_freq_filename"),

‎setup.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -81,7 +81,7 @@
8181
"graph_generation.output",
8282
"data",
8383
"data.subjects",
84-
"conf"
84+
"conf",
8585
]
8686
),
8787
test_suite="tests",

0 commit comments

Comments
 (0)