19136907997

Committed 06 Nov 2025 12:56PM UTC coverage: 68.835% (-0.4%) from 69.186%

Build # 19136907997

Build Type

push

github

Committed by

web-flow

Commit Message

Merge pull request #875 from cokelaer/dev

update HTML variant calling + others

Run Details

157 of 332 new or added lines in 14 files covered. (47.29%)

5 existing lines in 3 files now uncovered.

14551 of 21139 relevant lines covered (68.83%)

2.07 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

51.22

/sequana/scripts/main/html_report.py

#  This file is part of Sequana software
#
#  Copyright (c) 2016-2020 - Sequana Development Team
#
#  Distributed under the terms of the 3-clause BSD license.
#  The full license is in the LICENSE file, distributed with this software.
#
#  website: https://github.com/sequana/sequana
#  documentation: http://sequana.readthedocs.io
#
##############################################################################
import sys

import colorlog
import rich_click as click

from sequana.scripts.utils import CONTEXT_SETTINGS

logger = colorlog.getLogger(__name__)


groups = [
    {
        "name": "VCF options",
        "options": [
            "--freebayes-score",
            "--strand-ratio",
            "--frequency",
            "--forward-depth",
            "--keep-polymorphic",
            "--reverse-depth",
            "--min-depth",
            "--output-vcf-file",
            "--output-csv-file",
        ],
    },
    {
        "name": "General",
        "options": ["--output-directory"],
    },
]

click.rich_click.OPTION_GROUPS["html-report"] = groups.copy()


# not the hyphen html-report (not html_report). this is a 'bug/feature'
# @click.command(context_settings=CONTEXT_SETTINGS)
@click.command(context_settings=CONTEXT_SETTINGS)
@click.argument("name", type=click.Path(exists=True))
@click.option("--output-directory", default=".", show_default=True)
@click.option(
    "--output-vcf-file",
    "output_vcf_file",
    required=False,
    type=click.Path(),
    show_default=True,
    default="sequana.filter.vcf",
    help="Path to output VCF file.",
)
@click.option(
    "--output-csv-file",
    "output_csv_file",
    required=False,
    type=click.Path(),
    show_default=True,
    default="sequana.filter.csv",
    help="Path to output CSV file.",
)
@click.option(
    "--freebayes-score",
    "freebayes_score",
    default=20,
    show_default=True,
    type=float,
    help="Freebayes score threshold.",
)
@click.option(
    "--strand-ratio",
    "strand_ratio",
    default=0.2,
    show_default=True,
    type=float,
    help="Minimum strand ratio.",
)
@click.option(
    "--frequency",
    "frequency",
    default=0.1,
    show_default=True,
    type=float,
    help="Minimum allele frequency.",
)
@click.option(
    "--min-depth",
    default=10,
    show_default=True,
)
@click.option(
    "--forward-depth",
    default=3,
    show_default=True,
)
@click.option(
    "--reverse-depth",
    default=3,
    show_default=True,
)
@click.option(
    "--keep-polymorphic",
    default=True,
    show_default=True,
)
def html_report(**kwargs):
    """Create a HTML report for various type of data set

    # VCF

    The VCF module takes an input VCF file and filter it before creating an HTML report.
    The options are related to the variant quality. We assume that the VCF file was
    created with sequana_variant_calling pipeline and so relied on freebayes software.
    If so, you can provide a freebayes minimal score, remove variants below a given
    frequency (--frequency), etc (please see --help for other parameters




    """
    filename = kwargs["name"]
    if filename.endswith("fastq.gz") or filename.endswith(".fastq"):
        module = "fastq"
    elif filename.endswith("vcf"):
        module = "vcf"
    else:
        logger.error("Only extensions vcf is accepted for now")
        sys.exit(1)

    if module == "vcf":
        from sequana.freebayes_vcf_filter import VCF_freebayes
        from sequana.modules_report.variant_calling import VariantCallingModule
        from sequana.utils import config

        # use the sample name based on output HTML file otherwise, user can provide one
        report_dir = kwargs["output_directory"]
        config.output_dir = report_dir
        params = {
            "freebayes_score": kwargs["freebayes_score"],
            "frequency": kwargs["frequency"],
            "min_depth": kwargs["min_depth"],
            "forward_depth": kwargs["forward_depth"],
            "reverse_depth": kwargs["reverse_depth"],
            "strand_ratio": kwargs["strand_ratio"],
            "keep_polymorphic": kwargs["keep_polymorphic"],
        }

        v = VCF_freebayes(filename)
        filter_v = v.filter_vcf(params)
        filter_v.to_csv(kwargs["output_csv_file"])
        filter_v.to_vcf(kwargs["output_vcf_file"])

        # the HTML report
        VariantCallingModule(filter_v)


# Make the script runnable
if __name__ == "__main__":
    html_report()

1	# This file is part of Sequana software
2	#
3	# Copyright (c) 2016-2020 - Sequana Development Team
4	#
5	# Distributed under the terms of the 3-clause BSD license.
6	# The full license is in the LICENSE file, distributed with this software.
7	#
8	# website: https://github.com/sequana/sequana
9	# documentation: http://sequana.readthedocs.io
10	#
11	##############################################################################
12	import sys	3✔
13
14	import colorlog	3✔
15	import rich_click as click	3✔
16
17	from sequana.scripts.utils import CONTEXT_SETTINGS	3✔
18
19	logger = colorlog.getLogger(__name__)	3✔
20
21
22	groups = [	3✔
23	{
24	"name": "VCF options",
25	"options": [
26	"--freebayes-score",
27	"--strand-ratio",
28	"--frequency",
29	"--forward-depth",
30	"--keep-polymorphic",
31	"--reverse-depth",
32	"--min-depth",
33	"--output-vcf-file",
34	"--output-csv-file",
35	],
36	},
37	{
38	"name": "General",
39	"options": ["--output-directory"],
40	},
41	]
42
43	click.rich_click.OPTION_GROUPS["html-report"] = groups.copy()	3✔
44
45
46	# not the hyphen html-report (not html_report). this is a 'bug/feature'
47	# @click.command(context_settings=CONTEXT_SETTINGS)
48	@click.command(context_settings=CONTEXT_SETTINGS)	3✔
49	@click.argument("name", type=click.Path(exists=True))	3✔
50	@click.option("--output-directory", default=".", show_default=True)	3✔
51	@click.option(	3✔
52	"--output-vcf-file",
53	"output_vcf_file",
54	required=False,
55	type=click.Path(),
56	show_default=True,
57	default="sequana.filter.vcf",
58	help="Path to output VCF file.",
59	)
60	@click.option(	3✔
61	"--output-csv-file",
62	"output_csv_file",
63	required=False,
64	type=click.Path(),
65	show_default=True,
66	default="sequana.filter.csv",
67	help="Path to output CSV file.",
68	)
69	@click.option(	3✔
70	"--freebayes-score",
71	"freebayes_score",
72	default=20,
73	show_default=True,
74	type=float,
75	help="Freebayes score threshold.",
76	)
77	@click.option(	3✔
78	"--strand-ratio",
79	"strand_ratio",
80	default=0.2,
81	show_default=True,
82	type=float,
83	help="Minimum strand ratio.",
84	)
85	@click.option(	3✔
86	"--frequency",
87	"frequency",
88	default=0.1,
89	show_default=True,
90	type=float,
91	help="Minimum allele frequency.",
92	)
93	@click.option(	3✔
94	"--min-depth",
95	default=10,
96	show_default=True,
97	)
98	@click.option(	3✔
99	"--forward-depth",
100	default=3,
101	show_default=True,
102	)
103	@click.option(	3✔
104	"--reverse-depth",
105	default=3,
106	show_default=True,
107	)
108	@click.option(	3✔
109	"--keep-polymorphic",
110	default=True,
111	show_default=True,
112	)
113	def html_report(**kwargs):	3✔
114	"""Create a HTML report for various type of data set
115
116	# VCF
117
118	The VCF module takes an input VCF file and filter it before creating an HTML report.
119	The options are related to the variant quality. We assume that the VCF file was
120	created with sequana_variant_calling pipeline and so relied on freebayes software.
121	If so, you can provide a freebayes minimal score, remove variants below a given
122	frequency (--frequency), etc (please see --help for other parameters
123
124
125
126
127	"""
NEW 128	filename = kwargs["name"]	×
NEW 129	if filename.endswith("fastq.gz") or filename.endswith(".fastq"):	×
NEW 130	module = "fastq"	×
NEW 131	elif filename.endswith("vcf"):	×
NEW 132	module = "vcf"	×
133	else:
NEW 134	logger.error("Only extensions vcf is accepted for now")	×
NEW 135	sys.exit(1)	×
136
NEW 137	if module == "vcf":	×
NEW 138	from sequana.freebayes_vcf_filter import VCF_freebayes	×
NEW 139	from sequana.modules_report.variant_calling import VariantCallingModule	×
NEW 140	from sequana.utils import config	×
141
142	# use the sample name based on output HTML file otherwise, user can provide one
NEW 143	report_dir = kwargs["output_directory"]	×
NEW 144	config.output_dir = report_dir	×
NEW 145	params = {	×
146	"freebayes_score": kwargs["freebayes_score"],
147	"frequency": kwargs["frequency"],
148	"min_depth": kwargs["min_depth"],
149	"forward_depth": kwargs["forward_depth"],
150	"reverse_depth": kwargs["reverse_depth"],
151	"strand_ratio": kwargs["strand_ratio"],
152	"keep_polymorphic": kwargs["keep_polymorphic"],
153	}
154
NEW 155	v = VCF_freebayes(filename)	×
NEW 156	filter_v = v.filter_vcf(params)	×
NEW 157	filter_v.to_csv(kwargs["output_csv_file"])	×
NEW 158	filter_v.to_vcf(kwargs["output_vcf_file"])	×
159
160	# the HTML report
NEW 161	VariantCallingModule(filter_v)	×
162
163
164	# Make the script runnable
165	if __name__ == "__main__":	3✔
NEW 166	html_report()	×

sequana / sequana / 19136907997

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous