vis.py 13.8 KB
Newer Older
jhoogenboom's avatar
jhoogenboom committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
#!/usr/bin/env python
"""
Create a data visualisation web page or Vega graph specification.

With no optional arguments specified, a self-contained web page (HTML
file) is produced.  You can open this file in a web browser to view
interactive visualisations of your data.  The web page contains a file
selection element which can be used to select the data to be visualised.

Visualisations make use of the Vega JavaScript library
(https://vega.github.io).  The required JavaScript libraries (Vega and
D3) are embedded in the generated HTML file.  With the -O/--online
option specified, the HTML file will instead link to the latest version
of these libraries on the Internet.

Vega also supports generating visualisations on the command line, which
is useful if you wish to generate graphs automatically in your analysis
pipeline.  Specify the -V/--vega option to obtain a bare Vega graph
specification (a JSON file) instead of the full-featured HTML file.  You
can pass this file through Vega to generate a PNG or SVG image file.

If an input file is specified, the visualisation will be set up
specifically to visualise the contents of this file.  To this end, the
entire file contents are embedded in the generated visualisation.
"""
import argparse
import sys
import json
import re
import os
import cgi

from pkg_resources import resource_stream, resource_string

from ..lib import pos_int_arg

__version__ = "0.1dev"


# Default values for parameters are specified below.

# Default minimum number of reads to require.
# This value can be overridden by the -n command line option.
_DEF_THRESHOLD_ABS = 15

# Default minimum amount of reads to require, as a percentage of the
# highest allele of each marker.
# This value can be overridden by the -m command line option.
_DEF_THRESHOLD_PCT = 0.5

51 52 53 54 55
# Default minimum number of reads per orientation to require.
# This value can be overridden by the -s command line option.
_DEF_THRESHOLD_ORIENTATION = 0


jhoogenboom's avatar
jhoogenboom committed
56 57 58 59 60 61 62 63
# Default width of bars in bar graphs, in pixels.
# This value can be overridden by the -b command line option.
_DEF_BAR_WIDTH = 15

# Default amount of padding between subgraphs, in pixels.
# This value can be overridden by the -p command line option.
_DEF_SUBGRAPH_PADDING = 70

64 65 66 67
# Default graph width in pixels.
# This value can be overridden by the -w command line option.
_DEF_WIDTH = 600

jhoogenboom's avatar
jhoogenboom committed
68 69 70 71
# Default graph height in pixels.
# This value can be overridden by the -H command line option.
_DEF_HEIGHT = 400

72 73 74 75
# Default marker name matching regular expression.
# This value can be overridden by the -M command line option.
_DEF_MARKER_REGEX = ".*"

jhoogenboom's avatar
jhoogenboom committed
76 77 78 79
# Default repeat unit matching regular expression.
# This value can be overridden by the -U command line option.
_DEF_UNIT_REGEX = ".*"

jhoogenboom's avatar
jhoogenboom committed
80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
# Default data file that Vega will read when -V/--vega is specified
# without providing data to embed in the file.
# It is currently impossible to override this value.
_DEF_DATA_FILENAME = "data.csv"



_PAT_LOAD_SCRIPT = re.compile("<!--\s*BEGIN_LOAD_SCRIPT\s*-->\s*(.*?)\s*"
                              "<!--\s*END_LOAD_SCRIPT\s*-->", flags=re.DOTALL)
_PAT_LIBRARIES = re.compile("<!--\s*BEGIN_LIBRARIES\s*-->\s*(.*?)\s*"
                            "<!--\s*END_LIBRARIES\s*-->", flags=re.DOTALL)

_SCRIPT_BEGIN = '<script type="text/javascript">'
_SCRIPT_END = '</script>'

_EXTERNAL_LIBRARIES = ("vis/d3.min.js", "vis/vega.min.js")



def set_data_formula_transform_value(spec, dataname, fieldname, value):
    for data in spec["data"]:
        if data["name"] != dataname:
            continue
        if "transform" not in data:
            return False
        for transform in data["transform"]:
            if (transform["type"] == "formula" and
                    transform["field"] == fieldname):
                transform["expr"] = str(value)
                return True
    return False
#set_data_formula_transform_value


jhoogenboom's avatar
jhoogenboom committed
114 115 116 117 118 119 120 121 122 123 124 125 126 127 128
def set_axis_scale(spec, scalename, value):
    success = False
    for marks in spec["marks"]:
        if "scales" not in marks:
            continue
        for scale in marks["scales"]:
            if scale["name"] != scalename:
                continue
            scale["type"] = value;
            success = True
    return success
#set_axis_scale


def create_visualisation(vistype, infile, outfile, vega, online, tidy, min_abs,
129 130
                         min_pct, min_per_strand, bar_width, padding, marker,
                         width, height, log_scale, repeat_unit, no_alldata):
jhoogenboom's avatar
jhoogenboom committed
131 132 133 134 135 136 137 138 139 140 141
    # Get graph spec.
    spec = json.load(resource_stream(
        "fdstools", "vis/%svis/%svis.json" % (vistype, vistype)))
    if infile is not None:
        # Embed the given data file.
        spec["data"][0]["values"] = infile.read()
    elif vega:
        # Vega should load data from somewhere in headless mode.
        del spec["data"][0]["values"]
        spec["data"][0]["url"] = _DEF_DATA_FILENAME

jhoogenboom's avatar
jhoogenboom committed
142
    # Apply width, height, and padding settings.
143
    spec["width"] = width
jhoogenboom's avatar
jhoogenboom committed
144 145 146 147
    if vistype == "stuttermodel":
        set_data_formula_transform_value(spec, "yscale", "graphheight", height)
    else:
        set_data_formula_transform_value(spec, "yscale", "barwidth", bar_width)
148 149
    set_data_formula_transform_value(spec, "yscale", "subgraphoffset", padding)
    set_data_formula_transform_value(
jhoogenboom's avatar
jhoogenboom committed
150 151 152 153 154 155 156 157 158 159 160 161
        spec,
        "fitfunctions" if vistype == "stuttermodel" else "table",
        "filter_marker", "'" + marker + "'")

    # Apply type-specific settings.
    if vistype == "stuttermodel":
        set_data_formula_transform_value(
            spec, "fitfunctions", "filter_unit", "'" + repeat_unit + "'")
        set_data_formula_transform_value(
            spec, "fitfunctions", "show_all_data",
            "false" if no_alldata else "true")
    elif vistype == "sample" or vistype == "bgraw":
jhoogenboom's avatar
jhoogenboom committed
162 163 164 165
        set_data_formula_transform_value(
            spec, "table", "amplitude_threshold", min_abs)
        set_data_formula_transform_value(
            spec, "table", "amplitude_pct_threshold", min_pct)
166 167 168
    elif vistype == "profile":
        set_data_formula_transform_value(
            spec, "table", "filter_threshold", min_pct)
jhoogenboom's avatar
jhoogenboom committed
169 170
        set_data_formula_transform_value(
            spec, "table", "low", "0.001" if log_scale else "0")
171 172 173
    if vistype == "sample":
        set_data_formula_transform_value(
            spec, "table", "orientation_threshold", min_per_strand)
jhoogenboom's avatar
jhoogenboom committed
174 175 176 177 178 179 180 181 182

    # Apply axis scale settings.
    if vistype != "stuttermodel":
        if not log_scale:
            set_axis_scale(spec, "x", "linear")
        elif vistype == "sample":
            set_axis_scale(spec, "x", "sqrt")
        else:
            set_axis_scale(spec, "x", "log")
jhoogenboom's avatar
jhoogenboom committed
183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220

    # Stringify spec.
    if tidy:
        spec = json.dumps(spec, indent=2, separators=(",", ": "))
    else:
        spec = json.dumps(spec, separators=(",", ":"))

    if vega:
        # Vega graph spec is all that we were asked for.
        outfile.write(spec)
        return

    # Creating a fully self-contained HTML visualisation instead.
    html = resource_string("fdstools", "vis/%svis/index.html" % vistype)
    match = _PAT_LOAD_SCRIPT.search(html)
    if match:
        html = "".join([html[:match.start(1)],
                        _SCRIPT_BEGIN,
                        "var graph_spec=",
                        spec,
                        ";onLoadSpec(",
                        "true" if infile is not None else "false",
                        ");",
                        _SCRIPT_END,
                        html[match.end(1):]])

    if not online:
        # Replace external libraries with inline libraries.
        match = _PAT_LIBRARIES.search(html)
        if match:
            parts = [html[:match.start(1)]]
            for library in _EXTERNAL_LIBRARIES:
                parts += [_SCRIPT_BEGIN,
                          resource_string("fdstools", library),
                          _SCRIPT_END]
            parts.append(html[match.end(1):])
            html = "".join(parts)

221
    outfile.write(html)
jhoogenboom's avatar
jhoogenboom committed
222 223 224 225
#create_visualisation


def add_arguments(parser):
jhoogenboom's avatar
jhoogenboom committed
226
    parser.add_argument('type', metavar="TYPE",
jhoogenboom's avatar
jhoogenboom committed
227
        choices=("sample", "profile", "bgraw", "stuttermodel"),
228
        help="the type of data to visualise; use 'sample' to visualise "
jhoogenboom's avatar
jhoogenboom committed
229 230 231 232 233 234
             "sample data files and bgcorrect output; use 'profile' to "
             "visualise background noise profiles obtained with bgestimate, "
             "bghomstats, and bgpredict; use 'bgraw' to visualise raw "
             "background noise data obtained with bghomraw; use "
             "'stuttermodel' to visualise models of stutter obtained from "
             "stuttermodel")
jhoogenboom's avatar
jhoogenboom committed
235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256
    parser.add_argument('infile', metavar="IN", nargs="?",
        help="file containing the data to embed in the visualisation file; if "
             "not specified, HTML visualisation files will contain a file "
             "selection control, and Vega visualisation files will load data "
             "from a file called '%s'" % _DEF_DATA_FILENAME)
    parser.add_argument('outfile', metavar="OUT", nargs="?",
        type=argparse.FileType('w'),
        default=sys.stdout,
        help="file to write output to (default: write to stdout)")
    parser.add_argument('-V', '--vega', action="store_true",
        help="an HTML file containing an interactive visualisation is created "
             "by default; if this option is specified, a Vega graph "
             "specification (JSON file) is produced instead")
    parser.add_argument('-O', '--online', action="store_true",
        help="when generating an HTML visualisation file, required JavaScript "
             "libraries (D3 and Vega) are embedded in the file; if this "
             "option is specified, the HTML file will instead link to these "
             "libraries on the Internet, thereby always using the latest "
             "versions of D3 and Vega")
    parser.add_argument('-t', '--tidy', action="store_true",
        help="tidily indent the generated JSON")

257 258 259 260
    visgroup = parser.add_argument_group("visualisation options",
        description="words in [brackets] indicate applicable visualisation "
                    "types")
    visgroup.add_argument('-n', '--min-abs', metavar="N", type=pos_int_arg,
jhoogenboom's avatar
jhoogenboom committed
261
        default=_DEF_THRESHOLD_ABS,
jhoogenboom's avatar
jhoogenboom committed
262 263
        help="[sample, bgraw] only show sequences with this minimum number of "
             "reads (default: %(default)s)")
264
    visgroup.add_argument('-m', '--min-pct', metavar="PCT", type=float,
jhoogenboom's avatar
jhoogenboom committed
265
        default=_DEF_THRESHOLD_PCT,
jhoogenboom's avatar
jhoogenboom committed
266 267 268 269
        help="[sample, profile, bgraw] for sample: only show sequences with "
             "at least this percentage of the number of reads of the highest "
             "allele of a marker; for profile and bgraw: at least this "
             "percentage of the true allele (default: %(default)s)")
270 271 272 273
    visgroup.add_argument('-s', '--min-per-strand', metavar="N",
        type=pos_int_arg, default=_DEF_THRESHOLD_ORIENTATION,
        help="[sample] only show sequences with this minimum number of reads "
             "for both orientations (forward/reverse) (default: %(default)s)")
274 275
    visgroup.add_argument('-M', '--marker', metavar="REGEX",
        default=_DEF_MARKER_REGEX,
jhoogenboom's avatar
jhoogenboom committed
276 277 278 279 280 281 282 283
        help="[sample, profile, bgraw, stuttermodel] only show graphs for the "
             "markers that match the given regular expression; the default "
             "value '%(default)s' matches any marker name")
    visgroup.add_argument('-U', '--repeat-unit', metavar="REGEX",
        default=_DEF_UNIT_REGEX,
        help="[stuttermodel] only show graphs for the repeat units that match "
             "the given regular expression; the default value '%(default)s' "
             "matches any repeat unit sequence")
jhoogenboom's avatar
jhoogenboom committed
284 285 286
    visgroup.add_argument('-L', '--log-scale', action="store_true",
        help="[sample, profile, bgraw] use logarithmic scale (for sample: "
             "square root scale) instead of linear scale")
287
    visgroup.add_argument('-b', '--bar-width', metavar="N", type=pos_int_arg,
jhoogenboom's avatar
jhoogenboom committed
288
        default=_DEF_BAR_WIDTH,
jhoogenboom's avatar
jhoogenboom committed
289
        help="[sample, profile, bgraw] width of the bars in pixels (default: "
290 291
             "%(default)s)")
    visgroup.add_argument('-p', '--padding', metavar="N", type=pos_int_arg,
jhoogenboom's avatar
jhoogenboom committed
292
        default=_DEF_SUBGRAPH_PADDING,
jhoogenboom's avatar
jhoogenboom committed
293 294 295
        help="[sample, profile, bgraw, stuttermodel] amount of padding (in "
             "pixels) between graphs of different markers/alleles (default: "
             "%(default)s)")
296 297
    visgroup.add_argument('-w', '--width', metavar="N", type=pos_int_arg,
        default=_DEF_WIDTH,
jhoogenboom's avatar
jhoogenboom committed
298 299 300 301 302 303 304 305
        help="[sample, profile, bgraw, stuttermodel] width of the graph area "
             "in pixels (default: %(default)s)")
    visgroup.add_argument('-H', '--height', metavar="N", type=pos_int_arg,
        default=_DEF_HEIGHT,
        help="[stuttermodel] height of the graph area in pixels (default: "
             "%(default)s)")
    visgroup.add_argument('-A', '--no-alldata', action="store_true",
        help="[stuttermodel] if specified, show only marker-specific fits")
jhoogenboom's avatar
jhoogenboom committed
306 307 308 309 310 311 312 313 314 315 316 317 318
#add_arguments


def run(args):
    if args.infile == "-" and not sys.stdin.isatty():
        # User appears to want to pipe data in.
        args.infile = sys.stdin
    if (args.infile is not None and args.outfile == sys.stdout
            and not os.path.exists(args.infile)):
        # One filename given, and it does not exist.  Assume outfile.
        args.outfile = open(args.infile, 'w')
        args.infile = None

319 320 321 322 323
    if args.outfile.isatty():
        raise ValueError("Please specify a file name to write the %s to." %
                         ("Vega graph specification (JSON format)" if args.vega
                          else "HTML document"))

jhoogenboom's avatar
jhoogenboom committed
324 325 326 327 328 329
    if args.infile is not None and args.infile != sys.stdin:
        # Open the specified input file.
        args.infile = open(args.infile, 'r')

    create_visualisation(args.type, args.infile, args.outfile, args.vega,
                         args.online, args.tidy, args.min_abs, args.min_pct,
330 331 332
                         args.min_per_strand, args.bar_width, args.padding,
                         args.marker, args.width, args.height, args.log_scale,
                         args.repeat_unit, args.no_alldata)
jhoogenboom's avatar
jhoogenboom committed
333
#run