view plotly_blast_tool/plotlyblast.xml @ 1:6fbd48e9c950 draft

ready for beta testing?

<tool name="plotlyblast" id="plotlyblast" version="3.0">
  <!--Source in git at: https://github.com/fubar2/galaxy_tf_overlay-->
  <!--Created by [e-mail] at 26/07/2023 15:45:12 using the Galaxy Tool Factory.-->
  <description>Plotly plot generator</description>
  <requirements>
    <requirement type="package">pandas</requirement>
    <requirement type="package">plotly</requirement>
  </requirements>
  <stdio>
    <exit_code range="1:" level="fatal"/>
  </stdio>
  <version_command><![CDATA[echo "3.0"]]></version_command>
  <command><![CDATA[python
$runme
--input_tab
$input_tab
--htmlout
$htmlout
--xcol
"$xcol"
--ycol
"$ycol"
--colourcol
"$colourcol"
--hovercol
"$hovercol"
--title
"$title"
--header
"$header"]]></command>
  <configfiles>
    <configfile name="runme"><![CDATA[#raw

import argparse
import sys
import math
import plotly.express as px
import pandas as pd
# Ross Lazarus July 2023
# based on various plotly tutorials
# Expects a blastn search result passed in as the first command line parameter.
parser = argparse.ArgumentParser()
a = parser.add_argument
a('--input_tab',default='')
a('--header',default='qaccver,saccver,piden,length,mismatch,gapopen,qstart,qend,sstart,send,evalue,bitscore,sallseqid,score,nident,positive,gaps,ppos,qframe,sframe,qseq,sseq,qlen,slen,salltitles')
a('--htmlout',default="test_run.html")
a('--xcol',default='gaps')
a('--ycol',default='qseq')
a('--colourcol',default='qaccver')
a('--hovercol',default='qseq')
a('--title',default='test blast plot of mismatch by bitscore coloured by qaccver ')
args = parser.parse_args()
df = pd.read_csv(args.input_tab, sep='\t')
NCOLS = df.columns.size
MAXLEN = 30 # tricky way to truncate long axis tickmarks
defaultcols = ['col%d' % (x+1) for x in range(NCOLS)]
if len(args.header.strip()) > 0:
    newcols = args.header.split(',')
    if len(newcols) == NCOLS:
        df.columns = newcols
    else:
        sys.stderr.write('#### Supplied header %s has %d comma delimited header names - does not match the input tabular file %d columns - using col1,...coln' % (args.header, len(newcols), NCOLS))
        df.columns = defaultcols
else:
    df.columns = defaultcols
df['evalue'] = [-math.log(x + 1e-308) for x in df['evalue']] # convert so large values reflect statistical surprise
if len(args.colourcol.strip()) == 0:
    fig = px.scatter(df, x=args.xcol, y=args.ycol, hover_name=args.hovercol)
else:
    fig = px.scatter(df, x=args.xcol, y=args.ycol, color=args.colourcol, hover_name=args.hovercol)
if args.title:
    ftitle=dict(text=args.title, font=dict(size=40), automargin=True)
    fig.update_layout(title=ftitle)
for scatter in fig.data:
    scatter['x'] = [str(x)[:MAXLEN] + '..' if len(str(x)) > MAXLEN else x for x in scatter['x']]
    scatter['y'] = [str(x)[:MAXLEN] + '..' if len(str(x)) > MAXLEN else x for x in scatter['y']]
    if len(args.colourcol.strip()) == 0:
        sl = str(scatter['legendgroup'])
        if len(sl) > MAXLEN:
            scatter['legendgroup'] = sl[:MAXLEN]
fig.write_html(args.htmlout)


#end raw]]></configfile>
  </configfiles>
  <inputs>
    <param name="input_tab" type="data" optional="false" label="Tabular input file to plot" help="" format="tabular" multiple="false"/>
    <param name="xcol" type="text" value="qaccver" label="x axis for plat" help=""/>
    <param name="ycol" type="text" value="bitscore" label="y axis for plot" help=""/>
    <param name="colourcol" type="text" value="piden" label="column containing a groupable variable for colour. Default none." help="Adds a legend so choose wisely "/>
    <param name="hovercol" type="text" value="qseq" label="columname for hover string" help=""/>
    <param name="title" type="text" value="Qaccver by bitscore coloured by pident. Hover shows blast match." label="Title for the plot" help="Special characters will probably be escaped so do not use them"/>
    <param name="header" type="text" value="qaccver,saccver,piden,length,mismatch,gapopen,qstart,qend,sstart,send,evalue,bitscore,sallseqid,score,nident,positive,gaps,ppos,qframe,sframe,qseq,sseq,qlen,slen,salltitles" label="Use this comma delimited list of column header names for this tabular file. Default is None when col1...coln will be used" help="Default for Galaxy blast outputs with 25 columns. The column names supplied for xcol, ycol, hover and colour MUST match either the supplied list, or if none, col1...coln."/>
  </inputs>
  <outputs>
    <data name="htmlout" format="html" label="htmlout" hidden="false"/>
  </outputs>
  <tests>
    <test>
      <output name="htmlout" value="htmlout_sample" compare="sim_size" delta="5000"/>
      <param name="input_tab" value="input_tab_sample"/>
      <param name="xcol" value="qaccver"/>
      <param name="ycol" value="bitscore"/>
      <param name="colourcol" value="piden"/>
      <param name="hovercol" value="qseq"/>
      <param name="title" value="Qaccver by bitscore coloured by pident. Hover shows blast match."/>
      <param name="header" value="qaccver,saccver,piden,length,mismatch,gapopen,qstart,qend,sstart,send,evalue,bitscore,sallseqid,score,nident,positive,gaps,ppos,qframe,sframe,qseq,sseq,qlen,slen,salltitles"/>
    </test>
  </tests>
  <help><![CDATA[

This is a specialised version of a generic tabular file plotter tool plotlytabular



.. class:: warningmark

NOTE: Long strings in x and y tickmarks WILL BE TRUNCATED if they are too long - ".." is added to indicate truncation - otherwise some plots are squished.



.. class:: warningmark

NOTE: Blast evalues WILL BE TRANSFORMED using -log10(evalue), so they are scaled in a way that doesn't confuse plotly.express with the tiny values.



----



The main reason to run this tool is to have an interactive hover text specified so it appears when hovering over each data point to supply useful information. 



Assumes you want a hover display for an interactive plot to be informative



Column names are auto-generated as col1,...coln unless a comma separated list of column names is supplied as the header parameter.



For example, using a Galaxy blastn output with 25 columns, the following comma delimited string supplied as the "header" parameter will match the names of each column.

   qaccver,saccver,piden,length,mismatch,gapopen,qstart,qend,sstart,send,evalue,bitscore,sallseqid,score,nident,positive,gaps,ppos,qframe,sframe,qseq,sseq,qlen,slen,salltitles



When a header is supplied, the xcol and other column names must match one of those supplied column names.

So for example, xcol = "qaccver" for the blastn header example rather than xcol = "col1" when no header is supplied.



Relies on Plotly python code released under the MIT licence: https://github.com/plotly/plotly.py/blob/master/LICENSE.txt



.. image:: pbsample.png

    :height: 400

    :width: 400

 

------


Script::

    import argparse
    import sys
    import math
    import plotly.express as px
    import pandas as pd
    # Ross Lazarus July 2023
    # based on various plotly tutorials
    # Expects a blastn search result passed in as the first command line parameter.
    parser = argparse.ArgumentParser()
    a = parser.add_argument
    a('--input_tab',default='')
    a('--header',default='qaccver,saccver,piden,length,mismatch,gapopen,qstart,qend,sstart,send,evalue,bitscore,sallseqid,score,nident,positive,gaps,ppos,qframe,sframe,qseq,sseq,qlen,slen,salltitles')
    a('--htmlout',default="test_run.html")
    a('--xcol',default='gaps')
    a('--ycol',default='qseq')
    a('--colourcol',default='qaccver')
    a('--hovercol',default='qseq')
    a('--title',default='test blast plot of mismatch by bitscore coloured by qaccver ')
    args = parser.parse_args()
    df = pd.read_csv(args.input_tab, sep='\t')
    NCOLS = df.columns.size
    MAXLEN = 30 # tricky way to truncate long axis tickmarks
    defaultcols = ['col%d' % (x+1) for x in range(NCOLS)]
    if len(args.header.strip()) > 0:
        newcols = args.header.split(',')
        if len(newcols) == NCOLS:
            df.columns = newcols
        else:
            sys.stderr.write('#### Supplied header %s has %d comma delimited header names - does not match the input tabular file %d columns - using col1,...coln' % (args.header, len(newcols), NCOLS))
            df.columns = defaultcols
    else:
        df.columns = defaultcols
    df['evalue'] = [-math.log(x + 1e-308) for x in df['evalue']] # convert so large values reflect statistical surprise
    if len(args.colourcol.strip()) == 0:
        fig = px.scatter(df, x=args.xcol, y=args.ycol, hover_name=args.hovercol)
    else:
        fig = px.scatter(df, x=args.xcol, y=args.ycol, color=args.colourcol, hover_name=args.hovercol)
    if args.title:
        ftitle=dict(text=args.title, font=dict(size=40), automargin=True)
        fig.update_layout(title=ftitle)
    for scatter in fig.data:
        scatter['x'] = [str(x)[:MAXLEN] + '..' if len(str(x)) > MAXLEN else x for x in scatter['x']]
        scatter['y'] = [str(x)[:MAXLEN] + '..' if len(str(x)) > MAXLEN else x for x in scatter['y']]
        if len(args.colourcol.strip()) == 0:
            sl = str(scatter['legendgroup'])
            if len(sl) > MAXLEN:
                scatter['legendgroup'] = sl[:MAXLEN]
    fig.write_html(args.htmlout)

]]></help>
  <citations>
    <citation type="doi">10.1093/bioinformatics/bts573</citation>
  </citations>
</tool>