Mercurial > repos > nml > csvtk_replace

<tool id="csvtk_replace" name="csvtk-replace" version="@VERSION@+@GALAXY_VERSION@">
    <description> data of selected columns by regular expression</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements" />
    <expand macro="version_cmd" />
    <command detect_errors="exit_code"><![CDATA[

###################
## Start Command ##
###################
csvtk replace --num-cpus "\${GALAXY_SLOTS:-1}"

    ## Add additional flags as specified ##
    #######################################
    $ignore_case
    $global_param.illegal_rows
    $global_param.empty_rows
    $global_param.header
    $global_param.lazy_quotes

    ## Set Tabular input/output flag if first input is tabular ##
    #############################################################
    #if $in_1.is_of_type("tabular"):
        -t -T
    #end if

    ## Set input files ##
    #####################
    '$in_1'

    ## Specify fields to use ##
    ###########################
    -F -f '${column_text.in_text}'

    ## Specific Commands ##
    #######################
    -p '($pattern_string)'
    -r '$replacement_string'

    #if $input_kv
        -k '$input_kv'
    #end if

    #if $fill.how_fill == "key"
        -K
    #elif $fill.how_fill == "string"
        --key-miss-repl '$fill.fill_string'
    #end if

    ## To output ##
    ###############
    > replaced

    ]]></command>
    <inputs>
        <expand macro="singular_input"/>
        <expand macro="fields_input" />
        <param name="pattern_string" type="text" argument="-p"
            label="Pattern Regex"
            help="Regex to search column for. Input is structured as '(YOUR_INPUT_HERE)' so if your regex was just a period it would look like '(.)' as an input.">
            <expand macro="text_sanitizer" />
        </param>
        <param name="replacement_string" type="text" argument="-r"
            label="Replacement String">
            <help>
                <![CDATA[
        String to replace found data. Supports capture variables and special replacement symbols.

        - Capture Variables: $1 represents the text of the first submatch
        - {nr} inserts a record number starting from 1
        - {kv} uses corresponding value of the key (captured variable $n) of a key-value file

        If using the special replacement symbols, the capture variable must be specified as ${1}!
            ]]>
            </help>
            <expand macro="text_sanitizer" />
        </param>
        <param name="input_kv" type="data" format="tsv,tabular" argument="-k"
            optional="true"
            label="Key/Value file for replacement string"
            help="Only specify a file if using {kv} in replacement string. The file must be tab delimited with one key/value pair per line.
            An example can be found in the help section below"
        />
        <conditional name="fill">
            <param name="how_fill" type="select" label="Fill columns that don't get any regex matches">
                <option value="no">No</option>
                <option value="key">Yes - Fill with Original Value</option>
                <option value="string">Yes - Fill with String</option>
            </param>
            <when value="no" />
            <when value="key" />
            <when value="string" >
                <param name="fill_string" type="text" value="NA" argument="--key-miss-repl" label="Fill string">
                    <expand macro="text_sanitizer" />
                </param>
            </when>
        </conditional>
        <expand macro="ignore_case" />
        <expand macro="global_parameters" />
    </inputs>
    <outputs>
        <data format_source="in_1" name="replaced" from_work_dir="replaced" label='${in_1.name} with column ${column_text.in_text} replaced' />
    </outputs>
    <tests>
        <test>
            <param name="in_1" value="replace_input.csv" />
            <conditional name="column_text">
                <param name="select" value="string" />
                <param name="in_text" value="2" />
            </conditional>
            <param name="pattern_string" value=".+" />
            <param name="replacement_string" value="{nr}-$1" />
            <output name="replaced" value="replace_1.csv" />
        </test>
        <test>
            <param name="in_1" value="replace_input.csv" />
            <conditional name="column_text">
                <param name="select" value="string" />
                <param name="in_text" value="2" />
            </conditional>
            <param name="pattern_string" value=".+" />
            <param name="replacement_string" value="{kv}" />
            <param name="input_kv" value="kv.txt" />
            <conditional name="fill">
                <param name="how_fill" value="key" />
            </conditional>
            <output name="replaced" value="replace_2.csv" />
        </test>
    </tests>
    <help><![CDATA[

Csvtk - Replace Help
--------------------

Info
####
Csvtk-replace is a tool that uses Regular Expressions (Regex) to match data in the specified column and replace it with the replacement string.
Non-matched columns can be kept or filled with the Regex key or an input string

The regex input for this tool is structured such that your regular expression **does not** need to start with with quotes or brackets. You can
start your expression with a `^` or just go straight into it

For example:

::

    Using `.+` as an input would be used in the code as '(.+)'

    Using ^(.+)$ as an input would yield an input in the code as '(^(.+)$)'

.. class:: warningmark

    Single quotes are not allowed in text inputs!

-----


@HELP_INPUT_DATA@


Usage
#####
You can use csvtk replace to any matched regex expressions with your input replacement string.

The replacement string has some unique properties that you can use too to better replace your data:

- Replacement supports capture variables, like $1 which represents the text of the first submatch of the Regex

- \{nr} can be used to assign ascending numbers starting from 1 to each column

- \{kv} can be used to get the value of the key (captured variable $n) or a key-value file

A good Regular expressions cheat sheet that you can use to help yourself build regular expressions can be found at:
https://regexr.com/

**Replace Examples**

1. Replacement with {nr} and $1

Input file:

+---------+--------+
| Name    | Animal |
+=========+========+
| Bud     | Dog    |
+---------+--------+
| Mittens | Cat    |
+---------+--------+

Now if our regex was set to '.*' on column 2 and our replacement string was set to '{nr}-$1', the following output would be observed:

+---------+--------+
| Name    | Animal |
+=========+========+
| Bud     | 1-Dog  |
+---------+--------+
| Mittens | 2-Cat  |
+---------+--------+

---------------

2. Replacement with {kv} file

Suppose you set up a key-value TAB separated file that looked as such:

::

    Key     Value
    Dog     Big
    Cat     Small

And had a similar input file:

+---------+--------+
| Name    | Animal |
+=========+========+
| Bud     | Dog    |
+---------+--------+
| Mittens | Cat    |
+---------+--------+
| Fuzzy   | Gerbil |
+---------+--------+

Now if the regex was '.*' on column 2 with the replacement string as '{kv}'. Your output would look as such with 'No' fill specified:

+---------+--------+
| Name    | Animal |
+=========+========+
| Bud     | Big    |
+---------+--------+
| Mittens | Small  |
+---------+--------+
| Fuzzy   |        |
+---------+--------+

If you wanted to fill the blank cell you could set it to either:

- String - the string you input (ex. 'NA') would fill up the blank cell.

- Original value - would change the blank cell to 'Gerbil'

----

If your having trouble with the regular expressions, please play around with a builder, there are many others online
and they are great resources to improve your regex statements or test them before use!

----

@HELP_END_STATEMENT@


    ]]></help>
    <expand macro="citations" />
</tool>