--- a/metaphlan.xml	Wed Feb 02 22:00:29 2022 +0000
+++ b/metaphlan.xml	Mon Feb 27 06:59:29 2023 +0000
@@ -107,7 +107,7 @@
 #if $inputs.db.db_selector == "history"
 mkdir 'ref_db'
 &&
-bowtie2-build '$inputs.db.bowtie2db' 'ref_db/custom_db'
+bowtie2-build --large-index '$inputs.db.bowtie2db' 'ref_db/custom_db'
 &&
 python '$__tool_directory__/customizemetadata.py'
     transform_json_to_pkl
@@ -123,9 +123,15 @@
     --read_min_len $inputs.in.read_min_len
     --bt2_ps '$inputs.in.mapping.bt2_ps'
     --min_mapq_val $inputs.in.mapping.min_mapq_val
+    #if $ext == "sam"
+        --nreads \$(cat '$file_path' | grep -c -v '^@')
+    #end if
 #else
     '$inputs.in.in'
     --input_type '$inputs.in.selector'
+    #if $inputs.in.selector == "sam"
+        --nreads \$(cat '$inputs.in.in' | grep -c -v '^@')
+    #end if
 #end if
 #if $inputs.db.db_selector == "cached"
     --bowtie2db '$inputs.db.cached_db.fields.path'
@@ -257,15 +263,16 @@
                     <option value="history">From history</option>
                 </param>
                 <when value="cached">
-                    <param name="cached_db" label="Cached database with clade-specific marker genes" type="select">
-                        <options from_data_table="metaphlan_database">
-                            <validator message="No MetaPhlAn database is available" type="no_options" />
+                    <param name="cached_db" type="select" label="Cached database with clade-specific marker genes">
+                        <options from_data_table="@IDX_DATA_TABLE@">
+                            <filter type="static_value" column="4" value="@IDX_VERSION@"/>
+                            <validator message="No compatible MetaPhlAn database is available" type="no_options"/>
                         </options>
                     </param>
                 </when>
                 <when value="history">
                     <param argument="--bowtie2db" type="data" format="fasta" label="Database with clade-specific marker genes from history"/>
-                    <param argument="--mpa_pkl" type="data" format="json" label="Metadata associate to the database with clade-specific marker genes from history"/>
+                    <param argument="--mpa_pkl" type="data" format="json" label="Metadata associated to the database with clade-specific marker genes from history"/>
                 </when>
             </conditional>
         </section>
@@ -334,7 +341,7 @@
             <param argument="--sample_id" type="text" value="Metaphlan_Analysis" label="Sample ID for this analysis"/>
             <param argument="--use_group_representative" type='boolean' checked="false" truevalue='--use_group_representative' falsevalue='' 
                 label="Use a species as representative for species groups?"/>
-            <param name="legacy_output" argument="--legacy-output" type='boolean' checked="false" truevalue='--legacy-output' falsevalue='' 
+            <param argument="--legacy-output" type='boolean' checked="false" truevalue='--legacy-output' falsevalue='' 
                 label="Old MetaPhlAn2 two columns output?"/>
             <param argument="--CAMI_format_output" type='boolean' checked="false" truevalue='--CAMI_format_output' falsevalue='' 
                 label="Report the profiling using the CAMI output format?"/>
@@ -408,7 +415,7 @@
             </section>
             <output name="output_file" ftype="tabular">
                 <assert_contents>
-                    <has_text text="UNKNOWN"/>
+                    <has_text text="UNCLASSIFIED"/>
                 </assert_contents>
             </output>
             <output name="bowtie2out" ftype="tabular">
@@ -429,6 +436,7 @@
                 </assert_contents>
             </output>
             <output_collection name="levels" type="list" >
+            
                 <element name="all" ftype="tabular">
                     <assert_contents>
                         <has_text text="class"/>
@@ -436,21 +444,6 @@
                         <has_n_lines n="1"/>
                     </assert_contents>
                 </element>
-                <element name="kingdom" ftype="tabular">
-                    <assert_contents>
-                        <has_text text="kingdom_id"/>
-                        <has_n_columns n="3"/>
-                        <has_n_lines n="1"/>
-                    </assert_contents>
-                </element>
-                <element name="phylum" ftype="tabular">
-                    <assert_contents>
-                        <has_text text="phylum_id"/>
-                        <not_has_text text="kingdom_id"/>
-                        <has_n_columns n="3"/>
-                        <has_n_lines n="1"/>
-                    </assert_contents>
-                </element>
                 <element name="class" ftype="tabular">
                     <assert_contents>
                         <has_text text="class_id"/>
@@ -459,14 +452,6 @@
                         <has_n_lines n="1"/>
                     </assert_contents>
                 </element>
-                <element name="order" ftype="tabular">
-                    <assert_contents>
-                        <has_text text="order_id"/>
-                        <not_has_text text="class_id"/>
-                        <has_n_columns n="3"/>
-                        <has_n_lines n="1"/>
-                    </assert_contents>
-                </element>
                 <element name="family" ftype="tabular">
                     <assert_contents>
                         <has_text text="family_id"/>
@@ -483,6 +468,29 @@
                         <has_n_lines n="1"/>
                     </assert_contents>
                 </element>
+                <element name="kingdom" ftype="tabular">
+                    <assert_contents>
+                        <has_text text="kingdom_id"/>
+                        <has_n_columns n="3"/>
+                        <has_n_lines n="1"/>
+                    </assert_contents>
+                </element>
+                <element name="order" ftype="tabular">
+                    <assert_contents>
+                        <has_text text="order_id"/>
+                        <not_has_text text="class_id"/>
+                        <has_n_columns n="3"/>
+                        <has_n_lines n="1"/>
+                    </assert_contents>
+                </element>
+                <element name="phylum" ftype="tabular">
+                    <assert_contents>
+                        <has_text text="phylum_id"/>
+                        <not_has_text text="kingdom_id"/>
+                        <has_n_columns n="3"/>
+                        <has_n_lines n="1"/>
+                    </assert_contents>
+                </element>
                 <element name="species" ftype="tabular">
                     <assert_contents>
                         <has_text text="species_id"/>
@@ -503,7 +511,8 @@
             <output name="krona_output_file" ftype="tabular">
                 <assert_contents>
                     <not_has_text text="k__Bacteria"/>
-                    <has_n_lines n="0"/>
+                    <has_n_lines n="1" delta="1"/>
+                    <has_size value="1" delta="1"/>
                 </assert_contents>
             </output>
         </test>
@@ -581,21 +590,6 @@
                         <has_n_columns n="17"/>
                     </assert_contents>
                 </element>
-                <element name="kingdom" ftype="tabular">
-                    <assert_contents>
-                        <has_text text="kingdom_id"/>
-                        <has_text text="Bacteria"/>
-                        <has_n_columns n="3"/>
-                    </assert_contents>
-                </element>
-                <element name="phylum" ftype="tabular">
-                    <assert_contents>
-                        <has_text text="phylum_id"/>
-                        <not_has_text text="kingdom_id"/>
-                        <has_text text="Firmicutes"/>
-                        <has_n_columns n="3"/>
-                    </assert_contents>
-                </element>
                 <element name="class" ftype="tabular">
                     <assert_contents>
                         <has_text text="class_id"/>
@@ -604,14 +598,6 @@
                         <has_n_columns n="3"/>
                     </assert_contents>
                 </element>
-                <element name="order" ftype="tabular">
-                    <assert_contents>
-                        <has_text text="order_id"/>
-                        <not_has_text text="class_id"/>
-                        <has_text text="Propionibacteriales"/>
-                        <has_n_columns n="3"/>
-                    </assert_contents>
-                </element>
                 <element name="family" ftype="tabular">
                     <assert_contents>
                         <has_text text="family_id"/>
@@ -628,6 +614,29 @@
                         <has_n_columns n="3"/>
                     </assert_contents>
                 </element>
+                <element name="kingdom" ftype="tabular">
+                    <assert_contents>
+                        <has_text text="kingdom_id"/>
+                        <has_text text="Bacteria"/>
+                        <has_n_columns n="3"/>
+                    </assert_contents>
+                </element>
+                <element name="order" ftype="tabular">
+                    <assert_contents>
+                        <has_text text="order_id"/>
+                        <not_has_text text="class_id"/>
+                        <has_text text="Propionibacteriales"/>
+                        <has_n_columns n="3"/>
+                    </assert_contents>
+                </element>
+                <element name="phylum" ftype="tabular">
+                    <assert_contents>
+                        <has_text text="phylum_id"/>
+                        <not_has_text text="kingdom_id"/>
+                        <has_text text="Firmicutes"/>
+                        <has_n_columns n="3"/>
+                    </assert_contents>
+                </element>
                 <element name="species" ftype="tabular">
                     <assert_contents>
                         <has_text text="species_id"/>
@@ -981,20 +990,6 @@
                         <has_n_columns n="9"/>
                     </assert_contents>
                 </element>
-                <element name="kingdom" ftype="tabular">
-                    <assert_contents>
-                        <has_text text="kingdom"/>
-                        <has_text text="Bacteria"/>
-                        <has_n_columns n="2"/>
-                    </assert_contents>
-                </element>
-                <element name="phylum" ftype="tabular">
-                    <assert_contents>
-                        <has_text text="phylum"/>
-                        <has_text text="Firmicutes"/>
-                        <has_n_columns n="2"/>
-                    </assert_contents>
-                </element>
                 <element name="class" ftype="tabular">
                     <assert_contents>
                         <has_text text="class"/>
@@ -1002,13 +997,6 @@
                         <has_n_columns n="2"/>
                     </assert_contents>
                 </element>
-                <element name="order" ftype="tabular">
-                    <assert_contents>
-                        <has_text text="order"/>
-                        <has_text text="Propionibacteriales"/>
-                        <has_n_columns n="2"/>
-                    </assert_contents>
-                </element>
                 <element name="family" ftype="tabular">
                     <assert_contents>
                         <has_text text="family"/>
@@ -1023,6 +1011,27 @@
                         <has_n_columns n="2"/>
                     </assert_contents>
                 </element>
+                <element name="kingdom" ftype="tabular">
+                    <assert_contents>
+                        <has_text text="kingdom"/>
+                        <has_text text="Bacteria"/>
+                        <has_n_columns n="2"/>
+                    </assert_contents>
+                </element>
+                <element name="order" ftype="tabular">
+                    <assert_contents>
+                        <has_text text="order"/>
+                        <has_text text="Propionibacteriales"/>
+                        <has_n_columns n="2"/>
+                    </assert_contents>
+                </element>
+                <element name="phylum" ftype="tabular">
+                    <assert_contents>
+                        <has_text text="phylum"/>
+                        <has_text text="Firmicutes"/>
+                        <has_n_columns n="2"/>
+                    </assert_contents>
+                </element>
                 <element name="species" ftype="tabular">
                     <assert_contents>
                         <has_text text="species"/>
@@ -1051,7 +1060,7 @@
 ============
 
 MetaPhlAn is a computational tool for profiling the composition of microbial communities (Bacteria, 
-Archaea and Eukaryotes) from metagenomic shotgun sequencing data (i.e. not 16S) with species-level. 
+Archaea and Eukaryotes) from metagenomic shotgun sequencing data (i.e. not 16S) at species-level. 
 
 MetaPhlAn relies on ~1.1M unique clade-specific marker genes identified from ~100,000 reference genomes 
 (~99,500 bacterial and archaeal and ~500 eukaryotic), allowing:
@@ -1077,8 +1086,8 @@
 aggregated marker information. Such capability comes with several slightly different flavours and 
 are a way to perform strain tracking and comparison across multiple samples.
 
-Usually, MetaPhlAn is first ran with the default parameter for the type of analysis to profile the 
-species present in the community, and then a strain-level profiling can be performed to zoom-in into 
+Usually, MetaPhlAn is first ran with default parameters for the type of analysis to profile the 
+species present in the community, and then a strain-level profiling can be performed to zoom-in on 
 specific species of interest. This operation can be performed quickly as it exploits the bowtie2out 
 intermediate file saved during the execution of the default analysis type.
 
@@ -1087,16 +1096,16 @@
 
 Metaphlan takes as input either:
 
-- one or several sequence files in Fasta, FastQ (compressed or not)
+- one or several sequence files in Fasta, FastQ (whether compressed or not)
 - a BowTie2 produced SAM file
 - an intermediary mapping file of the microbiota generated by a previous MetaPhlAn run
 
-It also need the reference database, which can be locally installed or customized using the dedicated tools.
+It also needs the reference database, which can be locally installed or customized using the dedicated tools.
 
 Outputs
 =======
 
-The main output file is a tab-separated file with the predicted taxon relative abundances.
+The main output is a tab-separated file with the predicted taxon relative abundances.
 
 It also generates a BIOM file and some intermediate files (SAM and BowTie2out) if sequence files are given as inputs.
 
@@ -1106,7 +1115,7 @@
 
 To get more information about MetaPhlAn usage and use cases, please refer to the `Metaphlan documentation`_.
 
-.. _Metaphlan documentation: https://github.com/biobakery/MetaPhlAn/wiki/MetaPhlAn-3.0#basic-usage
+.. _Metaphlan documentation: https://github.com/biobakery/MetaPhlAn/wiki/MetaPhlAn-4#Basic-Usage
 
     ]]></help>
     <expand macro="citations"/>