diff ludwig_train.xml @ 0:f0be10937f5c draft default tip

planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
author goeckslab
date Tue, 07 Jan 2025 22:44:09 +0000
parents
children
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/ludwig_train.xml	Tue Jan 07 22:44:09 2025 +0000
@@ -0,0 +1,134 @@
+<tool id="ludwig_train" name="Ludwig Train" version="@VERSION@" profile="@PROFILE@">
+    <description>trains a deep learning model</description>
+    <macros>
+        <import>ludwig_macros.xml</import>
+    </macros>
+    <expand macro="python_requirements_gpu" />
+    <expand macro="macro_stdio" />
+    <version_command>echo "@VERSION@"</version_command>
+    <command>
+        <![CDATA[
+            #import re
+            #if $config
+            cp '$config' "./config.yml";
+            #end if
+
+            #if $dataset
+            #set $sanitized_dataset = re.sub('[^\w\-_\.]', '_', $dataset.element_identifier.strip())
+            ln -sf '$dataset' "./${sanitized_dataset}";
+            #end if
+
+            #if $training_set
+            #set $sanitized_training_set = re.sub('[^\w\-_\.]', '_', $training_set.element_identifier.strip())
+            ln -sf '$training_set' "./${sanitized_training_set}";
+            #end if
+
+            #if $validation_set
+            #set $sanitized_validation_set = re.sub('[^\w\-_\.]', '_', $validation_set.element_identifier.strip())
+            ln -sf '$validation_set' "./${sanitized_validation_set}";
+            #end if
+
+            #if $test_set
+            #set $sanitized_test_set = re.sub('[^\w\-_\.]', '_', $test_set.element_identifier.strip())
+            ln -sf '$test_set' "./${sanitized_test_set}";
+            #end if
+
+            #if $raw_data
+                unzip -o -q '$raw_data' -d ./;
+            #end if
+
+            python '$__tool_directory__/ludwig_train.py'
+                #if $config
+                --config "./config.yml"
+                #end if
+                #if $model_load_path
+                --model_load_path '${model_load_path.extra_files_path}'
+                #end if
+                #if $dataset
+                --dataset "./${sanitized_dataset}"
+                #end if
+                #if $training_set
+                --training_set "./${sanitized_training_set}"
+                #end if
+                #if $validation_set
+                --validation_set "./${sanitized_validation_set}"
+                #end if
+                #if $test_set
+                --test_set "./${sanitized_test_set}"
+                #end if
+                #if $training_set_metadata
+                --training_set_metadata '$training_set_metadata'
+                #end if
+                #if $disable_parallel_threads
+                --disable_parallel_threads
+                #end if
+                --output_directory "."
+                --data_format '$data_format'
+                --random_seed $random_seed
+                --backend local &&
+
+            mkdir -p '${output_model.extra_files_path}' &&
+            cp -r experiment_run/model/*.json experiment_run/model/model_weights '${output_model.extra_files_path}' &&
+            mkdir -p '$output_report.extra_files_path' &&
+            cp experiment_run/*.json '$output_report.extra_files_path' &&
+            cp -r visualizations '$output_report.extra_files_path' &&
+
+            echo "Training is Done!"
+        ]]>
+    </command>
+    <configfiles>
+        <inputs name="inputs" />
+    </configfiles>
+    <inputs>
+        <param name="config" type="data" format="yaml" label="Select the dataset containing model configuration" />
+        <param name="model_load_path" type="data" format="ludwig_model" optional="true" label="Load a pretrained model as initialization" help="Optional." />
+        <param name="dataset" type="data" format="tabular,csv,h5,json,txt" optional="true" label="Input dataset" />
+        <param name="training_set" type="data" format="tabular,csv,h5,json" optional="true" label="Input traning dataset" />
+        <param name="validation_set" type="data" format="tabular,csv,h5,json" optional="true" label="Input validation dataset" />
+        <param name="test_set" type="data" format="tabular,csv,h5,json" optional="true" label="Input test dataset" />
+        <param name="training_set_metadata" type="data" format="json" optional="true" label="Training set metadata" />
+        <param name="data_format" type="select" label="Data format">
+            <option value="auto" selected="true">auto</option>
+            <option value="tsv">tsv</option>
+            <option value="csv">csv</option>
+            <option value="h5">h5</option>
+            <option value="json">json</option>
+        </param>
+        <param name="random_seed" type="integer" value="42" label="Randomness seed" min="0" max="999999" />
+        <param name="disable_parallel_threads" type="boolean" checked="false" label="Whether to disable parallel threads for reproducibility?" />
+        <param name="raw_data" type="data" format="zip" optional="true" label="Raw data" help="Optional. Needed for images."/>
+    </inputs>       
+    <outputs>
+        <data format="ludwig_model" name="output_model" label="${tool.name} model on ${on_string}" />
+        <data format="html" name="output_report" from_work_dir="ludwig_train_report.html" label="${tool.name} report on ${on_string}" />
+        <collection type="list" name="output_pred_csv" label="${tool.name} training stat on ${on_string}" >
+            <discover_datasets pattern="(?P&lt;designation&gt;.+)\.json" format="json" directory="experiment_run" />
+        </collection>
+    </outputs>
+    <tests>
+        <test>
+            <param name="dataset" value="temperature_la.csv" ftype="csv" />
+            <param name="config" value="temperature_config.yml" ftype="yaml" />
+            <param name="data_format" value="csv" />
+            <output name="output_report" file="ludwig_train_report_test.html" compare="sim_size" delta="10" >
+                <assert_contents>
+                    <has_text text="Visualizations" />
+                </assert_contents>
+            </output>
+        </test>
+    </tests>
+    <help>
+        <![CDATA[
+**What it does**
+Train a model.
+
+
+**Output**
+One trained ludwig_model type composite dataset.
+One html containing the training report.
+
+
+        ]]>
+    </help>
+    <expand macro="macro_citations" />
+</tool>