diff --git a/src/main/java/eu/hbp/mip/model/Experiment.java b/src/main/java/eu/hbp/mip/model/Experiment.java
index c42fb6eb88381396d1e640a872e757d3048010d8..727f26ddb35b5a1c9890b5b4ac12a00e40980657 100644
--- a/src/main/java/eu/hbp/mip/model/Experiment.java
+++ b/src/main/java/eu/hbp/mip/model/Experiment.java
@@ -29,8 +29,10 @@ import scala.Tuple2;
public class Experiment {
public static final String WP_K_MEANS = "K_MEANS";
-
public static final String WP_LINEAR_REGRESSION = "WP_LINEAR_REGRESSION";
+ public static final String WP_VARIABLES_HISTOGRAM = "WP_VARIABLES_HISTOGRAM";
+ public static final String WP_REGRESSION_TREE = "PIPELINE_ISOUP_REGRESSION_TREE_SERIALIZER";
+ public static final String WP_MODEL_TREE = "PIPELINE_ISOUP_MODEL_TREE_SERIALIZER";
private static final Logger LOGGER = LoggerFactory.getLogger(Experiment.class);
@@ -134,12 +136,27 @@ public class Experiment {
validationsSeq, Option.empty());
}
+ private String chainsParams(List<String> params, String operator) {
+ StringBuilder sb = new StringBuilder();
+ int i = 0;
+ for (String s : params)
+ {
+ i++;
+ sb.append(s);
+ if (i < params.size()) {
+ sb.append(operator);
+ }
+ }
+ return sb.toString();
+ }
public String computeExaremeQuery(List<AlgorithmParam> params) {
List<ExaremeQueryElement> queryElements = new LinkedList<>();
// parameters
- String design = "";
+ String design = ""; // TODO: don't assign to a global
+ String nobuckets = "";
+
if (params != null) {
for (AlgorithmParam p : params)
{
@@ -152,6 +169,10 @@ public class Experiment {
if (p.getCode().equals("design")) {
design = p.getValue();
}
+
+ if (p.getCode().equals("nobuckets")) {
+ nobuckets = p.getValue();
+ }
}
}
@@ -167,56 +188,69 @@ public class Experiment {
String algoName = this.isExaremeAlgorithm()._2;
if (algoName.equals(WP_K_MEANS)) {
// columns
- List<String> columns = new ArrayList<>();
- columns.addAll(variables);
+ List<String> columns = new ArrayList<>(variables);
columns.addAll(covariables);
columns.addAll(groupings);
- StringBuilder sb = new StringBuilder();
- int i = 0;
- for (String s : columns)
- {
- i++;
- sb.append(s);
- if (i < columns.size()) {
- sb.append(",");
- }
- }
+
ExaremeQueryElement columnsEl = new ExaremeQueryElement();
columnsEl.setName("columns");
columnsEl.setDesc("");
- columnsEl.setValue(sb.toString());
+ columnsEl.setValue(chainsParams(columns, ";"));
queryElements.add(columnsEl);
} else if (algoName.equals(WP_LINEAR_REGRESSION)) {
- List<String> vars = new ArrayList<>();
- vars.addAll(variables);
- vars.addAll(covariables);
+ List<String> nominals = new ArrayList<>(covariables);
+ nominals.addAll(groupings);
String operator = design.equals("factorial") ? "*" : "+";
- StringBuilder sb = new StringBuilder();
- int i = 0;
- for (String s : vars)
- {
- i++;
- sb.append(s);
- if (i < vars.size()) {
- sb.append(operator);
- }
- }
+
ExaremeQueryElement xEl = new ExaremeQueryElement();
xEl.setName("x");
xEl.setDesc("");
- xEl.setValue(sb.toString());
+ xEl.setValue(chainsParams(nominals, operator));
queryElements.add(xEl);
- StringBuilder sby = new StringBuilder();
- for (String s : variables)
- {
- sby.append(s);
- }
ExaremeQueryElement yEl = new ExaremeQueryElement();
yEl.setName("y");
yEl.setDesc("");
- yEl.setValue(sby.toString());
+ yEl.setValue(chainsParams(variables, ";"));
+ queryElements.add(yEl);
+ } else if (algoName.equals(WP_VARIABLES_HISTOGRAM)) {
+ List<String> column1 = new ArrayList<>(variables);
+ ExaremeQueryElement columnsEl = new ExaremeQueryElement();
+ columnsEl.setName("column1");
+ columnsEl.setDesc("");
+ columnsEl.setValue(chainsParams(column1, ","));
+ queryElements.add(columnsEl);
+
+ List<String> column2 = new ArrayList<>(covariables);
+ ExaremeQueryElement columnsEl2 = new ExaremeQueryElement();
+ columnsEl2.setName("column2");
+ columnsEl2.setDesc("");
+ columnsEl2.setValue(chainsParams(column2, ","));
+ queryElements.add(columnsEl2);
+
+ if (!nobuckets.isEmpty()) {
+ ExaremeQueryElement columnsEl3 = new ExaremeQueryElement();
+ columnsEl3.setName("nobuckets");
+ columnsEl3.setDesc("");
+ columnsEl3.setValue(nobuckets);
+ queryElements.add(columnsEl3);
+ }
+ } else if (algoName.equals(WP_REGRESSION_TREE) || algoName.equals(WP_MODEL_TREE)) {
+ List<String> target = new ArrayList<>(variables);
+ List<String> descriptive = new ArrayList<>(covariables);
+ descriptive.addAll(groupings);
+
+ ExaremeQueryElement xEl = new ExaremeQueryElement();
+ xEl.setName("target_attributes");
+ xEl.setDesc("");
+ xEl.setValue(chainsParams(target, ","));
+ queryElements.add(xEl);
+
+ ExaremeQueryElement yEl = new ExaremeQueryElement();
+ yEl.setName("descriptive_attributes");
+ yEl.setDesc("");
+ yEl.setValue(chainsParams(descriptive, ","));
queryElements.add(yEl);
}
@@ -280,12 +314,15 @@ public class Experiment {
jsonObjectResult.add("code", algoObject.get("code"));
// add mime-type
- if (jsonData.get("Error") != null){
+ String algo = isExaremeAlgorithm._2;
+ if (jsonData.get("Error") != null) {
jsonObjectResult.add("type", new JsonPrimitive("text/plain+error"));
- } else if (isExaremeAlgorithm._2 == WP_K_MEANS){
+ } else if (algo.equals( WP_K_MEANS) || algo.equals( WP_VARIABLES_HISTOGRAM)) {
jsonObjectResult.add("type", new JsonPrimitive("application/vnd.highcharts+json"));
- } else if (isExaremeAlgorithm._2 == WP_LINEAR_REGRESSION){
+ } else if (algo.equals(WP_LINEAR_REGRESSION)) {
jsonObjectResult.add("type", new JsonPrimitive("application/vnd.dataresource+json"));
+ } else if (algo.equals(WP_REGRESSION_TREE) || algo.equals(WP_MODEL_TREE)) {
+ jsonObjectResult.add("type", new JsonPrimitive("application/vnd.visjs+javascript"));
}
jsonArrayResult.add(jsonObjectResult);
@@ -406,12 +443,22 @@ public class Experiment {
String algorithm = "";
String algorithms = this.algorithms;
+
if (algorithms.contains(WP_K_MEANS)) {
isExareme = true;
algorithm = WP_K_MEANS;
} else if (algorithms.contains(WP_LINEAR_REGRESSION)) {
isExareme = true;
algorithm = WP_LINEAR_REGRESSION;
+ } else if (algorithms.contains(WP_VARIABLES_HISTOGRAM)) {
+ isExareme = true;
+ algorithm = WP_VARIABLES_HISTOGRAM;
+ } else if (algorithms.contains(WP_REGRESSION_TREE)) {
+ isExareme = true;
+ algorithm = WP_REGRESSION_TREE;
+ } else if (algorithms.contains(WP_MODEL_TREE)) {
+ isExareme = true;
+ algorithm = WP_MODEL_TREE;
}
return new Tuple2(isExareme, algorithm);
diff --git a/src/main/resources/data/exareme_algorithms.json b/src/main/resources/data/exareme_algorithms.json
index 4f05bb102fe2c5305ddd820f0cf9a10e4b3ef20d..f1601397d77487b7623c9e2661d188feb3c8a6cc 100644
--- a/src/main/resources/data/exareme_algorithms.json
+++ b/src/main/resources/data/exareme_algorithms.json
@@ -1,7 +1,66 @@
[
+ {
+ "code": "WP_VARIABLES_HISTOGRAM",
+ "label": " Histogram",
+ "type": ["statistics"],
+ "environment": "Exareme",
+ "description": "Histograms",
+ "parameters": [{
+ "code": "nobuckets",
+ "label": "nobuckets",
+ "default_value": 4,
+ "type": "int",
+ "constraints": {
+ "min": 1,
+ "max": null
+ },
+ "description":
+ "The number of clusters. Typical values range from 2 to 10."
+ }],
+ "constraints": {
+ "variable": {
+ "binominal": true,
+ "integer": true,
+ "polynominal": true,
+ "real": true
+ }
+ }
+ },
+ {
+ "code": "PIPELINE_ISOUP_REGRESSION_TREE_SERIALIZER",
+ "label": " Regression Tree",
+ "type": ["statistics"],
+ "environment": "Exareme",
+ "description": "Regression Tree Serializer",
+ "parameters": [],
+ "constraints": {
+ "variable": {
+ "binominal": true,
+ "integer": true,
+ "polynominal": true,
+ "real": true
+ }
+ }
+ },
+ {
+ "code": "PIPELINE_ISOUP_MODEL_TREE_SERIALIZER",
+ "label": "Model Tree",
+ "type": ["statistics"],
+ "environment": "Exareme",
+ "description": "Model Tree Serializer",
+ "parameters": [],
+ "constraints": {
+ "variable": {
+ "binominal": true,
+ "integer": true,
+ "polynominal": true,
+ "real": true
+ }
+ }
+ },
{
"code": "WP_LINEAR_REGRESSION",
- "label": "Linear regression (Exareme)",
+ "label": "Linear regression",
"type": ["statistics"],
"environment": "Exareme",
"description": "Linear Regression using Exareme services",
@@ -9,13 +68,11 @@
{
"code": "design",
"default_value": "factorial",
- "description": "The type of multi-factors design. Choose 'factorial' to enable interactions analysis or 'additive' for a model without no interaction at all.",
+ "description":
+ "The type of multi-factors design. Choose 'factorial' to enable interactions analysis or 'additive' for a model without no interaction at all.",
"label": "design",
"type": "enumeration",
- "values": [
- "factorial",
- "additive"
- ]
+ "values": ["factorial", "additive"]
}
],
"constraints": {
@@ -29,7 +86,7 @@
},
{
"code": "K_MEANS",
- "label": "K_MEANS (Exareme)",
+ "label": "K_MEANS",
"type": ["statistics"],
"environment": "Exareme",
"description": "k-Means using Exareme services",