Request up to 5 principal components in pca process

docs/CHANGELOG.rst

@@ -16,6 +16,7 @@ Added
 
 Changed
 -------
+- Calculate up to 5 principal components in ``pca`` process
 
 Fixed
 -----

resolwe_bio/processes/clustering/pca.py

@@ -30,7 +30,7 @@ def component_top_factors(component, allgenes_array, max_size=20):
     return list(zip(np.array(allgenes_array)[ixs].tolist(), component[ixs].tolist()))
 
 
-def get_pca(expressions=pd.DataFrame(), gene_labels=[]):
+def get_pca(expressions=pd.DataFrame(), gene_labels=[], n_components=5):
     """Compute PCA."""
     if not gene_labels:
         gene_labels = expressions.index
@@ -41,11 +41,13 @@ def get_pca(expressions=pd.DataFrame(), gene_labels=[]):
         all_components = [[], []]
         all_explained_variance_ratios = [0.0, 0.0]
     else:
-        pca = PCA(n_components=2, whiten=True)
+        pca_components = min(expressions.shape[0], expressions.shape[1], n_components)
+        pca = PCA(n_components=pca_components, whiten=True)
         pca_expressions = pca.fit_transform(expressions.transpose())
 
         coordinates = [
-            t[:2].tolist() if len(t) > 1 else [t[0], 0.0] for t in pca_expressions
+            t[:pca_components].tolist() if len(t) > 1 else [t[0], 0.0]
+            for t in pca_expressions
         ]
         all_components = [
             component_top_factors(component=component, allgenes_array=gene_labels)
@@ -197,7 +199,7 @@ class PrinicipalComponentAnalysis(Process):
         },
     }
     data_name = "PCA"
-    version = "3.0.0"
+    version = "3.1.0"
     process_type = "data:pca"
     category = "Enrichment and Clustering"
     scheduling_class = SchedulingClass.INTERACTIVE

resolwe_bio/tests/processes/test_pca.py

@@ -53,17 +53,62 @@ def test_pca(self):
                 source="DICTYBASE",
                 species="Dictyostelium discoideum",
             )
+            expression_5 = self.prepare_expression(
+                f_rc="exp_6_rc.tab.gz",
+                f_exp="exp_6_tpm.tab.gz",
+                f_type="TPM",
+                source="DICTYBASE",
+                species="Dictyostelium discoideum",
+            )
+            expression_6 = self.prepare_expression(
+                f_rc="exp_7_rc.tab.gz",
+                f_exp="exp_7_tpm.tab.gz",
+                f_type="TPM",
+                source="DICTYBASE",
+                species="Dictyostelium discoideum",
+            )
+            expression_7 = self.prepare_expression(
+                f_rc="exp_8_rc.tab.gz",
+                f_exp="exp_8_tpm.tab.gz",
+                f_type="TPM",
+                source="DICTYBASE",
+                species="Dictyostelium discoideum",
+            )
 
         inputs = {
-            "exps": [expression_1.pk, expression_2.pk],
+            "exps": [
+                expression_1.pk,
+                expression_2.pk,
+                expression_5.pk,
+                expression_6.pk,
+                expression_7.pk,
+            ],
             "source": "DICTYBASE",
             "species": "Dictyostelium discoideum",
         }
         pca = self.run_process("pca", inputs)
         saved_json, test_json = self.get_json("pca_plot.json.gz", pca.output["pca"])
+        # returns 4 PCA components. Last component differs when testing on different systems
+        # and is not tested here.
         self.assertAlmostEqualGeneric(
-            round_elements(test_json["flot"]["data"]),
-            round_elements(saved_json["flot"]["data"]),
+            round_elements(test_json["flot"]["data"][0][:3]),
+            round_elements(saved_json["flot"]["data"][0][:3]),
+        )
+        self.assertAlmostEqualGeneric(
+            round_elements(test_json["flot"]["data"][1][:3]),
+            round_elements(saved_json["flot"]["data"][1][:3]),
+        )
+        self.assertAlmostEqualGeneric(
+            round_elements(test_json["flot"]["data"][2][:3]),
+            round_elements(saved_json["flot"]["data"][2][:3]),
+        )
+        self.assertAlmostEqualGeneric(
+            round_elements(test_json["flot"]["data"][3][:3]),
+            round_elements(saved_json["flot"]["data"][3][:3]),
+        )
+        self.assertAlmostEqualGeneric(
+            round_elements(test_json["flot"]["data"][4][:3]),
+            round_elements(saved_json["flot"]["data"][4][:3]),
         )
         self.assertAlmostEqualGeneric(
             round_elements(test_json["explained_variance_ratios"]),
@@ -84,8 +129,8 @@ def test_pca(self):
         pca = self.run_process("pca", inputs)
         saved_json, test_json = self.get_json("pca_plot_2.json.gz", pca.output["pca"])
         self.assertAlmostEqualGeneric(
-            round_elements(test_json["flot"]["data"]),
-            round_elements(saved_json["flot"]["data"]),
+            round_elements(test_json["flot"]["data"][0][:3]),
+            round_elements(saved_json["flot"]["data"][0][:3]),
         )
 
         self.assertEqual(len(pca.process_warning), 0)
@@ -95,8 +140,8 @@ def test_pca(self):
         pca = self.run_process("pca", inputs)
         saved_json, test_json = self.get_json("pca_plot_3.json.gz", pca.output["pca"])
         self.assertAlmostEqualGeneric(
-            round_elements(test_json["flot"]["data"]),
-            round_elements(saved_json["flot"]["data"]),
+            round_elements(test_json["flot"]["data"][0][:3]),
+            round_elements(saved_json["flot"]["data"][0][:3]),
         )
 
         self.assertEqual(len(pca.process_warning), 0)