add biii parsing and allow for multiple xref parsing

bin/create_interactive_table.py

@@ -4,7 +4,7 @@
 
 import pandas as pd
 
-# TODO maybe allow comunities to modify
+# TODO maybe allow communities to modify
 COLUMNS = [
     "Expand",
     "Galaxy wrapper id",
@@ -18,6 +18,7 @@
     "EDAM topic",
     "Description",
     "bio.tool description",
+    "biii",
     "Status",
     "Source",
     "ToolShed categories",

bin/extract_galaxy_tools.py

@@ -187,11 +187,13 @@ def get_xref(el: et.Element, attrib_type: str) -> Optional[str]:
     :param el: Element object
     :attrib_type: the type of the xref (e.g.: bio.tools or biii)
     """
+
     xrefs = el.find("xrefs")
     if xrefs is not None:
-        xref = xrefs.find("xref")
-        if xref is not None and xref.attrib["type"] == attrib_type:
-            return xref.text
+        xref_items = xrefs.findall("xref")  # check all xref items
+        for xref in xref_items:
+            if xref is not None and xref.attrib["type"] == attrib_type:
+                return xref.text
     return None
 
 
@@ -298,9 +300,14 @@ def get_tool_metadata(tool: ContentFile, repo: Repository) -> Optional[Dict[str,
                         metadata["Galaxy wrapper version"] = child.text
                     elif child.attrib["name"] == "requirements":
                         metadata["Conda id"] = get_conda_package(child)
+                    # bio.tools
                     biotools = get_xref(child, attrib_type="bio.tools")
                     if biotools is not None:
                         metadata["bio.tool id"] = biotools
+                    # biii
+                    biii = get_xref(child, attrib_type="biii")
+                    if biii is not None:
+                        metadata["biii"] = biii
     # parse XML file and get meta data from there, also tool ids
     for file in file_list:
         if file.name.endswith("xml") and "macro" not in file.name:
@@ -329,6 +336,12 @@ def get_tool_metadata(tool: ContentFile, repo: Repository) -> Optional[Dict[str,
                     biotools = get_xref(root, attrib_type="bio.tools")
                     if biotools is not None:
                         metadata["bio.tool id"] = biotools
+                        # bio.tools
+                # biii
+                if metadata["biii"] is None:
+                    biii = get_xref(root, attrib_type="biii")
+                    if biii is not None:
+                        metadata["biii"] = biii
                 # conda package
                 if metadata["Conda id"] is None:
                     reqs = get_conda_package(root)