Merge pull request #20466 from pvanheus/add_auspicejson_datatype

[25.0] Add Auspice JSON datatype

Author: mvdbeek

lib/galaxy/config/sample/datatypes_conf.xml.sample

@@ -148,6 +148,7 @@
     <datatype extension="vitessce.json" type="galaxy.datatypes.text:VitessceJson" mimetype="application/json" display_in_upload="True">
       <visualization plugin="vitessce" />
     </datatype>
+    <datatype extension="auspice.json" type="galaxy.datatypes.text:AuspiceJson" mimetype="application/json" display_in_upload="True" />
     <datatype extension="data_manager_json" type="galaxy.datatypes.text:DataManagerJson" mimetype="application/json" subclass="true" display_in_upload="false"/>
     <datatype extension="dbn" type="galaxy.datatypes.sequence:DotBracket" display_in_upload="true" description="Dot-Bracket format is a text-based format for storing both an RNA sequence and its corresponding 2D structure." description_url="https://wiki.galaxyproject.org/Learn/Datatypes#Dbn"/>
     <datatype extension="fai" type="galaxy.datatypes.tabular:Tabular" display_in_upload="true" subclass="true" description="A Fasta Index File is a text file consisting of lines each with five TAB-delimited columns : Name, Length, offset, linebases, Linewidth" description_url="http://www.htslib.org/doc/faidx.html"/>
@@ -1387,6 +1388,7 @@
     <sniffer type="galaxy.datatypes.text:CytoscapeJson"/>
     <sniffer type="galaxy.datatypes.text:GeoJson"/>
     <sniffer type="galaxy.datatypes.text:VitessceJson"/>
+    <snipper type="galaxy.datatypes.text:AuspiceJson"/>
     <sniffer type="galaxy.datatypes.text:PithyaResult"/>
     <sniffer type="galaxy.datatypes.text:BCSLts"/>
     <sniffer type="galaxy.datatypes.text:Json"/>

lib/galaxy/datatypes/test/1.auspicejson

@@ -0,0 +1,14 @@
+{
+  "version": "v2",
+  "meta": {
+    "title": "Minimal AuspiceJSON",
+    "updated": "2025-02-05",
+    "panels": ["tree"]
+  },
+  "tree": {
+    "name": "1",
+    "node_attrs": {
+      "div": 1
+    }
+  }
+}

lib/galaxy/datatypes/text.py

@@ -695,6 +695,57 @@ def _looks_like_is_vitesscejson(self, file_prefix: FilePrefix, load_size: int =
         return False
 
 
+@build_sniff_from_prefix
+class AuspiceJson(Json):
+    """
+    Auspice is a visualization tool for phylogenetic trees and associated data.
+    It uses JSON format to represent the tree structure and metadata.
+    """
+
+    file_ext = "auspice.json"
+
+    def set_peek(self, dataset: DatasetProtocol, **kwd) -> None:
+        super().set_peek(dataset)
+        if not dataset.dataset.purged:
+            dataset.blurb = "AuspiceJSON"
+
+    def sniff_prefix(self, file_prefix: FilePrefix) -> bool:
+        """
+        Determines whether the file is in Auspice v2 JSON by looking for keys
+        like "version", "meta" and "updated" that are both required by the
+        https://docs.nextstrain.org/projects/auspice/en/stable/releases/v2.html format
+        and also will be in the first part of the file
+
+        >>> from galaxy.datatypes.sniff import get_test_fname
+        >>> fname = get_test_fname( '1.json' )
+        >>> AuspiceJson().sniff( fname )
+        False
+        >>> fname = get_test_fname( '1.auspicejson' )
+        >>> AuspiceJson().sniff( fname )
+        True
+        """
+        is_auspicejson = False
+        if self._looks_like_json(file_prefix):
+            is_auspicejson = self._looks_like_is_auspicejson(file_prefix)
+        return is_auspicejson
+
+    def _looks_like_is_auspicejson(self, file_prefix: FilePrefix, load_size: int = 20000) -> bool:
+        """
+        Expects JSON to start with { and 'meta', 'tree', 'updated' and 'nodes' to be present as keys in the JSON structure.
+        """
+        try:
+            with open(file_prefix.filename) as fh:
+                segment_str = fh.read(load_size)
+
+                if segment_str.startswith("{") and all(
+                    x in segment_str for x in ["version", "meta", "updated", "panels"]
+                ):
+                    return True
+        except Exception:
+            pass
+        return False
+
+
 @build_sniff_from_prefix
 class Obo(Text):
     """