ose-plugin-hierarchical-spreadsheets 0.2.5__py3-none-any.whl → 0.3.1__py3-none-any.whl

ose_plugin_hierarchical_spreadsheets/GenerateHierarchicalSpreadsheetReleaseStep.py

@@ -2,6 +2,8 @@ import dataclasses
 import re
 from typing import List, Optional, Dict, Callable, Any, Tuple
 
+import networkx as nx
+
 import openpyxl
 import pyhornedowl
 from flask_github import GitHub
@@ -12,8 +14,9 @@ from ose.release.ReleaseStep import ReleaseStep
 from ose.model.ReleaseScript import ReleaseScript, ReleaseScriptFile
 from ose.model.Result import Result
 from ose.services.ConfigurationService import ConfigurationService
-from ose.utils import letters
-from ose.utils.github import parse_spreadsheet
+
+from ose.model.ExcelOntology import ExcelOntology
+from ose.model.TermIdentifier import TermIdentifier
 
 
 @dataclasses.dataclass
@@ -43,7 +46,10 @@ class Node:
 
 def form_tree(edges: List[Tuple[Tuple[str, str | None, str | None], Optional[str]]]) -> List[Node]:
     all_nodes = set(n for n, _ in edges)
-    item_to_node = dict((c, Node(item=c, label=lbl if lbl is not None else c, definition=d or "<no definition>")) for (c, lbl, d) in all_nodes)
+    item_to_node = dict(
+        (c, Node(item=c, label=lbl if lbl is not None else c, definition=d or "<no definition>"))
+        for (c, lbl, d) in all_nodes
+    )
 
     for (child, _, _), parent in edges:
         if parent is None:
@@ -62,9 +68,17 @@ def form_tree(edges: List[Tuple[Tuple[str, str | None, str | None], Optional[str
 
 
 class GenerateHierarchicalSpreadsheetReleaseStep(ReleaseStep):
-
-    def __init__(self, db: SQLAlchemy, gh: GitHub, release_script: ReleaseScript, release_id: int, tmp: str,
-                 config: ConfigurationService, *, included_files: Dict[str, str]) -> None:
+    def __init__(
+        self,
+        db: SQLAlchemy,
+        gh: GitHub,
+        release_script: ReleaseScript,
+        release_id: int,
+        tmp: str,
+        config: ConfigurationService,
+        *,
+        included_files: Dict[str, str],
+    ) -> None:
         super().__init__(db, gh, release_script, release_id, tmp, config)
 
         self._included_files = included_files
@@ -78,29 +92,42 @@ class GenerateHierarchicalSpreadsheetReleaseStep(ReleaseStep):
         for file in files:
             self._next_item(item=file.target.file, message="Generating hierarchical spreadsheet for")
 
-            hierarchies, ontology = self.build_hierarchy(file)
+            G, ontology = self.build_hierarchy(file)
 
             wb = openpyxl.Workbook()
             assert wb.active is not None
             sheet = wb.active
 
-            height = max(h.height() for h in hierarchies)
-            annotations = list({k for h in hierarchies for k in h.annotations.keys()})
+            height = nx.dag_longest_path_length(G) + 1
+            # height = max(h.height() for h in hierarchies)
+            annotations = list(
+                {
+                    k
+                    for n, d in G.nodes(data=True)
+                    for k in d.keys() if k not in {"label", "definition"}
+                }
+            )
+            # annotations = list({k for h in hierarchies for k in h.annotations.keys()})
 
             sheet.append(["ID", "Label"] + [""] * (height - 1) + ["Definition"] + annotations)
 
-            def write_line(n: Node, depth: int) -> None:
-                sheet.append([ontology.get_id_for_iri(n.item)] +
-                             [""] * depth +
-                             [n.label] + [""] * (height - depth - 1) +
-                             [n.definition] +
-                             [n.annotations.get(a, None) for a in annotations])
-
-                for child in n.children:
-                    write_line(child, depth + 1)
-
-            for hierarchy in hierarchies:
-                write_line(hierarchy, 0)
+            def write_line(n: str, d: dict, depth: int) -> None:
+                sheet.append(
+                    [ontology.get_id_for_iri(n)]
+                    + [""] * depth
+                    + [d.get("label", n)]
+                    + [""] * (height - depth - 1)
+                    + [d.get("definition", "")]
+                    + [d.get(a, None) for a in annotations]
+                )
+
+            roots = [(n, d) for n, d in G.nodes(data=True) if G.in_degree(n) == 0]
+            for root, root_data in roots:
+                write_line(root, root_data, 0)
+                successors = nx.dfs_preorder_nodes(G, root)
+                for c in successors:
+                    d = G.nodes[c]
+                    write_line(c, d, nx.shortest_path_length(G, root, c))
 
             [path, name] = file.target.file.rsplit("/", 1)
             sub_name = name.rsplit(".", 1)[0]
@@ -111,33 +138,40 @@ class GenerateHierarchicalSpreadsheetReleaseStep(ReleaseStep):
             wb.save(self._local_name(file_name))
 
             self._store_artifact(self._local_name(file_name), f"{path}/{file_name}")
-
+        
+        
+        result.value = None
         result.warnings = []
         self._set_release_result(result)
         return result.ok()
+        
+        
 
     @classmethod
     def name(cls) -> str:
         return "HIERARCHICAL_SPREADSHEETS"
 
-    def build_hierarchy(self, file: ReleaseScriptFile) -> Tuple[List[Node], pyhornedowl.PyIndexedOntology]:
-        # Excel files to extract annotations
-        excel_files: List[str]
+    def build_hierarchy(self, file: ReleaseScriptFile) -> Tuple[nx.DiGraph, pyhornedowl.PyIndexedOntology]:
         release_file: str
 
-        excel_files = [self._local_name(s.file) for s in file.sources]
-        release_file = next((a.local_path for a in self._artifacts()
-                             if a.target_path == file.target.file and a.kind == 'final'),
-                            self._local_name(file.target.file))
+        release_file = next(
+            (a.local_path for a in self._artifacts() if a.target_path == file.target.file and a.kind == "final"),
+            self._local_name(file.target.file),
+        )
 
         ontology = pyhornedowl.open_ontology_from_file(release_file)
 
         for p, d in self._repo_config.prefixes.items():
             ontology.prefix_mapping.add_prefix(p, d)
 
-        classes = [(c, ontology.get_annotation(c, "http://www.w3.org/2000/01/rdf-schema#label"),
-                    ontology.get_annotation(c, "http://purl.obolibrary.org/obo/IAO_0000115")) for c in
-                   ontology.get_classes()]
+        classes = [
+            (
+                c,
+                ontology.get_annotation(c, "http://www.w3.org/2000/01/rdf-schema#label"),
+                ontology.get_annotation(c, "http://purl.obolibrary.org/obo/IAO_0000115"),
+            )
+            for c in ontology.get_classes()
+        ]
         child_parent: List[Tuple[Tuple[str, Optional[str], Optional[str]], Optional[str]]] = []
         for c in classes:
             for p in ontology.get_superclasses(c[0]):
@@ -145,41 +179,42 @@ class GenerateHierarchicalSpreadsheetReleaseStep(ReleaseStep):
             else:
                 child_parent.append((c, None))
 
-        # child_parent = [(c, p) for c in classes for p in ontology.get_superclasses(c[0])]
-        hierarchies = form_tree(child_parent)
-
-        # If we do not collapse (or do not import) imported ontologies a root will always contain no label
-        # We remove these roots as they only indicate where the subontology should be mounted in the overall ontology
-        hierarchies = [c for h in hierarchies for c in h.children]
-
-        for excel_file in excel_files:
-            with open(excel_file, "rb") as f:
-                file_data = f.read()
-
-            rows, header = parse_spreadsheet(file_data)
-
-            data = dict((r["ID"], r) for r in rows if "ID" in r)
-
-            def annotate(n: Node):
-                id = ontology.get_id_for_iri(n.item)
-                if id is None:
-                    return
+        excel_ontology = ExcelOntology(file.target.iri)
+        for s in file.sources:
+            if s.type == "classes":
+                excel_ontology.add_terms_from_excel(s.file, self._local_name(s.file)).ok_or_raise()
+            elif s.type == "relations":
+                excel_ontology.add_relations_from_excel(s.file, self._local_name(s.file)).ok_or_raise()
+
+        def node_annotations(iri: str) -> dict[str, str | None]:
+            id = ontology.get_id_for_iri(iri)
+            if id is None:
+                return dict()
+
+            term = excel_ontology._raw_term_by_id(id)
+            if term is None or isinstance(term, TermIdentifier):
+                return dict()
+
+            return {
+                "comment": term.get_relation_value(TermIdentifier(id="rdfs:comment")),
+                "subontology": term.get_relation_value(TermIdentifier(label="lowerLevelOntology")),
+                "examples": "; ".join(term.get_relation_values(TermIdentifier(id="IAO:0000112"))),
+                "synonyms": "; ".join(term.get_relation_values(TermIdentifier(id="IAO:0000118"))),
+                "crossreference": "; ".join(term.get_relation_values(TermIdentifier(label="crossReference"))),
+                "informaldefinition": term.get_relation_value(TermIdentifier(label="informalDefinition")),
+            }
                 
-                fields = {
-                    "comment": "Comment",
-                    "subontology": "Sub-ontology",
-                    "examples": "Examples",
-                    "synonyms": "Synonyms",
-                    "crossreference": "Cross reference",
-                    "informaldefinition": "Informal definition",
-                }
-                for field_key, field in fields.items():
-                    node_data = data.get(id, dict())
-                    key = next((k for k in node_data.keys() if letters(k) == field_key), None)
-                    if key is not None and n.annotations.get(field, None) is None:
-                        n.annotations[field] = node_data[key]
-
-            for h in hierarchies:
-                h.recurse(annotate)
-
-        return hierarchies, ontology
+        G = nx.DiGraph()
+        G.add_nodes_from([(iri, {**node_annotations(iri), "label": label, "definition": definition}) for iri, label, definition in classes])
+        G.add_edges_from([(p, c) for (c, _, _), p in child_parent if p is not None])
+        
+        # The only terms that should be included. All other may only be included if they are on a path from root to one of these terms.
+        required_leaf_terms = {ontology.get_iri_for_id(t.id) for t in excel_ontology._terms if t.id is not None}
+        
+        # Remove all other nodes (e.g. external or from a higher level ontology)
+        for root in [n for n in G if G.in_degree(n) == 0]:
+            for n in list(nx.dfs_postorder_nodes(G, root)):
+                if G.out_degree(n) == 0 and n not in required_leaf_terms:
+                    G.remove_node(n)
+        
+        return G, ontology

ose_plugin_hierarchical_spreadsheets/__init__.py

@@ -1,14 +1,20 @@
+from importlib.metadata import version, PackageNotFoundError
+
+try:
+    __version__ = version("ose-plugin-hierarchical-spreadsheets")
+except PackageNotFoundError:
+    __version__ = "unknown"
+
 from ose.model.Plugin import Plugin
 from .GenerateHierarchicalSpreadsheetReleaseStep import GenerateHierarchicalSpreadsheetReleaseStep
 
 
-
 HierarchicalSpreadsheetsPlugin = Plugin(
     id="org.bssofoundry.hierarchicalspreadsheets",
     name="HierarchicalSpreadsheet Plugin",
     version="0.1.0",
     description="Plugin to generate hierarchical spreadsheets during release.",
     contents=[
-      GenerateHierarchicalSpreadsheetReleaseStep,
+        GenerateHierarchicalSpreadsheetReleaseStep,
     ],
-)
+)

{ose_plugin_hierarchical_spreadsheets-0.2.5.dist-info → ose_plugin_hierarchical_spreadsheets-0.3.1.dist-info}/METADATA

@@ -1,10 +1,12 @@
 Metadata-Version: 2.4
 Name: ose-plugin-hierarchical-spreadsheets
-Version: 0.2.5
+Version: 0.3.1
 Summary: OntoSpreadEd plugin for hierarchical spreadsheet generation
 Requires-Python: >=3.12
 Description-Content-Type: text/markdown
-Requires-Dist: ose-core==0.2.5
+Requires-Dist: networkx>=3.6.1
+Requires-Dist: ose-core==0.3.1
+Requires-Dist: scipy>=1.17.0
 
 # OSE Plugin: Hierarchical Spreadsheets
 

ose_plugin_hierarchical_spreadsheets-0.3.1.dist-info/RECORD

@@ -0,0 +1,7 @@
+ose_plugin_hierarchical_spreadsheets/GenerateHierarchicalSpreadsheetReleaseStep.py,sha256=yEjYsJD0xNnBK03la0NAwuUUn8GhFdflzmPCxUW9-WI,8318
+ose_plugin_hierarchical_spreadsheets/__init__.py,sha256=SJfQCWp7Jzg6N3-VZ5JDCUOsjKEZ5yPabevF2S1Me2Q,640
+ose_plugin_hierarchical_spreadsheets-0.3.1.dist-info/METADATA,sha256=oGNDPkARV5e4ZraiB15ET4CnzKpyJFLMWnSQ2xGiNEs,1008
+ose_plugin_hierarchical_spreadsheets-0.3.1.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+ose_plugin_hierarchical_spreadsheets-0.3.1.dist-info/entry_points.txt,sha256=a6pod_osiwl4-QPdRnl2wk4HQpNGYS_7jQ6_C6aJgrc,110
+ose_plugin_hierarchical_spreadsheets-0.3.1.dist-info/top_level.txt,sha256=aYtBQC3URKw1ZKCmEvBWYGsITjMnjmbKIrkevgzQcI4,37
+ose_plugin_hierarchical_spreadsheets-0.3.1.dist-info/RECORD,,

{ose_plugin_hierarchical_spreadsheets-0.2.5.dist-info → ose_plugin_hierarchical_spreadsheets-0.3.1.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (80.10.1)
+Generator: setuptools (80.10.2)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

ose_plugin_hierarchical_spreadsheets-0.2.5.dist-info/RECORD

@@ -1,7 +0,0 @@
-ose_plugin_hierarchical_spreadsheets/GenerateHierarchicalSpreadsheetReleaseStep.py,sha256=5HFAlpDKykhtH6WeAr9asasqHfjwtVfxS_3QrkeDWrA,7158
-ose_plugin_hierarchical_spreadsheets/__init__.py,sha256=4X14epIVlHO_3MY9uA1qK5LWXLqDaMqX2lEp3dywUp4,447
-ose_plugin_hierarchical_spreadsheets-0.2.5.dist-info/METADATA,sha256=MI2CTDzDdW9tpQSH1xQ0Kcbr5GJ7hOs0KeF_D4sRIFE,948
-ose_plugin_hierarchical_spreadsheets-0.2.5.dist-info/WHEEL,sha256=qELbo2s1Yzl39ZmrAibXA2jjPLUYfnVhUNTlyF1rq0Y,92
-ose_plugin_hierarchical_spreadsheets-0.2.5.dist-info/entry_points.txt,sha256=a6pod_osiwl4-QPdRnl2wk4HQpNGYS_7jQ6_C6aJgrc,110
-ose_plugin_hierarchical_spreadsheets-0.2.5.dist-info/top_level.txt,sha256=aYtBQC3URKw1ZKCmEvBWYGsITjMnjmbKIrkevgzQcI4,37
-ose_plugin_hierarchical_spreadsheets-0.2.5.dist-info/RECORD,,