From af76b58246f87c439d04c28713589f68e2bcd0d4 Mon Sep 17 00:00:00 2001
From: Charles Tapley Hoyt <cthoyt@gmail.com>
Date: Wed, 5 Jul 2023 12:46:59 +0200
Subject: [PATCH] Additional updates

---
 mira/dkg/construct.py          | 17 +++++++++++++----
 mira/dkg/construct_registry.py |  1 -
 mira/dkg/resources/probonto.py | 10 ++++++++--
 3 files changed, 21 insertions(+), 7 deletions(-)

diff --git a/mira/dkg/construct.py b/mira/dkg/construct.py
index 631e25415..e1a62dd0f 100644
--- a/mira/dkg/construct.py
+++ b/mira/dkg/construct.py
@@ -464,7 +464,7 @@ def construct(
                 property_values="",
                 xref_types="",  # TODO
                 synonym_types=";".join(
-                    synonym.type or "skos:exactMatch" for synonym in term.synonyms or []
+                    synonym.type.curie for synonym in term.synonyms or []
                 ),
             )
             for parent in term.parents:
@@ -781,6 +781,16 @@ def _get_edge_name(curie_: str, strict: bool = False) -> str:
                 edge.pred for edge in graph.edges if edge.pred.startswith("http")
             )
 
+            clean_edges = (
+                edge
+                for edge in graph.edges
+                if (
+                    edge.subject is not None
+                    and edge.predicate is not None
+                    and edge.object is not None
+                    and edge.object.curie not in OBSOLETE
+                )
+            )
             edges.extend(
                 (
                     edge.sub,
@@ -792,9 +802,8 @@ def _get_edge_name(curie_: str, strict: bool = False) -> str:
                     version or "",
                 )
                 for edge in tqdm(
-                    sorted(graph.edges, key=methodcaller("as_tuple")), unit="edge", unit_scale=True
+                    sorted(clean_edges, key=methodcaller("as_tuple")), unit="edge", unit_scale=True
                 )
-                if edge.obj not in OBSOLETE
             )
 
         for sub, obj, pred_label, pred, *_ in edges:
@@ -970,7 +979,7 @@ def get_node_info(term: pyobo.Term, type: EntityType = "class"):
         property_values="",
         xref_types="",
         synonym_types=";".join(
-            synonym.type or "skos:exactMatch" for synonym in term.synonyms or []
+            synonym.type.curie for synonym in term.synonyms or []
         ),
     )
 
diff --git a/mira/dkg/construct_registry.py b/mira/dkg/construct_registry.py
index 1bfe368da..bde7c12bd 100644
--- a/mira/dkg/construct_registry.py
+++ b/mira/dkg/construct_registry.py
@@ -9,7 +9,6 @@
 from typing import Optional, Set
 
 import bioregistry
-import bioregistry.app.impl
 import click
 from bioregistry import Manager
 from tqdm import tqdm
diff --git a/mira/dkg/resources/probonto.py b/mira/dkg/resources/probonto.py
index 187075340..3bc71dafb 100644
--- a/mira/dkg/resources/probonto.py
+++ b/mira/dkg/resources/probonto.py
@@ -1,9 +1,12 @@
 import json
 from collections import defaultdict
+from pathlib import Path
 
 import bioontologies
 import rdflib
 
+HERE = Path(__file__).parent.resolve()
+PROBONTO_PATH = HERE.joinpath("probonto.json")
 
 def get_data_properties(rdf_graph: rdflib.Graph, identifier):
     return {
@@ -24,7 +27,10 @@ def get_instances(obo_graph, probonto_identifier: str):
     ]
 
 
-def get_probonto_terms():
+def get_probonto_terms(*, refresh: bool = False):
+    if PROBONTO_PATH.is_file() and not refresh:
+        return json.loads(PROBONTO_PATH.read_text())
+
     obo_graph = (
         bioontologies.get_obograph_by_prefix("probonto")
         .guess("probonto")
@@ -96,7 +102,7 @@ def get_probonto_terms():
 
 def main():
     results = get_probonto_terms()
-    with open("probonto.json", "w") as file:
+    with open(PROBONTO_PATH, "w") as file:
         json.dump(results, file, indent=2, ensure_ascii=False)