Trying to optimize the queries

lotusnprod · Nov 20, 2023 · 291e3f0 · 291e3f0
1 parent 7ebca25
commit 291e3f0
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 10 deletions.
diff --git a/update/download_couples_referenced.py b/update/download_couples_referenced.py
@@ -7,10 +7,10 @@
 
 def run(root: Path, retry: int = 3) -> None:
     query = """
-    SELECT DISTINCT ?compound ?taxon ?reference WHERE {
-      ?compound p:P703 ?statement;
-        # We use InChIKey (P235) instead of SMILES as some of them are incomplete.
-        wdt:P235 [].
+    SELECT DISTINCT ?compound ?taxon ?reference WHERE {    
+      # We use InChIKey (P235) instead of SMILES as some of them are incomplete.
+      ?compound wdt:P235 [].
+      ?compound p:P703 ?statement.
       ?statement ps:P703 ?taxon;
         (prov:wasDerivedFrom/pr:P248) ?reference.
     }

diff --git a/update/download_smiles.py b/update/download_smiles.py
@@ -8,12 +8,13 @@
 def run(root: Path, retry: int = 3) -> None:
     query = """
     SELECT DISTINCT ?structure ?structure_smiles ?canonical_smiles WHERE {
-        ?structure wdt:P703 ?taxon;
-          # Using InChIKey (P235) to recognize chemicals. 
-          # Could also be
-          # P31 wd:Q113145171 `type of a chemical entity`
-          # P31 wd:Q59199015 `group of stereoisomers`
-          wdt:P235 [].
+        # Using InChIKey (P235) to recognize chemicals. 
+        # Could also be
+        # P31 wd:Q113145171 `type of a chemical entity`
+        # P31 wd:Q59199015 `group of stereoisomers`
+        ?structure wdt:P235 [].
+        ?structure wdt:P703 ?taxon.
+          
         # All P2017 should also have P233 but some of them are not complete.
         OPTIONAL {
             ?structure wdt:P233 ?canonical_smiles.