From 631277510fc820047fb07a268461aa33d1594225 Mon Sep 17 00:00:00 2001
From: mboudet <mateo.boudet@gmail.com>
Date: Thu, 10 Aug 2023 17:19:22 +0200
Subject: [PATCH] Move default query back to faldo ontology

---
 CHANGELOG.md                        |  1 +
 askomics/libaskomics/SparqlQuery.py |  6 +++---
 docs/abstraction.md                 | 19 +++++++++++++++++--
 3 files changed, 21 insertions(+), 5 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 8e34d082..f36a7ae7 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -52,6 +52,7 @@ This changelog was started for release 4.2.0.
 - 'Same strand' queries will now match 'BothStrand' with a forward or reverse strand
 - Use '+', '-' and '.' for strand values in CSV instead of raw value (for homogenization)
 - Now allows 'infinite' recursive blocks (ie, a Minus block inside a Union block, or the opposite)
+- Overhaul faldo relation: add 'direct shortcut' triples between entities and values, to avoid using slow property paths when using faldo queries
 
 ### Security
 
diff --git a/askomics/libaskomics/SparqlQuery.py b/askomics/libaskomics/SparqlQuery.py
index 222f9128..a2ee56d4 100644
--- a/askomics/libaskomics/SparqlQuery.py
+++ b/askomics/libaskomics/SparqlQuery.py
@@ -1538,7 +1538,7 @@ def build_query_from_json(self, preview=False, for_editor=False):
                 if attribute["visible"] or Utils.check_key_in_list_of_dict(attribute["filters"], "filterValue") or attribute["id"] in start_end or attribute["id"] in linked_attributes:
                     subject = self.format_sparql_variable("{}{}_uri".format(attribute["entityLabel"], attribute["nodeId"]))
                     if attribute["faldo"]:
-                        predicate = "askomics:{}".format("faldoBegin" if attribute["faldo"].endswith("faldoStart") else "faldoEnd")
+                        predicate = "faldo:location/faldo:{}/faldo:position".format("begin" if attribute["faldo"].endswith("faldoStart") else "end")
                     else:
                         predicate = "<{}>".format(attribute["uri"])
                     obj = self.format_sparql_variable("{}{}_{}".format(attribute["entityLabel"], attribute["nodeId"], attribute["label"]))
@@ -1623,7 +1623,7 @@ def build_query_from_json(self, preview=False, for_editor=False):
                     category_value_uri = self.format_sparql_variable("{}{}_{}Category".format(attribute["entityLabel"], attribute["nodeId"], attribute["label"]))
                     category_label = self.format_sparql_variable("{}{}_{}".format(attribute["entityLabel"], attribute["humanNodeId"], attribute["label"]))
                     if attribute["faldo"] and attribute["faldo"].endswith("faldoReference"):
-                        category_name = 'askomics:faldoReference'
+                        category_name = 'faldo:location/faldo:begin/faldo:reference'
                         self.store_triple({
                             "subject": node_uri,
                             "predicate": category_name,
@@ -1640,7 +1640,7 @@ def build_query_from_json(self, preview=False, for_editor=False):
                                 "nested_end": True if attribute["optional"] else False
                             }, block_id, sblock_id, pblock_ids, depth)
                     elif attribute["faldo"] and attribute["faldo"].endswith("faldoStrand"):
-                        category_name = 'askomics:faldoStrand'
+                        category_name = 'faldo:location/faldo:begin/rdf:type'
                         self.store_triple({
                             "subject": node_uri,
                             "predicate": category_name,
diff --git a/docs/abstraction.md b/docs/abstraction.md
index d4f782b5..4bdb23ca 100644
--- a/docs/abstraction.md
+++ b/docs/abstraction.md
@@ -111,6 +111,19 @@ A FALDO entity have to be declared as FALDO on the abstraction. If attribute are
 
 Four FALDO attributes are supported by AskOmics: reference, strand, start and end.
 
+!!! warning
+    AskOmics expect faldo entities to follow the faldo ontology for triple definition. Ex:
+    ```turtle
+    # Reference
+    :Entity faldo:location/faldo:begin/faldo:reference "value"
+    # strand
+    :Entity faldo:location/faldo:begin/rdf:type "value"
+    # Start
+    :Entity faldo:location/faldo:begin/faldo:position "value"
+    # Stop.
+    :Entity faldo:location/faldo:end/faldo:position "value"
+    ```
+
 ### faldo:reference
 
 A faldo:reference attribute derive from a Category attribute.
@@ -190,8 +203,10 @@ The default faldo ontology uses a chain of triple to describe the position (ex,
 This make *faldo queries* (included_in/overlap_with/distant_from) extremely slow. To improve query time, AskOmics can use 'shortcut triples', direct relations between the Entity and the reference/strand, to quickly filter entities on the same reference/strand/both. For example:
 
 ```turtle
-:EntityName faldo:reference reference_uri .
-:EntityName faldo:strand strand_uri .
+:EntityName askomics:faldoReference reference_uri .
+:EntityName askomics:faldoBegin begin_value .
+:EntityName askomics:faldoEnd end_value .
+:EntityName askomics:faldoStrand strand_uri .
 :EntityName askomics:referenceStrand reference_strand_uri .
 ```