feat(taxref): add taxref_tree materialized view (#567)

* feat(taxref): add parent relationship
* feat(taxref): add taxref_tree materialized view
* feat(taxref): taxons comparison through path

.github/workflows/pytest.yml

@@ -61,6 +61,7 @@ jobs:
         psql -h localhost -U taxhubadmin -d taxhub -tc 'CREATE EXTENSION "unaccent";'
         psql -h localhost -U taxhubadmin -d taxhub -tc 'CREATE EXTENSION "uuid-ossp";'
         psql -h localhost -U taxhubadmin -d taxhub -tc 'CREATE EXTENSION "pg_trgm";'
+        psql -h localhost -U taxhubadmin -d taxhub -tc 'CREATE EXTENSION "ltree";'
       env:
         PGPASSWORD: taxhubpwd
     - uses: actions/checkout@v3

apptax/migrations/versions/83d7105edb76_taxref_tree.py

@@ -0,0 +1,78 @@
+"""create vm_taxref_tree
+
+Revision ID: 83d7105edb76
+Revises: 44447746cacc
+Create Date: 2024-10-05 17:40:11.302423
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision = "83d7105edb76"
+down_revision = "6a20cd1055ec"
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    op.execute("CREATE EXTENSION IF NOT EXISTS ltree")
+    op.execute(
+        """
+        CREATE MATERIALIZED VIEW taxonomie.vm_taxref_tree AS
+        WITH RECURSIVE childs AS (
+                SELECT
+                    t.cd_nom,
+                    t.cd_ref::TEXT::ltree AS path,
+                    1 AS path_length,
+                    t_ref.cd_sup AS cd_sup
+                FROM
+                    taxonomie.taxref t
+                JOIN taxonomie.taxref t_ref ON
+                    t.cd_ref = t_ref.cd_nom
+            UNION ALL
+                SELECT
+                    child.cd_nom AS cd_nom,
+                    parent.cd_ref::TEXT || child.path AS path,
+                    child.path_length + 1 AS path_length,
+                    parent_ref.cd_sup AS cd_sup
+                FROM
+                    childs child
+                JOIN taxonomie.taxref parent ON
+                    child.cd_sup = parent.cd_nom
+                JOIN taxonomie.taxref parent_ref ON
+                    parent.cd_ref = parent_ref.cd_nom
+        )
+        SELECT
+            DISTINCT ON
+            (cd_nom) cd_nom,
+            path
+        FROM
+            childs
+        ORDER BY
+            cd_nom,
+            path_length DESC
+        WITH DATA;
+        """
+    )
+    op.create_index(
+        index_name="taxref_tree_cd_nom_idx",
+        schema="taxonomie",
+        table_name="vm_taxref_tree",
+        columns=["cd_nom"],
+        unique=True,
+    )
+    # required for these operators: <, <=, =, >=, >, @>, <@, @, ~, ?
+    op.create_index(
+        index_name="taxref_tree_path_idx",
+        schema="taxonomie",
+        table_name="vm_taxref_tree",
+        columns=["path"],
+        postgresql_using="gist",
+    )
+
+
+def downgrade():
+    op.execute("DROP MATERIALIZED VIEW taxonomie.vm_taxref_tree")

apptax/taxonomie/commands/utils.py

@@ -231,6 +231,7 @@ def refresh_taxref_vm():
     db.session.execute("REFRESH MATERIALIZED VIEW taxonomie.vm_phylum")
     db.session.execute("REFRESH MATERIALIZED VIEW taxonomie.vm_regne")
     db.session.execute("REFRESH MATERIALIZED VIEW taxonomie.vm_taxref_list_forautocomplete")
+    db.session.execute("REFRESH MATERIALIZED VIEW taxonomie.vm_taxref_tree")
 
 
 def get_csv_field_names(f, encoding, delimiter):

apptax/taxonomie/models.py

@@ -3,7 +3,7 @@
 
 from sqlalchemy.ext.hybrid import hybrid_property
 from sqlalchemy.schema import FetchedValue
-from sqlalchemy.orm import deferred, raiseload, joinedload
+from sqlalchemy.orm import backref, deferred, raiseload, joinedload, foreign, remote
 
 from utils_flask_sqla.serializers import serializable
 from ref_geo.models import LAreas
@@ -176,6 +176,7 @@ class Taxref(db.Model):
     synonymes = db.relationship(
         "Taxref", foreign_keys=[cd_ref], primaryjoin="Taxref.cd_ref == Taxref.cd_ref", uselist=True
     )
+    parent = db.relationship("Taxref", primaryjoin=foreign(cd_sup) == remote(cd_ref))
     attributs = db.relationship("CorTaxonAttribut", back_populates="taxon")
     listes = db.relationship("BibListes", secondary=cor_nom_liste, back_populates="noms")
     medias = db.relationship("apptax.taxonomie.models.TMedias", back_populates="taxon")
@@ -229,6 +230,9 @@ def where_params(cls, filters=None, *, query):
                 query = query.filter(col.ilike(value + "%"))
         return query
 
+    def __le__(self, other):
+        return self.tree <= other.tree
+
 
 @serializable
 class BibListes(db.Model):
@@ -538,6 +542,19 @@ class TMetaTaxref(db.Model):
     update_date = db.Column(db.DateTime, default=db.func.now(), nullable=False)
 
 
+class TaxrefTree(db.Model):
+    __tablename__ = "vm_taxref_tree"
+    __table_args__ = {"schema": "taxonomie"}
+    cd_nom = db.Column(db.Integer, ForeignKey("taxonomie.taxref.cd_nom"), primary_key=True)
+    taxref = db.relationship(Taxref, backref=backref("tree", uselist=False))
+    path = db.Column(db.String, nullable=False)
+
+    def __le__(self, other):
+        # self <= other means taxon other is the same or a parent of self
+        p1, p2 = self.path.split("."), other.path.split(".")
+        return len(p1) >= len(p2) and p1[: len(p2)] == p2
+
+
 # Taxref deffered properties
 
 Taxref.nb_medias = deferred(

apptax/tests/test_models.py

@@ -0,0 +1,46 @@
+import pytest
+import sqlalchemy as sa
+
+from .fixtures import *
+from apptax.taxonomie.models import Taxref, TaxrefTree
+
+
+@pytest.mark.usefixtures("client_class", "temporary_transaction")
+class TestModels:
+    def test_taxref_tree_comparison(self):
+        animalia = db.session.execute(
+            sa.select(TaxrefTree).where(TaxrefTree.cd_nom == 183716)
+        ).scalar_one()
+        capra_ibex = db.session.execute(
+            sa.select(TaxrefTree).where(TaxrefTree.cd_nom == 61098)
+        ).scalar_one()
+        cinnamon = db.session.execute(
+            sa.select(TaxrefTree).where(TaxrefTree.cd_nom == 706584)
+        ).scalar_one()
+
+        assert animalia <= animalia
+        assert capra_ibex <= animalia
+        assert not animalia <= capra_ibex
+        assert not cinnamon <= animalia
+        assert not animalia <= cinnamon
+        assert not cinnamon <= capra_ibex
+        assert not capra_ibex <= cinnamon
+
+    def test_taxref_comparison(self):
+        animalia = db.session.execute(
+            sa.select(Taxref).where(Taxref.cd_nom == 183716)
+        ).scalar_one()
+        capra_ibex = db.session.execute(
+            sa.select(Taxref).where(Taxref.cd_nom == 61098)
+        ).scalar_one()
+        cinnamon = db.session.execute(
+            sa.select(Taxref).where(Taxref.cd_nom == 706584)
+        ).scalar_one()
+
+        assert animalia <= animalia
+        assert capra_ibex <= animalia
+        assert not animalia <= capra_ibex
+        assert not cinnamon <= animalia
+        assert not animalia <= cinnamon
+        assert not cinnamon <= capra_ibex
+        assert not capra_ibex <= cinnamon

docs/changelog.md

@@ -27,6 +27,7 @@ Si vous utilisez GeoNature, TaxHub sera désormais intégré à celui-ci et il n
 - Dépreciation de la route `/taxoninfo` au profit de la route `/taxref` (#554)
 - Ajout d'une route `/bdc_statuts/status_symbologies` renvoyant la symbologie de statuts des taxons (couleurs des valeurs des listes rouges) (#510, par @edelclaux)
 - Amélioration de la vue `taxonomie.vm_taxref_list_forautocomplete` pour afficher tous les noms d'un taxon (#332, par @JulienCorny et @andriacap)
+- Ajout d’une vue matérialisée `vm_taxref_tree` contenant pour chaque `cd_nom` la liste complète des `cd_ref` parents menant jusqu’au vivant (#567)
 
 **⚠️ Notes de version**
 
@@ -46,6 +47,11 @@ Si vous utilisez GeoNature, TaxHub sera désormais intégré à celui-ci et il n
 - L'image Docker de TaxHub n'est plus générée automatiquement en raison de son intégration à GeoNature (#519)
 - Suppression du code spécifique Amazon S3. Pour utiliser des services S3 de stockage des médias, il est toujours possible de monter un volume pour y déposer directement les médias.
 - Les branches `taxhub` et `taxhub-admin` ont été renommées en `taxhub-standalone` et `taxhub-standalone-sample`.
+- Si votre utilisateur PostgreSQL n’a pas la permission `CREATE EXTENSION`, vous devez manuellement créer l’extension `ltree` :
+
+```bash
+sudo -n -u postgres -s psql -d $db_name -c 'CREATE EXTENSION ltree;'
+```
 
 - Déplacement des médias à préciser/clarifier ? Avec GN ou sans c'est différent ? De /static/medias/ à media/taxhub/ ?
 

install_db.sh

@@ -58,6 +58,8 @@ then
 
     sudo -n -u postgres -s psql -d $db_name -c 'CREATE EXTENSION IF NOT EXISTS "uuid-ossp";' &>> $LOG_FILE
 
+    sudo -n -u postgres -s psql -d $db_name -c 'CREATE EXTENSION IF NOT EXISTS ltree;' &>> $LOG_FILE
+
     sudo -n -u postgres -s psql -d $db_name -c 'CREATE EXTENSION IF NOT EXISTS "postgis";' &>> $LOG_FILE
 
     echo "Extracting PostGIS version..."