From 008a5e3579e6ae34705386ef46d3cc56d399709f Mon Sep 17 00:00:00 2001
From: "amaury.zarzelli" <amaury.zarzelli@ign.fr>
Date: Tue, 27 Feb 2024 10:43:43 +0100
Subject: [PATCH] change(pivot2pgr): do not use clean_graph() procedure

---
 changelog.md          |  5 +++++
 r2gg/__about__.py     |  2 +-
 r2gg/_pivot_to_pgr.py | 34 ++++++++++++++++++++++++++++++++--
 3 files changed, 38 insertions(+), 3 deletions(-)

diff --git a/changelog.md b/changelog.md
index 2d20644..c5d0e45 100644
--- a/changelog.md
+++ b/changelog.md
@@ -2,6 +2,11 @@
 
 ## x.y.z
 
+## 2.2.4
+
+CHANGED:
+- Pivot to pgr: Not using clean_grpah method anymore, but a SQL query
+
 ## 2.2.3
 
 CHANGED:
diff --git a/r2gg/__about__.py b/r2gg/__about__.py
index 425afd0..10db08c 100644
--- a/r2gg/__about__.py
+++ b/r2gg/__about__.py
@@ -34,7 +34,7 @@
 __uri_tracker__ = f"{__uri_repository__}issues/"
 __uri__ = __uri_repository__
 
-__version__ = "2.2.3"
+__version__ = "2.2.4"
 __version_info__ = tuple(
     [
         int(num) if num.isdigit() else num
diff --git a/r2gg/_pivot_to_pgr.py b/r2gg/_pivot_to_pgr.py
index ded60ff..2d0686a 100644
--- a/r2gg/_pivot_to_pgr.py
+++ b/r2gg/_pivot_to_pgr.py
@@ -395,13 +395,43 @@ def pivot_to_pgr(source, cost_calculation_file_path, connection_work, connection
 
     for profile_name in profile_names:
         logger.info("Cleaning isolated edges for profile {}...".format(profile_name))
-        clean_graph_query = "SELECT {0}.clean_graph('{1}')".format(schema, profile_name)
+        clean_graph_query = """
+        WITH connected_components AS (
+          SELECT * FROM pgr_connectedComponents(
+            'SELECT id, source, target, cost_s_{1} as cost, reverse_cost_s_{1} as reverse_cost FROM {0}.ways'
+          )
+        ),
+        remove_nodes AS (
+          SELECT node
+          FROM
+            connected_components
+          WHERE
+          component = ANY(
+            SELECT DISTINCT component
+            FROM
+            (
+              SELECT component, count(*) AS nb
+              FROM
+                connected_components
+              GROUP BY component
+            )
+            AS components
+            WHERE nb <= 10
+          )
+        )
+        UPDATE {0}.ways
+        SET cost_s_{1} = -1,
+          reverse_cost_s_{1} = -1,
+          cost_m_{1} = -1,
+          reverse_cost_m_{1} = -1
+        WHERE {0}.ways.target = ANY(SELECT * from remove_nodes) OR {0}.ways.source = ANY(SELECT * from remove_nodes);
+        """.format(schema, profile_name)
         logger.info("SQL: {}".format(clean_graph_query))
         cursor_isolated.execute(clean_graph_query)
+        connection_out.commit()
 
     et_execute = time.time()
     logger.info("Execution ended. Elapsed time : %s seconds." %(et_execute - st_execute))
-    connection_out.commit()
     cursor_isolated.close()
 
     end_time = time.time()