From d8588c25f3ac0e912179b1eb192ce4c5b70758b0 Mon Sep 17 00:00:00 2001
From: erezalg <52002872+erezalg@users.noreply.github.com>
Date: Mon, 5 Dec 2022 13:15:01 +0200
Subject: [PATCH] Add path substitution docs (#394)

---
 docs/integrations/storage.md | 28 ++++++++++++++++++++++++++--
 1 file changed, 26 insertions(+), 2 deletions(-)

diff --git a/docs/integrations/storage.md b/docs/integrations/storage.md
index 2f3b099e..501e0ef3 100644
--- a/docs/integrations/storage.md
+++ b/docs/integrations/storage.md
@@ -184,6 +184,31 @@ content directly from code.
 
 See [Storage Examples](../guides/storage/examples_storagehelper.md).
 
+### Path Substitution
+The ClearML StorageManager supports local path substitution when fetching files.<br/>
+This is especially useful when managing data using [clearml-data](https://clear.ml/docs/latest/docs/clearml_data/clearml_data_cli)! If different data consumers have the data physically stored in different locations, path 
+substitution allows for registering the data into clearml-data once, and then storing it, and accessing it, in multiple locations.
+
+To enable path substitution, modify the clearml.conf file and configure:
+
+```bash
+sdk{
+    storage{
+        path_substitution = [
+            # Replace registered links with local prefixes,
+            # Solve mapping issues, and allow for external resource caching.
+            # {
+            #     registered_prefix = "s3://bucket/research"
+            #     local_prefix = "file:///mnt/shared/bucket/research
+            # },
+            # {
+            #     registered_prefix = "file:///mnt/shared/folder/"
+            #     local_prefix = "file:///home/user/shared/folder"
+            # }
+        ]
+    }
+}
+```
 
 ## Caching
 ClearML also manages a cache of all downloaded content so nothing is duplicated, and code won't need to download the same
@@ -213,5 +238,4 @@ By default, all artifacts (Models / Artifacts / Datasets) are automatically down
 Some storage mediums (NFS / Local storage) allows for direct access,
 which means that the code would work with the object where it's originally stored and not downloaded to cache first.
 
-To enable direct access, specify the urls to access directly.
-
+To enable direct access, specify the urls to access directly.
\ No newline at end of file