From 196dcf9eabcc1c870c5c93b82a5eab0d039b35b7 Mon Sep 17 00:00:00 2001
From: "Haoyu (Daniel)" <yanghaoyu97@outlook.com>
Date: Sun, 20 Oct 2024 11:02:11 +0800
Subject: [PATCH] lazily import torch for MontyDecoder

---
 src/monty/json.py | 29 +++++++++++++++--------------
 1 file changed, 15 insertions(+), 14 deletions(-)

diff --git a/src/monty/json.py b/src/monty/json.py
index e05aab0b..25ae42a5 100644
--- a/src/monty/json.py
+++ b/src/monty/json.py
@@ -52,11 +52,6 @@
     orjson = None
 
 
-try:
-    import torch
-except ImportError:
-    torch = None
-
 __version__ = "3.0.0"
 
 
@@ -803,15 +798,21 @@ def process_decoded(self, d):
                             d = {k: self.process_decoded(v) for k, v in data.items()}
                             return cls_(**d)
 
-                elif torch is not None and modname == "torch" and classname == "Tensor":
-                    if "Complex" in d["dtype"]:
-                        return torch.tensor(  # pylint: disable=E1101
-                            [
-                                np.array(r) + np.array(i) * 1j
-                                for r, i in zip(*d["data"])
-                            ],
-                        ).type(d["dtype"])
-                    return torch.tensor(d["data"]).type(d["dtype"])  # pylint: disable=E1101
+                elif modname == "torch" and classname == "Tensor":
+                    try:
+                        import torch
+
+                        if "Complex" in d["dtype"]:
+                            return torch.tensor(  # pylint: disable=E1101
+                                [
+                                    np.array(r) + np.array(i) * 1j
+                                    for r, i in zip(*d["data"])
+                                ],
+                            ).type(d["dtype"])
+                        return torch.tensor(d["data"]).type(d["dtype"])  # pylint: disable=E1101
+
+                    except ImportError:
+                        pass
 
                 elif np is not None and modname == "numpy" and classname == "array":
                     if d["dtype"].startswith("complex"):