From a721d2015da29d686a3ac52ddb7ccafdefcb5b37 Mon Sep 17 00:00:00 2001
From: Yoann Schneider <yschneider@teklia.com>
Date: Thu, 16 Feb 2023 09:33:19 +0100
Subject: [PATCH] clearer comments about config serialization
---
dan/ocr/document/train.py | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/dan/ocr/document/train.py b/dan/ocr/document/train.py
index 3eb15552..9552d042 100644
--- a/dan/ocr/document/train.py
+++ b/dan/ocr/document/train.py
@@ -239,7 +239,10 @@ def get_config():
def serialize_config(config):
"""
- Serialize a dictionary to transform it into json and remove the credentials
+ Make every field of the configuration JSON-Serializable and remove sensitive information.
+
+ - Classes are transformed using their name attribute
+ - Functions are casted to strings
"""
# Create a copy of the original config without erase it
serialized_config = deepcopy(config)
@@ -282,6 +285,7 @@ def serialize_config(config):
"synthetic_data" in config["dataset_params"]["config"]
and config["dataset_params"]["config"]["synthetic_data"]
):
+ # The Probability scheduler is a function and needs to be casted to string
serialized_config["dataset_params"]["config"]["synthetic_data"][
"proba_scheduler_function"
] = str(
--
GitLab