Fixed json.loads() decode error when LLM produces code block

[FoliageOwO]

Description

Using some models may make mem0 internal responses generating JSONDecodeError:

json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)

The response from Mem0 LLM is below (using diff as code block format because the triple backticks "```" does not show at the code block below, so ignore these + that at the first of a line).

+ ```json
+ {
+     "memory": [
+         {
+             "id": "(UUID)",
+             "text": "I'm visiting Paris",
+             "event": "ADD"
+         }
+     ]
+ }
+ ```

Fixes #1854

Type of change

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

Please refer to the issue #1854.

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• Any dependent changes have been merged and published in downstream modules
• I have checked my code and corrected any misspellings

Maintainer Checklist

• Closes JSONDecodeError: AI sometimes produces JSON with a code block format #1854
• Made sure Checks passed

[prateekchhikara]

do you think we should still do "json.loads"?

[FoliageOwO]

@prateekchhikara Hi. Thanks for reviewing.

I think it may be the name of the function that confused you.

The function parse_json_response only extracts the JSON content from the potential code block from the given origin response string, it doesn't call json.loads(), which means it doesn't return a dictionary but a string.

So we HAVE to call json.loads() after the parsing progress.