Fix character encoding issues when loading JSON configuration files

[vicenciomf2]

Problem Description

When loading configuration files containing non-ASCII characters (like Spanish accents or special characters), the text is incorrectly encoded, resulting in mojibake (e.g., "ó" appearing as "ó"). This affects all text loaded from JSON configuration files, including prompts for LLM services.

Example of the issue:

{
    "LLMImageDescriptionProcessor_image_description_prompt": "Eres un experto en análisis de documentos especializado en crear descripciones de texto para imágenes. (...)"
}

Was being loaded as:

Eres un experto en análisis de documentos especializado en crear descripciones de texto para imágenes. (...)

And the LLM's response was, for example:

Esta imagen muestra el escudo de la Pontificia Universidad Católica de Chile.

Root Cause

The JSON configuration files were being opened without explicitly specifying UTF-8 encoding, causing the text to be read with the system's default encoding.

Solution

Modified the configuration file loading in config/parser.py to explicitly use UTF-8 encoding when opening JSON files. This is particularly important for non-English LLM prompts and responses.

[github-actions]

CLA Assistant Lite bot All contributors have signed the CLA ✍️ ✅

[vicenciomf2]

I have read the CLA Document and I hereby sign the CLA