#176 - smart output file renaming

[Cubix33]

Closes #176

📝 Description

This PR introduces Smart Output Renaming, replacing static timestamped filenames with dynamic, context-aware names based on the data extracted by the LLM.

🛠️ Key Changes

• filler.py: Refactored the fill_form method to run the LLM extraction before generating the output filename.
• Regex Sanitization: Added a robust string cleaner using the re module to strip illegal file characters and convert spaces to underscores, ensuring safe saves across all operating systems.
• Dynamic Naming Logic: The script now searches the JSON dictionary for common identifiers (e.g., "Employee's name", "Name"). If found, it appends this to the base filename.
• Fallback Preserved: If the LLM fails to extract a name (or returns "-1"), the system safely falls back to the original _YYYYMMDD_HHMMSS_filled.pdf format.

📸 Evidence

Before:
temp_outfile_20260304_120000_filled.pdf

After:
temp_outfile_John_Doe.pdf

🧪 Quality Check

• Tested with standard names (John Doe).
• Tested with special characters/plural lists (safely extracts the first string and removes illegal chars).
• Tested fallback logic when no name is present in the transcript.