refactor: Centralize logger import to src.logger_config across various modules.

Refactor: Standardize logging by replacing print statements with logger calls and adjusting verbosity levels.

feat: Add AI-driven verification and correction for speech-to-text timed transcripts.

feat: Improve JSON parsing robustness by switching to json5 and explicitly instructing the model to avoid trailing commas in its output.

refactor: Update GCS file uploads to use a `folder` parameter for temporary files and enable old video deletion.

feat: Standardize temporary GCS uploads to a dedicated bucket, refactor video deletion logic to preserve recent files or clear folders, and fix a workflow environment typo.

feat: Implement centralized logging with colored output, verbose control, and updated log levels across the application.

fix: Move product detection before video generation to eliminate double-gen

feat: Implement refined product integration flow that splits segments based on AI product detection for dynamic video generation and selection.

feat: Implement product video detection and merging into the voiceover pipeline using Gemini and MoviePy.

chore: Disable video library matching to always generate new videos.

feat: Merge system and negative prompts into a single enriched prompt for Grok video generation and refine prompt templates to prevent text.

feat: Calculate and apply audio durations per script segment using timed transcripts for more accurate video generation.

feat: Calculate and return the actual duration of generated videos and increase the default video generation duration to 4.0 seconds.

feat: skip audio compliance check in VoiceOverAIPipeline

feat: Refactor video generation to centralize provider selection and pass provider-specific arguments via `generate_video_process`.

feat: Implement VoiceOver AI Pipeline and Refactor Config

feat: Refine b-roll video title generation prompt with more detailed instructions and a random seed for variation.

feat: Add temporary method to override the audio library with specific tracks.

feat: enforce strict 9:16 video aspect ratio during sync and upload, and add Drive file content update functionality.

refactor: Standardize imports to use src as the root package.

refactor: Adjust import paths to use `src.config` and enhance text clip rendering with dynamic font sizing and vertical positioning.

feat: Add error handling for b-roll video title generation to log failures and return None.

feat: Allow specifying GCS destination blob name with uniqueness handling, generate AI-powered b-roll filenames, and update Gemini model to flash-latest.

feat: Improve gcloud ADC setup with client secret and default prompts, and add test automation skip for video cropping.

feat: Implement text overlay management from Google Sheets and refactor pipeline selection based on AI generation and video merge type.

refactor: merge render_interval_video and render_random_video into unified render_video_cuts

feat: async-only downloads with parallel GCS, sequential Drive

refactor: Pipeline architecture revamp with proper separation