feat: add audio support with docling pipelines

[MRGuziX]

Description

• add ASR pipeline
• add new audio formats to support

Motivation

Why is this change needed?

Changes

• Bullet point list of changes
• Another change

Testing

How did you test this?

• use some🔮 spells 🪄 on the ContainerImage:
  to install libs

    pip install -U --index-url "https://pypi.org/simple" --no-cache-dir \
        'docling-slim[standard,feat-chunking,format-opendocument,asr]~=2.107.0' \
        'openai-whisper'

to pull needed model:

# Pre-cache Whisper turbo model for offline/air-gapped use
RUN mkdir -p /opt/app-root/docling-artifacts/models && \
    curl -L -o /opt/app-root/docling-artifacts/models/large-v3-turbo.pt \
        https://openaipublic.azureedge.net/main/whisper/models/aff26ae408abcbd652a906f2a3b109c163511b0df4ba0bbafcdc0a3dbb2f053e/large-v3-turbo.pt && \
    chown -R 1001:0 /opt/app-root/docling-artifacts/models && \
    chmod -R g+rw /opt/app-root/docling-artifacts/models

and to install FFMPEG lib to process the audio/video files:

RUN curl -L https://johnvansickle.com/ffmpeg/releases/ffmpeg-release-amd64-static.tar.xz | \
    tar -xJ --strip-components=1 -C /usr/local/bin/ --wildcards '*/ffmpeg' '*/ffprobe'

• created dev image with autorag
• uploaded image to quay
• created the audio samples
• run experiment

Checklist

• Tests added/updated
• Documentation updated
• Code follows style guide
• All checks passing