Document to_png() function usage and figure display mechanism

[Copilot]

Analysis of how _to_png() is called and how figures are rendered in the Jupyter frontend.

Findings

_to_png() Definition

Located in alpaca_kernel/alpaca/utils.py:

def _to_png(fig):
    imgdata = BytesIO()
    fig.savefig(imgdata, format='png', bbox_inches='tight')
    imgdata.seek(0)
    return imgdata.read()

Call Site

ipkernel.py:905 in sendPLOT():

png = _to_png(self.fig)

Display Mechanism

Figures are sent to Jupyter via IOPub socket using standard messaging protocol:

• New plots: send_response(iopub_socket, 'display_data', content)
• Updates: send_response(iopub_socket, 'update_display_data', content)

Content structure:

{
    'data': {'image/png': png},
    'metadata': {'image/png': {'width': 600, 'height': 400}},
    'transient': {'display_id': str(plot_uuid)}
}

Plot Modes

• Mode 1: Matplotlib-style (array plotting)
• Mode 2: Live plotting
• Mode 3: Scope-style with trigger
• Mode 4: EEG-style with auto-scroll

Original prompt

find where to_png() function is called and where and how these figures are shown in jupyter frontend

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.