replace infinite source width in dummy adjoint source time to avoid runtime warning for overflow error

[groberts-flex]

Recently, a change was made that started causing the runtime warnings to pop up when making the adjoint source dummy source_time for hashing. The warnings look like the one pasted below. The fix is to not use a consistent, but finite width source that covers all the frequencies necessary (if not, validators will fail for some sources like custom field source).

Greptile Overview

Updated On: 2025-10-13 15:01:59 UTC

Greptile Summary

This PR fixes runtime overflow warnings in the adjoint source hashing logic by replacing infinite frequency width with computed finite bounds. The change modifies the dummy adjoint source creation in tidy3d/components/data/sim_data.py to calculate appropriate frequency bounds based on existing adjoint sources rather than using fwidth=inf. The fix determines min_freq_tmp_src and max_freq_tmp_src from the frequency ranges of all adjoint sources, then sets the temporary source's center frequency and width to cover this range. This maintains the original hashing functionality while eliminating numerical overflow warnings that were being triggered during calculations with infinite values.

Important Files Changed

Changed Files

Filename	Score	Overview
tidy3d/components/data/sim_data.py	4/5	Replaces infinite source width with computed finite frequency bounds in dummy adjoint source creation

Confidence score: 4/5

• This PR is safe to merge with minimal risk as it addresses a specific runtime warning without changing core functionality
• Score reflects a targeted fix that maintains existing behavior while solving numerical overflow issues; no major algorithmic changes or breaking modifications
• Pay close attention to the frequency bound calculations to ensure they properly cover all necessary frequency ranges for different source types

Sequence Diagram

sequenceDiagram
    participant User
    participant SimulationData
    participant AdjointSource
    participant GaussianPulse

    User->>SimulationData: "_adjoint_src_width_broadband(adj_srcs)"
    SimulationData->>SimulationData: "extract frequencies from adj_srcs"
    SimulationData->>SimulationData: "calculate middle_f0 = 0.5 * (max + min)"
    SimulationData->>SimulationData: "calculate decay_by_f0_fwidth = middle_f0 / NUM_ADJOINT_FWIDTH_TO_ZERO"
    SimulationData->>SimulationData: "calculate fwidth_to_min_f0 = (middle_f0 - min_f0) / NUM_ADJOINT_FWIDTH_TO_FMIN"
    SimulationData->>SimulationData: "check if fwidth_to_min_f0 > decay_by_f0_fwidth and log warning if needed"
    SimulationData->>SimulationData: "adj_src_fwidth = max(decay_by_f0_fwidth, fwidth_to_min_f0)"
    SimulationData-->>User: "return (middle_f0, adj_src_fwidth)"

    User->>SimulationData: "_make_broadband_source(adj_srcs)"
    SimulationData->>SimulationData: "_adjoint_src_width_broadband(adj_srcs)"
    SimulationData->>SimulationData: "get source_index from normalize_index"
    SimulationData->>GaussianPulse: "updated_copy(amplitude=1.0, phase=0.0)"
    GaussianPulse-->>SimulationData: "src_time_base"
    SimulationData->>GaussianPulse: "updated_copy(freq0=adj_src_f0, fwidth=adj_src_fwidth)"
    GaussianPulse-->>SimulationData: "updated source_time"
    SimulationData->>AdjointSource: "updated_copy(source_time=updated_source_time)"
    AdjointSource-->>SimulationData: "src_broadband"
    SimulationData-->>User: "return src_broadband"

Loading

[greptile-apps]

_{1 file reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[github-actions]

Diff Coverage

Diff: origin/develop...HEAD, staged and unstaged changes

• tidy3d/components/data/sim_data.py (100%)

Summary

• Total: 5 lines
• Missing: 0 lines
• Coverage: 100%

[yaugenst-flex]

Thanks @groberts-flex! LGTM but need to update the commit message (fix: )