Accelerate write

[JingChen-Thu]

Accelerate sr.write:

1m20s -> 1.6 s

by preparing text info before loop.

Frequently using 'itertuples' is computationally expensive.

[Copilot]

Pull request overview

This PR optimizes SrcRec.write() in pytomoatt by avoiding per-source DataFrame filtering/itertuples() over receiver data, aiming to significantly reduce runtime when writing large src_rec files.

Changes:

• Precomputes receiver/common-source/common-receiver output lines grouped by src_index to avoid repeated DataFrame filtering inside the source loop.
• Switches source iteration to itertuples(name=None) and writes prebuilt receiver blocks via dictionary lookups.

Suppressed comments (2)

pytomoatt/src_rec.py:532

• Common-source receiver record formatting reduces weight to 3 decimals (%7.3f), which changes persisted values compared to the previous 4-decimal formatting. If weights are meant to be preserved across write→read, keep the previous precision.

                    "%7d %7d %6s %9.4f %9.4f %9.4f %7d %6s %9.4f %9.4f %9.4f %s %8.4f %7.3f\n" % row

pytomoatt/src_rec.py:559

• Common-receiver receiver record formatting reduces weight to 3 decimals (%7.3f), which changes persisted values compared to the previous 4-decimal formatting. If weights are meant to be preserved across write→read, keep the previous precision.

                    "%7d %7d %6s %9.4f %9.4f %9.4f %7d %6s %9.4f %9.4f %9.4f %s %8.4f %7.3f\n" % row

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[xumi1993]

@copilot Fix the code for all comments in this review thread.

When a review comment includes a suggested change, apply the suggestion exactly.

Do not make changes beyond what is described in the linked review thread.

[Copilot]

@copilot Fix the code for all comments in this review thread.

When a review comment includes a suggested change, apply the...

All issues from the review have been addressed in commit $(git rev-parse --short HEAD):

1. Weight precision (lines 495, 532, 559): Restored %7.4f (4 decimal places) for receiver weight fields.
2. Source header format (line 585): Restored the strftime("%Y_%m_%d_%H_%M_%S.%f") approach to preserve full microsecond precision for second, and {:.4f} for mag and weight.
3. Column order dependency (lines 566–570): Replaced positional src[1], src[2], … indexing with an explicit column list (_src_cols) passed to itertuples, so the mapping is stable regardless of DataFrame column reordering.
4. Regression test: Added test_write_roundtrip_preserves_microsecond_and_weight_precision to assert that write→read roundtrips preserve microsecond origin-time precision and 4-decimal weight precision for both source and receiver records.