Skip to content

SUYA 0.1.11

Choose a tag to compare

@github-actions github-actions released this 13 Nov 08:48
· 10 commits to main since this release

Download / 내려받기

OS File / 파일 Size First run / 처음 실행
Windows 10/11 x64 SUYA-0.1.11-windows-x64-setup.exe ≈ 50 MB Installer (English/Korean). SmartScreen: More info → Run anyway / "추가 정보 → 실행"
macOS 12+ Apple Silicon (M1–M4) SUYA-0.1.11-macos-arm64.dmg ≈ 58 MB Drag to Applications. macOS 15+: open once, then System Settings → Privacy & Security → Open Anyway; macOS 12–14: right-click → Open / macOS 15 이상: 한 번 연 뒤 시스템 설정 → 개인정보 보호 및 보안 → 그래도 열기, 12~14: 우클릭 → 열기
macOS 12+ Intel SUYA-0.1.11-macos-x86_64.dmg ≈ 64 MB Same / 위와 같음
Linux x86_64 (glibc ≥ 2.35) SUYA-0.1.11-linux-x86_64.tar.gz ≈ 109 MB Extract, run ./install.sh / 풀고 ./install.sh

Not sure which Mac you have? Apple menu → About This Mac: "Chip Apple M…" → Apple Silicon, "Processor … Intel" → Intel. / Apple 메뉴 → 이 Mac에 관하여: "칩 Apple M…"이면 Apple Silicon, "프로세서 … Intel"이면 Intel.

Already on 0.1.2 to 0.1.10? You do not need to download anything. SUYA finds this release by itself when it starts and at its regular check: every 6 hours on 0.1.10, once a day on earlier versions. 0.1.5 to 0.1.10 install it by themselves after a 20-second prompt that lets you postpone; 0.1.2, 0.1.3 and 0.1.4 ask before installing (Help ▸ Check for updates… looks right away).
0.1.2~0.1.10을 쓰고 있으면 내려받을 필요가 없습니다. 시작할 때와 정기 확인 때 이 릴리스를 스스로 찾습니다. 정기 확인은 0.1.10이 6시간마다, 그 전 버전이 하루 한 번입니다. 0.1.5~0.1.10은 20초 동안 미룰 수 있게 물어본 뒤 스스로 설치하고, 0.1.2·0.1.3·0.1.4는 설치 전에 물어봅니다(*도움말 ▸ 업데이트 확인…*으로 바로 확인).

Checksums: SHA256SUMS.txt (below, under Assets). Every file is built by GitHub Actions from this tag (.github/workflows/release.yml) and installed on the build machine, where the installed copy has to pass its self-test before the file is attached. The builds are unsigned, hence the first-run warning.
Updating from 0.1.0 or 0.1.1: install over the old version. Your database, settings and Excel workbooks stay where they are.
Python users: pip install "suya[gui] @ git+https://github.com/jongsu-king/SUYA.git@v0.1.11" then run suya. Without [gui] you get the collector alone, with no Qt.
Containers: docker build -t suya https://github.com/jongsu-king/SUYA.git#v0.1.11, then docker run -d --name suya --restart unless-stopped -v suya-data:/data suya.
macOS with Homebrew: brew install --cask jongsu-king/suya/suya installs the same disk image, with the same first-run warning.
Homebrew를 쓰는 macOS: brew install --cask jongsu-king/suya/suya로 같은 디스크 이미지를 설치합니다. 처음 실행 경고도 같습니다.


SUYA 0.1.11 — Antarctic Ocean Monitor: Southern Ocean temperature, salinity and air temperature collected continuously from public APIs into SQLite and monthly Excel; south-polar map, profiles, T-S diagrams; English/Korean.
SUYA 0.1.11 — 남극해 수온·염분·기온 실시간 수집 프로그램.

Copyright © 2026 Jongsu Lim (Department of Physics, HMOL, Incheon National University; imjongsu1102@gmail.com; ORCID 0009-0000-3395-1484) and Suyeon Park (Department of Materials Science and Engineering, CELLMATE, Incheon National University; dbsnt110@inu.ac.kr) · MIT License

DOI — all versions (concept DOI): 10.5281/zenodo.23937852 · this version: 10.5281/zenodo.24189873 · archived on Zenodo
DOI — 전체 버전(개념 DOI): 10.5281/zenodo.23937852 · 이 버전: 10.5281/zenodo.24189873

This release keeps collection going through network outages and storage failures. A source that cannot reach its server is tried again within 5 minutes, a database that cannot be written no longer stops a source until the next start, and the log records when the program starts and stops.
이번 릴리스는 네트워크가 끊기거나 저장에 실패해도 수집이 이어지게 합니다. 서버에 닿지 못한 소스는 5분 안에 다시 시도하고, 데이터베이스에 쓸 수 없어도 소스가 다음 시작 때까지 멈춰 있지 않으며, 프로그램이 시작하고 끝난 때를 로그에 남깁니다.

Added / 더한 것

  • Start and stop are written to the log. monitor.log records when the collecting program starts, with its version and process number, and when it stops. When the run before did not stop in the usual way, the next start says so and gives the time of the last collection attempt, so a gap in the data can be traced to the moment the program went away. suya daemon writes the same lines.
    시작과 종료를 로그에 남깁니다. monitor.log에 수집 프로그램이 시작한 때를 버전·프로세스 번호와 함께 적고, 끝난 때도 적습니다. 앞선 실행이 정상적으로 끝나지 않았으면 다음 시작 때 그 사실과 마지막 수집 시도 시각을 적으므로, 데이터가 빈 구간을 프로그램이 사라진 때까지 거슬러 찾을 수 있습니다. suya daemon도 같은 줄을 남깁니다.
  • Errors nothing handled are written to the log. An error that reaches the top of the window program, or that ends one of its threads, is recorded in monitor.log with its traceback and then reported the usual way.
    처리되지 않은 오류를 로그에 남깁니다. 창 프로그램의 맨 위까지 올라온 오류나 스레드 하나를 끝낸 오류를 monitor.log에 트레이스백과 함께 적고, 그 뒤에는 전과 같이 알립니다.

Changed / 바꾼 것

  • A source that cannot reach its server is tried again within 5 minutes. A failed collection was repeated after 1, 2, 4, 8 … minutes up to the source's own interval, so after a long network outage a source could wait hours for its next attempt. A failure in which no request reached a server now waits 5 minutes at most, and as soon as one source collects again, the sources that were waiting for the same reason are woken straight away. A failure in which the server answered keeps the old schedule.
    서버에 닿지 못한 소스는 5분 안에 다시 시도합니다. 실패한 수집은 1, 2, 4, 8 … 분 뒤에 다시 했고 이 간격은 소스의 원래 주기까지 늘어났습니다. 그래서 네트워크가 오래 끊긴 뒤에는 다음 시도까지 몇 시간을 기다리기도 했습니다. 이제 요청이 하나도 서버에 닿지 못한 실패는 길어야 5분을 기다리고, 소스 하나가 다시 수집하면 같은 이유로 기다리던 소스들을 바로 깨웁니다. 서버가 답한 실패는 전과 같은 일정을 따릅니다.
  • A long outage no longer fills the log. While a source stays unreachable only its first failure is logged as a warning, and the log says when collection works again. A database that refuses writes is reported the same way: once, and once more when it accepts them again.
    긴 장애가 로그를 채우지 않습니다. 소스가 계속 닿지 않는 동안에는 첫 실패만 경고로 적고, 수집이 다시 되면 그 사실을 적습니다. 쓰기를 거부하는 데이터베이스도 같은 방식으로 알립니다. 한 번 적고, 다시 쓸 수 있게 되면 한 번 더 적습니다.
  • HYCOM, Open-Meteo Marine and Windy tell an unreachable server from a failed answer. These collectors reported every failure alike. They now report "unreachable" when none of their requests reached a server, which is what the shorter retry above relies on.
    HYCOM, Open-Meteo Marine, Windy가 닿지 않는 서버와 실패한 응답을 구별합니다. 이 수집기들은 모든 실패를 똑같이 보고했습니다. 이제 요청이 하나도 서버에 닿지 못했으면 "닿지 않음"으로 보고하며, 위의 짧은 재시도가 이 구별에 기대고 있습니다.
  • The write-ahead log beside the database stays small. SQLite cuts the -wal file back to 16 MB after a checkpoint and no longer leaves it at its largest size.
    데이터베이스 옆의 미리 쓰기 로그(WAL)가 작게 유지됩니다. SQLite가 체크포인트 뒤에 -wal 파일을 16 MB로 줄이고, 가장 커졌던 크기로 남겨 두지 않습니다.

Fixed / 고친 것

  • A server that cannot be reached is given up on within about a minute. Every request passed its own response limit (90 to 180 seconds, 900 for a download) in a way that also replaced the 20-second limit for opening the connection, so three attempts at a server that did not answer the connection took up to 9 minutes before the failure was recorded. On 2026-11-13 one Open-Meteo run took 4 minutes 37 seconds to fail this way. The connection limit is 20 seconds again and the response limits are unchanged.
    닿지 않는 서버는 1분쯤 안에 포기합니다. 요청마다 응답 제한 시간(90~180초, 내려받기는 900초)을 넘겨주던 방식이 연결을 여는 20초 제한까지 바꿔 버렸습니다. 그래서 연결에 답하지 않는 서버에 세 번 시도하면 실패가 기록되기까지 길게는 9분이 걸렸습니다. 2026-11-13에는 Open-Meteo 수집 한 번이 이렇게 실패하는 데 4분 37초가 걸렸습니다. 연결 제한은 다시 20초이고 응답 제한은 그대로입니다.
  • A write that failed midway is undone. When a write to the database failed partway, on a full disk for example, the part already done stayed pending and was saved together with the next successful write. Each write is now saved whole or undone whole.
    중간에 실패한 쓰기는 되돌립니다. 디스크가 가득 찼을 때처럼 데이터베이스에 쓰다가 중간에 실패하면, 이미 된 부분이 남아 있다가 다음에 성공한 쓰기와 함께 저장되었습니다. 이제 쓰기는 통째로 저장되거나 통째로 되돌려집니다.
  • Collection goes on when the database cannot be written. An error while saving the result of a run or the time of the next run ended that source's loop until the program was restarted, and an error in the daily clean-up ended the clean-up. Each loop now logs the failure and carries on, and the retry interval keeps growing from the loop's own count while the stored count cannot be updated.
    데이터베이스에 쓸 수 없어도 수집을 계속합니다. 수집 결과나 다음 수집 시각을 저장하다가 오류가 나면 그 소스의 반복이 프로그램을 다시 시작할 때까지 끝나 있었고, 하루 한 번의 정리에서 오류가 나면 정리가 끝나 있었습니다. 이제 각 반복은 실패를 로그에 적고 계속 돌며, 저장된 실패 횟수를 고칠 수 없는 동안에는 반복이 스스로 센 횟수로 재시도 간격을 계속 늘립니다.

After updating / 올린 뒤

  • Nothing has to be done by hand. From this version on, monitor.log has a line for each start and each stop of the program.
    손으로 할 일은 없습니다. 이 버전부터 monitor.log에 프로그램이 시작할 때와 끝날 때마다 한 줄씩 남습니다.

Full list: CHANGELOG.md.