JJ1WTL selected-source mirror

Current capture

All 43 selected articles are archived as original Seesaa HTML under /data/brific/jj1wtl/mirror/pages/<article-id>/article.html. The Japanese proxy returned HTTP 200 for every article; the capture was verified on 18 September 2026.

The archive contains 1,860 image files (1,857 JPEGs and three PNGs), plus 43 copies of the linked site stylesheet. Both thumbnails and linked full-size originals are saved. The HTML includes 17 served comments.

All 2,018 manifest records passed file-size and SHA-256 checks. The manifest includes 43 article/link-inventory pairs, 1,903 media records, 28 failed direct-access responses and one supplied reference image.

The 43 full English article translations link to the original Japanese HTML and include all 17 comments with author names and timestamps. They translate article text, captions and tables; writing inside photographs and diagrams is not fully transcribed. The original images remain available alongside the translations.

Running the mirror

config/jj1wtl_urls.txt selects the articles. The script does not recursively crawl cross-references.

uv run --locked python tools/mirror_jj1wtl.py

tools/mirror_jj1wtl.py reads OXYLABS_PROXY_USERNAME and OXYLABS_PROXY_PASSWORD from the ignored creds/oxylabs_proxy.env, resolved relative to the repository. Environment variables take precedence. Article requests use ddc.oxylabs.io:8001 when credentials are present; without them the script tries direct access. Media requests bypass the proxy and go directly to jj1wtl.up.seesaa.net.

The script stores no proxy credentials in the archive. It parses HTML links locally with Python's HTMLParser; no Reader service is used. An HTTP 200 page without Seesaa article content is rejected. Known downloaded media are reused, and running a smaller --urls selection retains other manifest entries. Failures produce a nonzero exit status.

The output defaults to $BRIFIC_DATA_DIR/jj1wtl/mirror, or /data/brific/jj1wtl/mirror. Use --data-dir to change the data root, or --output to choose a specific mirror directory. --provenance-file can additionally record a researcher-supplied reference.

Layout

Each pages/<article-id>/ contains:

  • article.html: original HTML bytes, including the comments served with that page.
  • html-links.txt: URLs extracted from HTML href and src attributes.
  • media/: downloaded images and linked site stylesheet.
  • direct-response.html, where present: an earlier failed direct response, recorded as an error in the manifest.

manifest.json records URLs, retrieval timestamps, HTTP status, content type, response metadata, byte length, SHA-256 and the article associated with each media file. Raw HTML preserves the original external URLs; it is not rewritten into an offline website.

Central diagrams and screenshot provenance

The original blog decoder diagram was recovered from article 519927784 (English translation):

  • Source URL: https://jj1wtl.up.seesaa.net/image/20260208boso-peninsula-map.jpg
  • Local path: /data/brific/jj1wtl/mirror/pages/519927784/media/20260208boso-peninsula-map.jpg
  • HTTP status/content type: 200, image/jpeg
  • Bytes: 386,162
  • SHA-256: 8a650978d4617a6419df5db7eb2ccadbd3cfd76023056e4629e063741d5aad1b

The exact user-supplied PNG is preserved separately:

  • Origin: /home/alex/.codex/attachments/c1b74f5d-5c91-47c1-a56b-385d68d8b193/codex-clipboard-3336f690-6b59-43a9-8f81-579742885f71.png
  • Local path: /data/brific/jj1wtl/mirror/researcher-supplied/codex-clipboard-3336f690-6b59-43a9-8f81-579742885f71.png
  • Bytes: 1,677,823
  • SHA-256: 3a8fa0a633837a6b4a601e52a15557bf5371a61cf64dfa067981efdb3bab11ec

ImageMagick pixel comparison reports 0 (0) differing pixels between the source JPEG after decode and the supplied PNG. The PNG is therefore a pixel-identical lossless rendering of the recovered blog image, although their encoded bytes and hashes differ.

The two central August 23 assets are also present as original JPEG bytes (article translation):

  • Updated operator/site map: /data/brific/jj1wtl/mirror/pages/521403607/media/20260823boso-peninsula-map.jpg; 387,219 bytes; SHA-256 4be599424dc85eb59bb273002c673d75c3907dfb3b0b188bac7b035c22466894
  • Chikura/urban/backhaul route schematic: /data/brific/jj1wtl/mirror/pages/521403607/media/20260823boso-peninsula-table.jpg; 385,553 bytes; SHA-256 d7bc2c67e408864bdc3b0c461148c661cca4773c1c9807ddf959007547b61f29