Skip to main content

August 2026 Updates

· 10 min read
Brian Pepple
Brian Pepple
Founder of the Metron Project / Code Monkey

During August, Metron added its first internationalization support — a fully translatable UI with an Italian pilot translation, plus data model changes to support non-USD, non-English comics — alongside Redis-backed caching for the API's read endpoints, an average series rating with a fractional star display, and four new series genres. On the tooling side, Trok had its first public release, a new .NET library for reading and writing MetronInfo.xml. Here's everything that landed, plus the usual bug fixes and quality-of-life improvements.

Monthly Statistics​

During August the Metron Project added the following to its database:

  • Users: 583
  • Issues: 8,817
  • Creators: 876
  • Characters: 1,191
  • Reading Lists: 19

Internationalization​

Metron now has the infrastructure to serve the site in languages other than English, with Italian as the pilot translation.

i18n infrastructure. LANGUAGES/LOCALE_PATHS settings, LocaleMiddleware, the i18n template context processor, the set_language URL, and a cookie-based language switcher in the navbar lay the groundwork for translating the site.

Pilot slice, then full markup coverage. User-facing strings in the base template, navbar/footer/messages partials, and the Issue model/form/view were wrapped in gettext calls first, to prove the translation pipeline end-to-end (including generating and translating an Italian catalog) before committing to the rest of the app. With that validated, the remaining apps followed in phases — timeline, api, polls, pull_list, wish_list, user_collection, users (including both formats of the account activation e-mail), and reading_lists — wrapping templates, views, forms, and model verbose names/help texts/choices, and converting f-string messages to %-style formatting so they stay translatable. Vote counts and other countable strings use proper {% blocktrans count %}/ngettext pluralization rather than an informal "(s)" suffix. The Italian catalog now covers several hundred additional strings picked up across the rollout. With the whole app now translatable, additional languages beyond Italian will be added as needed — e.g. alongside future international publisher support — rather than as a blanket rollout.

Build and CI. GitHub Actions runners and the python:3.14-slim base image don't ship GNU gettext by default, which compilemessages needs — CI and the Containerfile build now install it before compiling (and remove it again afterward in the image, to keep the runtime slim). The PR template also gained a checklist reminder for contributors to wrap new user-facing strings and run makemessages -a before opening a PR.

Italian Comics Support​

Alongside the UI translation work, the data model picked up support for non-English, non-USD comics.

New fields. Publisher.alt_names tracks alternate publisher names, the same pattern already used for series alt names; Series.language records a series' language; and Issue.price's max_digits was widened to accommodate cover prices in currencies other than USD.

EUR/ITL currencies and language filtering. New EUR/ITL currency choices and a COUNTRIES setting support non-USD pricing, and the series list gained language filter chips. alt_names and language are exposed through the admin, forms, filters, and the Publisher/Series API serializers and templates, with a GIN trigram index on Publisher.alt_names so alternate-name search stays fast, and both web and API search now match against alt_names alongside name.

API Response Caching​

Redis-backed caching for read endpoints. Detail responses now cache under a self-versioning key derived from the object's modified timestamp, so writes invalidate automatically with no explicit cache-busting. List responses cache under a per-model version counter, bumped by the same modified-cascade signals already used elsewhere in the app (extended to cover Publisher/Imprint/Universe/Series/Arc/Character/Team/Issue, plus a new Credits → Issue cascade). User-scoped endpoints — Collection, PullList, WishList, ReadingList — are intentionally excluded from list caching so one user's data can't leak into another's cached response. Every cached response now also carries an X-Cache: HIT/MISS header.

Shaking out staleness and collision bugs. Rolling caching out surfaced a run of edge cases, now fixed: embedded fields (like Issue's nested Series/Publisher/Imprint names) that didn't cascade a modified bump onto the cached parent; a Series-rename dependency on IssueViewSet that was over-broad enough to invalidate every cached issue on any unrelated issue edit in the series; a cache-key collision between an object's own retrieve response and a detail-scoped list action (e.g. /character/<id>/issue_list/) that could return one endpoint's cached response for the other; further key collisions across pagination pages and request hosts; a condition()-wrapped method that silently rebound request to the viewset instance instead of the real DRF request; a race that could cache an issue with an empty role list; and a deleted-publisher cache hit that kept returning 200 instead of 404 until the list-cache TTL expired.

Ratings​

Average series rating. Series detail pages now show an aggregate rating across all of a series' issues, alongside the existing per-issue ratings.

Fractional star fill. The star rating display now renders a partially-filled star for decimal averages instead of rounding to the nearest whole star — a 3.4 average now shows roughly 40% of the fourth star filled in.

New Series Genres​

Four new genres are now available for tagging series:

  • Action - Comics which focus on physical action, often with violence.
  • Adventure - Comics which focus on a character, or multiple characters, who embark on a quest or mission.
  • Drama - Comics which have a more serious tone to them.
  • Mystery - Comics which focus on solving a crime or mystery.

API Improvements​

Collection add/remove endpoints. POST /api/collection/add/ adds an issue to a user's collection without marking it as read, and DELETE /api/collection/{id}/ removes an item — closing the gap where PullList and WishList already had add/remove endpoints but Collection did not.

Rate Limit Monitoring​

No more duplicate notices for the same spike. The repeat-offender scan added last month only looked at the trailing lookback window with no memory of what had already been reported, so a single throttling spike could trigger a second notice on a later run purely because the window still overlapped the same day. The notifier now tracks each account's most recent notice date and ignores throttling on or before it, so only new activity after the last notice can trigger another one — a prerequisite for eventually moving this to an unattended daily schedule.

Clearer notice wording. The throttle notice and enforcement e-mails (HTML and text) previously read oddly for a single extreme day — "over the past several days" / "across 1 day, including X in a single day" — and now word single-day triggers naturally. The per-candidate summary in notify_throttled_clients also now shows how many times an account has already been notified, so admins can see escalation history at a glance instead of inferring it from the enforce/warn verb.

Account Security​

Client IP logged for signup, activation, and API auth. Signup and account-activation log lines now include the client IP in the same (user=, ip=) format already used for API authentication, and API auth logging now includes the client IP too — making it possible to grep logs and correlate a single IP behind multiple usernames, or against Basic Auth deprecation metrics. DEPLOYMENT.md documents the journalctl commands for that lookup.

Bug Fixes​

Duplicate read-date calendar widgets. Unrelated htmx swaps on the collection detail page (e.g. rating an issue) re-scanned the whole document and stacked a new calendar widget onto the read-date input each time, since the vendored calendar library's own "already attached" check doesn't work. Re-initialization is now scoped to just the swapped fragment.

Empty sort dropdown on secondary issue lists. Only the main issue, week, next-week, and future list views wired up the sort dropdown; every other view that renders the issue list template — series, arc, creator, character, universe, and team pages — left it unset, so the dropdown appeared empty and sorting silently did nothing. All of them now share the same sort helpers.

Publisher detail page crash on imprints/universes. The publisher detail page and its HTMX load-more endpoints raised a TemplateSyntaxError on any publisher with imprints or universes, since the {% blocktrans count %} tags in the template referenced series_count/issue_count annotations the view's querysets didn't actually provide. A first fix annotated the querysets, but missed that a correlated subquery returns NULL rather than 0 for an imprint with no series or a universe with no issues — which still crashed on publishers with empty imprints/universes, like DC Comics. Both counts are now coalesced to 0.

UI Improvements​

Copy-confirmation feedback on the API token page. The "copy" button on the API token page now shows the same checkmark/"Copied" confirmation already used for the identification number copy buttons on the issue detail page.

Developer Experience​

Faster test suite. Switching to MD5 password hashing under pytest and reusing the test database between runs (--reuse-db) cut the full test suite from roughly 13m40s to 51s — full PBKDF2 hashing on every user fixture had been the dominant cost by far. An unused issue-ratings backfill management command (and its tests) was also removed.

Tooling Releases​

Mokkari 4.6.0​

  • 4.5.0 - Adds alt_names (Publisher) and language (Series) fields, and widens PricePost.currency to include EUR/ITL, matching the Italian comics support above. Also adds collection_add and collection_delete methods for the new Collection add/remove endpoints.
  • 4.6.0 - Tracks Metron's new X-Cache header via Session.last_cache_status, reflecting whether the most recent cacheable request was served from cache ("HIT") or generated fresh ("MISS").

Metron-Tagger 4.15.0​

  • 4.15.0 - Updates to Mokkari 4.6.0 and wires series language through to Metadata.series.language, written as the lang attribute on MetronInfo.xml's Series element.

New: Trok 0.1.0​

Trok is a new .NET library for reading, writing, and validating MetronInfo.xml files, targeting .NET 10 — the MetronInfo.xsd counterpart to Darkseid on the Python side. Read/Write work against POCOs under the MetronInfo.Xml.Models namespace that mirror the schema's elements and attributes, and Validate checks a document against an embedded copy of the schema, plus separately enforces the "at most one primary=\"true\"" rules on IDS/ID and URLs/URL that the schema expresses as XSD 1.1 xs:assert rules .NET's built-in XSD 1.0 validator can't evaluate on its own. Published to NuGet.org. See the docs for full usage examples.

OpenCollective​

A huge thank you to everyone who has contributed to our Open Collective! Your support makes a real difference in keeping the Metron Project running and growing.

What Your Contributions Support​

Funds from Open Collective go directly toward:

  • Server hosting costs - Keeping the Metron website and API available
  • Domain registration - Annual domain name renewals
  • Future capacity increases - Scaling resources as the database and user base grows

All expenses are transparent and publicly viewable on our Open Collective page, so you can see exactly where every dollar goes.

Support the Project​

As covered in our supporter rate limits post, donors now automatically get an elevated daily API rate limit. Any contribution, at any tier, genuinely helps keep Metron free for the whole community.

Thanks to everyone who contributed issues, pull requests, and feedback this month. As always, the project is open source and community contributions are welcome.