Search inside documents: PDF text, Tesseract OCR for scans (in the image), optional Claude; nightly catch-up

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
rickyandClaude Opus 5.5 committed 2026-11-11 21:09:12 -05:00
1 parent 2953b9a18d
commit 08ba017c7e
14 files changed
+271 -6

No files matched your search

+8
View File
@@ -38,5 +38,13 @@
<p class="hint">With off-site backup to Google Drive, this many of the newest nightly database copies are kept there. Uploaded files are copied once and kept.</p>
</div>
<h2 class="card-title pt-2">Document search</h2>
<div>
<%= label_tag "settings[documents_claude_text]", "Read scanned documents with", class: "label" %>
<%= select_tag "settings[documents_claude_text]", options_for_select([ [ "Tesseract (free, on this server)", 0 ], [ "Claude (better on handwriting; uses the Anthropic API)", 1 ] ], @values["documents_claude_text"].to_i),
id: "settings_documents_claude_text", class: "input" %>
<p class="hint">Text in documents is read so search can find it. PDFs that contain text are read directly either way. Claude is used only when an ANTHROPIC_API_KEY is set<%= " (it isn't)" if ENV["ANTHROPIC_API_KEY"].blank? %>, and costs API credit per page.</p>
</div>
<div class="flex gap-2"><%= f.submit "Save Settings", class: "btn-primary" %><%= link_to "Cancel", admin_about_path, class: "btn-secondary" %></div>
<% end %>