Skip to content

Add document-image OCR workflow and clue-driven search query generation - #5

Merged
potemkin666 merged 13 commits into
mainfrom
copilot/remove-image-and-workflow
May 5, 2026
Merged

Add document-image OCR workflow and clue-driven search query generation#5
potemkin666 merged 13 commits into
mainfrom
copilot/remove-image-and-workflow

Conversation

Copilot AI commented May 5, 2026

Copy link
Copy Markdown
Contributor

BlueLens handled OCR as raw text extraction, but it did not treat document-style images as first-class evidence surfaces or turn extracted clues into ready-to-run searches. This change adds a document-image workflow for structured OCR review and a query generator that composes higher-signal searches from text, brands, locations, usernames, filenames, dimensions, and language hints.

  • Document-image workflow

    • Adds a document_image mission and direct Text-tab action for document-heavy images: letters, forms, IDs, posters, flyers, menus, receipts, certificates, and similar material.
    • Structures OCR output into:
      • document-kind inference
      • layout summary (headings, key-value rows, tabular rows)
      • logo/brand lookup candidates
      • entity summary across orgs, people, locations, dates, domains, handles, phones

  • Search query generator

    • Adds a query_generator mission and direct Text-tab action.
    • Generates focused follow-up searches from available clues, including:
      • brand + city
      • OCR text + logo
      • object + language
      • file name + dimensions
      • visible username + platform
      • logo + domain
      • quoted OCR line

  • UI surface

    • Extends the workflow selector with:
      • Document Image
      • Search Query Generator
    • Adds dedicated output panels in the Text tab so document analysis and generated searches stay visible independent of generic mission output.

  • Export/report integration

    • Includes document-mode and query-generator outputs in report JSON/markdown exports so OCR interpretation is preserved with the rest of the investigation record.

  • Shared logic

    • Moves pure document/query heuristics into bluelens-helpers.js so the browser and test suite use the same logic for:
      • layout summarization
      • document-kind inference
      • logo candidate extraction
      • search query composition
const output = buildSearchQueryGeneratorOutput({
  ent,
  handles,
  domains,
});

// Example generated searches:
// - "Acme Festival" "Berlin"
// - "JOIN US THIS FRIDAY" "Acme Festival"
// - "@acmefest" instagram.com

Copilot AI and others added 13 commits May 5, 2026 17:26
@potemkin666
Remove landing promo copy
41d21a8 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/6cd3f2cb-45d7-4938-9b7f-bcaaafbfe12f

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Align UI regression coverage
3e0dbad 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/6cd3f2cb-45d7-4938-9b7f-bcaaafbfe12f

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Simplify the default search flow and move power-user controls behind …
dd71357 
…Advanced

Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/6cd3f2cb-45d7-4938-9b7f-bcaaafbfe12f

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Add AI suspicion review panel
6edd4bd 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/63cdc1b1-d6ec-45bc-9cb6-304f7c230e30

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Document AI review heuristics
1857d55 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/63cdc1b1-d6ec-45bc-9cb6-304f7c230e30

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Name hi-res AI review threshold
60c2499 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/63cdc1b1-d6ec-45bc-9cb6-304f7c230e30

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Refine AI square-ratio heuristic
d05124a 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/63cdc1b1-d6ec-45bc-9cb6-304f7c230e30

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Add engine relay autopsy and README refresh
b5551c8 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/419821fd-96da-4593-99b0-98b32ee0cc59

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Polish relay review followups
5db50cb 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/419821fd-96da-4593-99b0-98b32ee0cc59

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Add contradiction panel and batch summaries
e8e950d 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/9f2fb05c-7bc4-41b5-9e84-d5dfb49bffb6

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Polish contradiction review followups
4f34355 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/9f2fb05c-7bc4-41b5-9e84-d5dfb49bffb6

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Add document image mode and query generator
29c39d1 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/e1b7b5f9-72f5-432a-a7be-edccc449d31d

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
Add helper coverage for document workflows
34abf0f 
Agent-Logs-Url: https://github.com/potemkin666/bluelens/sessions/e1b7b5f9-72f5-432a-a7be-edccc449d31d

Co-authored-by: potemkin666 <183807833+potemkin666@users.noreply.github.com>
@potemkin666
potemkin666 marked this pull request as ready for review May 5, 2026 18:37
Copilot AI review requested due to automatic review settings May 5, 2026 18:37
@potemkin666
potemkin666 merged commit c5283c4 into main May 5, 2026
1 of 3 checks passed
@potemkin666
potemkin666 deleted the copilot/remove-image-and-workflow branch May 5, 2026 18:38
Copilot AI reviewed May 5, 2026
View reviewed changes

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR expands BlueLens’s OCR capabilities by adding a document-image review workflow and a clue-driven search query generator, while also broadening the multi-engine relay and updating the UI/export surfaces to keep these new outputs visible and exportable.

Changes:

  • Add document_image and query_generator workflows with structured OCR/layout output and generated follow-up searches.
  • Expand reverse-search engine coverage (incl. Pinterest/SauceNAO/IQDB/Baidu/ASCII2D) across UI, config, OSINT lib, and doctor checks.
  • Enhance investigation UX and exports (deduped intake labeling/suppression, no-result autopsy, contradiction surfacing, batch CSV + enriched JSON export).

Reviewed changes

Copilot reviewed 11 out of 12 changed files in this pull request and generated 3 comments.

Show a summary per file
File Description
test/ui-regressions.test.js Updates UI regression assertions for new workflows, panels, engine set, and copy changes.
test/osint.test.js Adds test coverage for new reverse-search engine URL builders.
test/helpers.test.js Adds tests for new document/layout/logo/query helper functions.
styles.css Adds styling for quick/advanced action groups, result suppression, autopsy/contradiction cards, and AI suspicion panel.
server.js Expands doctor engine availability checks to the broader engine list.
README.md Rewrites README with new workflow framing, expanded engine list, and updated run/test guidance.
osint-lib.js Adds reverse-search URL and upload-page mappings for additional engines.
index.html Adds quick actions + advanced drawer, new OCR panels (document/query), new engine buttons, and other UI restructuring.
bluelens-helpers.js Introduces shared document/layout inference and query-spec composition helpers for browser + tests.
bluelens-config.js Updates default engine order/labels/icons to include the new engines.
assets/bluelens-ocean-banner.svg Adds a new README banner asset.
app.js Implements document-mode output, query generator output, intake suppression/autopsy/contradictions, AI suspicion checklist, engine wiring, and batch CSV export.

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

Comment thread index.html
Comment on lines +198 to +203
<img
class="table-art"
src="https://github.com/user-attachments/assets/59733571-36f8-4cec-b18c-52e88b2fb9ea"
alt="BlueLens visual reference"
title="External GitHub attachment retained as a visual reference"
loading="lazy"
Comment thread app.js
Comment on lines +2474 to +2481
<div class="mission-card">
<div class="mission-card-head">
<div>${escapeHtml(item?.label || "Item")}</div>
<div class="mission-card-meta">${escapeHtml(item?.meta || "")}</div>
</div>
<div class="mission-card-lines">${lines.map((line) => `<div>${escapeHtml(line)}</div>`).join("")}</div>
${links.length ? `<div class="mission-card-links">${links.map((link) => `<a class="chip chip-link" href="${escapeAttr(link.url)}" target="_blank" rel="noreferrer">${escapeHtml(link.label || link.url)}</a>`).join("")}</div>` : ""}
</div>
Comment thread app.js
Comment on lines +2514 to +2533
function parseFilenameStem(name) {
if (appHelpers.parseFilenameStem) return appHelpers.parseFilenameStem(name);
const raw = String(name || "").trim().replace(/\.[^.]+$/, "");
return raw.replace(/[_-]+/g, " ").replace(/\s+/g, " ").trim();
}

function getCurrentReconContext() {
return normalizeReconEntities(state.ocrText || "");
}

function summarizeDocumentLayout(text) {
if (appHelpers.summarizeDocumentLayout) return appHelpers.summarizeDocumentLayout(text);
const lines = String(text || "")
.split(/\r?\n/)
.map((line) => line.trim())
.filter(Boolean);
const paragraphBreaks = String(text || "").split(/\n\s*\n/).filter((block) => block.trim()).length;
const headingCandidates = lines.filter((line) => line.length >= 4 && line.length <= 72 && (line === line.toUpperCase() || /^[A-Z][A-Za-z0-9 '&/-]{3,}$/.test(line))).slice(0, 4);
const kvRows = lines.filter((line) => /[:|]/.test(line) || /\b(total|date|name|address|price|receipt|invoice|menu|certificate|issued|expires)\b/i.test(line)).slice(0, 6);
const tableRows = lines.filter((line) => /\d/.test(line) && /[$€£¥]|(?:\s{2,}|\t)|\bqty\b|\btotal\b|\bsubtotal\b/i.test(line)).slice(0, 6);
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

Copilot code review Copilot Copilot left review comments

Assignees

@potemkin666 potemkin666

Copilot code review Copilot

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@potemkin666