Polish import content review: markdown rendering, enriched meta, complete fixtures

- Render markdown content with marked.js instead of raw <pre> tags
- Add sanitizeMarkdown() to strip dangerous HTML from user-imported content
- Add CSS for .import-content-markdown with constrained heading/list styling
- Strip markdown syntax from preview text in frontend (stripMarkdown helper)
- Enrich ICS importer manifest meta: time_range, attendee_names, recurrence
- Generate plain-text preview from structured data when description is empty
- Add event-specific card display: time range, attendee names, location
- Display tags for Obsidian notes in card meta
- Fix Gemini importer missing message_count in manifest meta
- Strip markdown from Obsidian preview text in importer
- Complete fixture chains for all 6 source types:
- Add import.json + imported.json + content_manifest.jsonl for
Kindle, Obsidian, Claude, and Gemini
- Update ICS manifest with enriched meta fields
- ChatGPT fixtures verified (already correct)

Co-Authored-By: Claude Opus 4.6 <[e-mail]>

apps/import/_detail.html

···
 .import-content-item-body { display: none; padding: 0.75em 1em; border-top: 1px solid #e5e5e5; background: #fafafa; }
 .import-content-item.expanded .import-content-item-body { display: block; }
 .import-content-item-body pre { white-space: pre-wrap; font-size: 0.9em; margin: 0; }
+.import-content-markdown { font-size: 0.95em; line-height: 1.6; }
+.import-content-markdown h2 { font-size: 1.1em; margin: 0 0 0.5em 0; }
+.import-content-markdown h3 { font-size: 1em; margin: 0.75em 0 0.25em 0; }
+.import-content-markdown ul, .import-content-markdown ol { margin: 0.25em 0; padding-left: 1.5em; }
+.import-content-markdown li { margin: 0.15em 0; }
+.import-content-markdown blockquote { border-left: 3px solid #ddd; margin: 0.5em 0; padding: 0.25em 0.75em; color: #555; }
+.import-content-markdown p { margin: 0.4em 0; }
+.import-content-markdown hr { border: none; border-top: 1px solid #e5e5e5; margin: 1em 0; }
+.import-content-markdown code { background: #f0f0f0; padding: 1px 4px; border-radius: 3px; font-size: 0.9em; }
+.import-content-markdown pre code { background: none; padding: 0; }
 .import-content-item-body .import-content-message { margin: 0.5em 0; }
 .import-content-item-body .import-content-speaker { font-weight: 600; margin-right: 0.5em; }
 .import-content-item-body .import-content-speaker.human { color: #007bff; }
···
 
 </div>
 
+<script src="{{ vendor_lib('marked') }}"></script>
 <script>
+// Configure marked for safe rendering of imported content
+marked.setOptions({ breaks: true, gfm: true });
+const markedRenderer = new marked.Renderer();
+// Sanitize: strip raw HTML tags from markdown input to prevent XSS from user-imported content
+function sanitizeMarkdown(md) {
+  return md.replace(/<(script|iframe|object|embed|form|input|textarea|button|select|style|link|meta)[^>]*>[\s\S]*?<\/\1>/gi, '')
+    .replace(/<(script|iframe|object|embed|form|input|textarea|button|select|style|link|meta)[^>]*\/?>/gi, '')
+    .replace(/\bon\w+\s*=\s*["'][^"']*["']/gi, '')
+    .replace(/javascript\s*:/gi, 'blocked:');
+}
+
 // Tab switching
 const tabs = document.querySelectorAll('.tab');
 const contents = document.querySelectorAll('.tab-content');
···
         <div class="import-content-item-meta">
           ${item.meta?.message_count ? `${item.meta.message_count} messages` : ''}
           ${item.meta?.highlight_count ? `${item.meta.highlight_count} highlights` : ''}
-          ${item.meta?.duration_minutes ? `${item.meta.duration_minutes} min` : ''}
+          ${item.meta?.time_range ? `${escapeContentHtml(item.meta.time_range)}` : (item.meta?.duration_minutes ? `${item.meta.duration_minutes} min` : '')}
+          ${item.meta?.attendee_names?.length ? ` · ${item.meta.attendee_names.map(n => escapeContentHtml(n)).join(', ')}` : (item.meta?.attendee_count ? ` · ${item.meta.attendee_count} attendees` : '')}
+          ${item.meta?.location ? ` · ${escapeContentHtml(item.meta.location)}` : ''}
           ${item.meta?.author ? `· ${escapeContentHtml(item.meta.author)}` : ''}
+          ${item.meta?.tags?.length ? `· ${item.meta.tags.map(t => escapeContentHtml(t)).join(', ')}` : ''}
           ${calLink ? ` · <a href="${calLink}" class="import-content-cal-link" onclick="event.stopPropagation();">View in calendar →</a>` : ''}
         </div>
-        ${item.preview ? `<div class="import-content-item-preview">${escapeContentHtml(item.preview)}</div>` : ''}
+        ${item.preview ? `<div class="import-content-item-preview">${escapeContentHtml(stripMarkdown(item.preview))}</div>` : ''}
         <div class="import-content-item-body"><div class="no-data">Loading...</div></div>
       </div>`;
   });
···
 
   return data.content.map(part => {
     if (part.type === 'markdown') {
-      return `<pre>${escapeContentHtml(part.content)}</pre>`;
+      return `<div class="import-content-markdown">${marked.parse(sanitizeMarkdown(part.content))}</div>`;
     }
     const speaker = part.speaker || 'Unknown';
     const cls = speaker === 'Human' ? 'human' : 'assistant';
···
     .replace(/</g, '&lt;')
     .replace(/>/g, '&gt;')
     .replace(/"/g, '&quot;');
+}
+
+function stripMarkdown(str) {
+  if (!str) return '';
+  return str
+    .replace(/^#{1,6}\s+/gm, '')
+    .replace(/\*\*([^*]+)\*\*/g, '$1')
+    .replace(/\*([^*]+)\*/g, '$1')
+    .replace(/^[-*+]\s+/gm, '')
+    .replace(/^\d+\.\s+/gm, '')
+    .replace(/\[([^\]]+)\]\([^)]+\)/g, '$1')
+    .replace(/`([^`]+)`/g, '$1')
+    .replace(/^>\s+/gm, '')
+    .trim();
 }
 
 document.getElementById('contentSearch').addEventListener('input', function() {

tests/fixtures/journal/imports/20260101_090000/content_manifest.jsonl

···
-{"id": "event-0", "title": "Weekly Engineering Standup", "date": "20260101", "type": "event", "preview": "- **When:** Thursday, January 1, 2026 09:00–09:30 MST", "meta": {}, "segments": [{"day": "20260101", "key": "090000_300"}]}
-{"id": "event-1", "title": "1:1 with David Park (Betaworks)", "date": "20260101", "type": "event", "preview": "- **When:** Thursday, January 1, 2026 10:00–10:45 MST", "meta": {}, "segments": [{"day": "20260101", "key": "090000_300"}]}
-{"id": "event-2", "title": "Product Review — Solstone v0.4 Roadmap", "date": "20260101", "type": "event", "preview": "- **When:** Thursday, January 1, 2026 11:00–12:00 MST", "meta": {}, "segments": [{"day": "20260101", "key": "090000_300"}]}
-{"id": "event-3", "title": "Lunch with Sarah Chen", "date": "20260101", "type": "event", "preview": "- **When:** Thursday, January 1, 2026 12:00–13:00 MST", "meta": {}, "segments": [{"day": "20260101", "key": "093000_300"}]}
-{"id": "event-4", "title": "Demo Prep — Betaworks Studio Day", "date": "20260101", "type": "event", "preview": "- **When:** Thursday, January 1, 2026 14:00–15:00 MST", "meta": {}, "segments": [{"day": "20260101", "key": "093000_300"}]}
+{"id": "event-0", "title": "Weekly Engineering Standup", "date": "20260101", "type": "event", "preview": "9:00 AM–9:30 AM · Google Meet · jer, maya, erik, carlos · Weekly on Thu", "meta": {"time_range": "9:00 AM–9:30 AM", "location": "Google Meet (meet.google.com/xwr-knvq-paj)", "duration_minutes": 30, "attendee_count": 4, "attendee_names": ["jer", "maya", "erik", "carlos"], "recurrence": "Weekly on Thu"}, "segments": [{"day": "20260101", "key": "090000_300"}]}
+{"id": "event-1", "title": "1:1 with David Park (Betaworks)", "date": "20260101", "type": "event", "preview": "Quarterly check-in. David wants to discuss follow-on timing and intro to a potential design partner in the media space.", "meta": {"time_range": "10:00 AM–10:45 AM", "location": "Zoom (us06web.zoom.us/j/84729163055)", "duration_minutes": 45, "attendee_count": 2, "attendee_names": ["[e-mail]", "[e-mail]"]}, "segments": [{"day": "20260101", "key": "090000_300"}]}
+{"id": "event-2", "title": "Product Review — Solstone v0.4 Roadmap", "date": "20260101", "type": "event", "preview": "Walk through v0.4 milestone plan. Key topics: journal segment format finalization, import pipeline reliability, and facet extraction accuracy targets.", "meta": {"time_range": "11:00 AM–12:00 PM", "location": "Google Meet (meet.google.com/bmf-tsyh-rea)", "duration_minutes": 60, "attendee_count": 4, "attendee_names": ["jer", "maya", "erik", "lin"]}, "segments": [{"day": "20260101", "key": "090000_300"}]}
+{"id": "event-3", "title": "Lunch with Sarah Chen", "date": "20260101", "type": "event", "preview": "Catch up with Sarah on outstanding items. Review updated operating agreement language around benefit purpose reporting obligations.", "meta": {"time_range": "12:00 PM–1:00 PM", "location": "The Kitchen, 1530 16th St, Denver, CO 80202", "duration_minutes": 60, "attendee_count": 2, "attendee_names": ["[e-mail]", "[e-mail]"]}, "segments": [{"day": "20260101", "key": "093000_300"}]}
+{"id": "event-4", "title": "Demo Prep — Betaworks Studio Day", "date": "20260101", "type": "event", "preview": "Dry run for the Betaworks Studio Day demo on Jan 15.", "meta": {"time_range": "2:00 PM–3:00 PM", "location": "Google Meet (meet.google.com/qjz-wmvx-dck)", "duration_minutes": 60, "attendee_count": 3, "attendee_names": ["jer", "maya", "erik"]}, "segments": [{"day": "20260101", "key": "093000_300"}]}

tests/fixtures/journal/imports/20260101_110000/content_manifest.jsonl

···
+{"id": "activity-0", "title": "Kubernetes vs Nomad for small teams", "date": "20260101", "type": "conversation", "preview": "We're a small team at Meridian Labs, about six engineers, and we're still running docker-compose in production. It's starting to fall apart as we add more services.", "meta": {"message_count": 8}, "segments": [{"day": "20260101", "key": "110000_300"}, {"day": "20260101", "key": "110500_300"}]}

tests/fixtures/journal/imports/20260101_110000/import.json

···
+{
+  "original_filename": "My Activity.zip",
+  "upload_timestamp": 1767265200000,
+  "upload_datetime": "2026-01-01T11:00:00",
+  "user_timestamp": "20260101_110000",
+  "file_size": 52340,
+  "mime_type": "application/zip",
+  "facet": "work",
+  "setting": "ai_conversation",
+  "file_path": "imports/20260101_110000/My Activity.zip"
+}

tests/fixtures/journal/imports/20260101_110000/imported.json

···
+{
+  "processed_timestamp": "20260101_110000",
+  "processing_completed": "2026-01-01T11:03:00",
+  "total_files_created": 2,
+  "all_created_files": [
+    "20260101/import.gemini/110000_300/conversation_transcript.jsonl",
+    "20260101/import.gemini/110500_300/conversation_transcript.jsonl"
+  ],
+  "segments": [
+    "110000_300",
+    "110500_300"
+  ],
+  "source_type": "gemini",
+  "source_display": "Gemini",
+  "entries_written": 8,
+  "entities_seeded": 0,
+  "date_range": [
+    "20260101",
+    "20260101"
+  ],
+  "target_day": "20260101"
+}

tests/fixtures/journal/imports/20260101_140000/content_manifest.jsonl

···
+{"id": "conv-0", "title": "Event sourcing trade-offs for Solstone", "date": "20260101", "type": "conversation", "preview": "I'm evaluating whether to store domain events in Postgres directly or use a dedicated event store like EventStoreDB. What are the main trade-offs?", "meta": {"message_count": 15}, "segments": [{"day": "20260101", "key": "140000_300"}, {"day": "20260101", "key": "140500_300"}, {"day": "20260101", "key": "141000_300"}]}

tests/fixtures/journal/imports/20260101_140000/import.json

···
+{
+  "original_filename": "claude-conversations.json",
+  "upload_timestamp": 1767276000000,
+  "upload_datetime": "2026-01-01T14:00:00",
+  "user_timestamp": "20260101_140000",
+  "file_size": 67890,
+  "mime_type": "application/json",
+  "facet": "work",
+  "setting": "ai_conversation",
+  "file_path": "imports/20260101_140000/claude-conversations.json"
+}

tests/fixtures/journal/imports/20260101_140000/imported.json

···
+{
+  "processed_timestamp": "20260101_140000",
+  "processing_completed": "2026-01-01T14:04:00",
+  "total_files_created": 3,
+  "all_created_files": [
+    "20260101/import.claude/140000_300/conversation_transcript.jsonl",
+    "20260101/import.claude/140500_300/conversation_transcript.jsonl",
+    "20260101/import.claude/141000_300/conversation_transcript.jsonl"
+  ],
+  "segments": [
+    "140000_300",
+    "140500_300",
+    "141000_300"
+  ],
+  "source_type": "claude",
+  "source_display": "Claude Chat",
+  "entries_written": 15,
+  "entities_seeded": 0,
+  "date_range": [
+    "20260101",
+    "20260101"
+  ],
+  "target_day": "20260101"
+}

tests/fixtures/journal/imports/20260101_160000/content_manifest.jsonl

···
+{"id": "note-0", "title": "Trust Architecture for Decentralized Identity", "date": "20260101", "type": "note", "preview": "Sketching out the trust layer for Solstone. The core idea is that identity verification shouldn't depend on a single authority — it needs to be distributed and composable.", "meta": {"tags": ["trust", "identity", "solstone", "architecture"]}, "segments": [{"day": "20260101", "key": "160000_300"}]}
+{"id": "note-1", "title": "Book Rec from Nadia", "date": "20260101", "type": "note", "preview": "Nadia recommended Governing the Commons by Elinor Ostrom over coffee today. Relevant to what we're building with Solstone.", "meta": {"tags": ["reading", "governance", "commons"]}, "segments": [{"day": "20260101", "key": "160000_300"}]}
+{"id": "note-2", "title": "Betaworks Call Prep", "date": "20260101", "type": "note", "preview": "Meeting with David Park from Betaworks on Thursday. Need to cover key points about composable trust signals, the protocol layer, and identity providers.", "meta": {"tags": ["meetings", "betaworks", "partnerships"]}, "segments": [{"day": "20260101", "key": "160500_300"}]}
+{"id": "note-3", "title": "Stream Processing Patterns", "date": "20260101", "type": "note", "preview": "Documenting the patterns we keep reaching for. These are emerging from actual implementation, not theoretical.", "meta": {"tags": ["engineering", "streams", "architecture", "solstone"]}, "segments": [{"day": "20260101", "key": "160500_300"}]}

tests/fixtures/journal/imports/20260101_160000/import.json

···
+{
+  "original_filename": "vault-export.zip",
+  "upload_timestamp": 1767283200000,
+  "upload_datetime": "2026-01-01T16:00:00",
+  "user_timestamp": "20260101_160000",
+  "file_size": 89012,
+  "mime_type": "application/zip",
+  "facet": "work",
+  "setting": "obsidian",
+  "file_path": "imports/20260101_160000/vault-export.zip"
+}

tests/fixtures/journal/imports/20260101_160000/imported.json

···
+{
+  "processed_timestamp": "20260101_160000",
+  "processing_completed": "2026-01-01T16:03:00",
+  "total_files_created": 2,
+  "all_created_files": [
+    "20260101/import.obsidian/160000_300/note_transcript.md",
+    "20260101/import.obsidian/160500_300/note_transcript.md"
+  ],
+  "segments": [
+    "160000_300",
+    "160500_300"
+  ],
+  "source_type": "obsidian",
+  "source_display": "Obsidian / Logseq Vault",
+  "entries_written": 4,
+  "entities_seeded": 7,
+  "date_range": [
+    "20260101",
+    "20260101"
+  ],
+  "target_day": "20260101"
+}

tests/fixtures/journal/imports/20260101_200000/content_manifest.jsonl

···
+{"id": "book-0", "title": "The Art of Doing Science and Engineering by Richard Hamming", "date": "20260101", "type": "highlight_group", "preview": "Most people spend their careers working on problems that don't really matter. They choose safe, incremental topics where they can guarantee results rather than risking failure on something significant.", "meta": {"author": "Richard Hamming", "highlight_count": 5}, "segments": [{"day": "20260101", "key": "200000_300"}]}
+{"id": "book-1", "title": "Governing the Commons by Elinor Ostrom", "date": "20260101", "type": "highlight_group", "preview": "Communities that have successfully managed shared resources over long periods almost never did so by following the prescriptions of outside experts.", "meta": {"author": "Elinor Ostrom", "highlight_count": 4}, "segments": [{"day": "20260101", "key": "200500_300"}]}

tests/fixtures/journal/imports/20260101_200000/import.json

···
+{
+  "original_filename": "My Clippings.txt",
+  "upload_timestamp": 1767297600000,
+  "upload_datetime": "2026-01-01T20:00:00",
+  "user_timestamp": "20260101_200000",
+  "file_size": 34521,
+  "mime_type": "text/plain",
+  "facet": "reading",
+  "setting": "kindle",
+  "file_path": "imports/20260101_200000/My Clippings.txt"
+}

tests/fixtures/journal/imports/20260101_200000/imported.json

···
+{
+  "processed_timestamp": "20260101_200000",
+  "processing_completed": "2026-01-01T20:02:00",
+  "total_files_created": 2,
+  "all_created_files": [
+    "20260101/import.kindle/200000_300/highlights_transcript.md",
+    "20260101/import.kindle/200500_300/highlights_transcript.md"
+  ],
+  "segments": [
+    "200000_300",
+    "200500_300"
+  ],
+  "source_type": "kindle",
+  "source_display": "Kindle Highlights",
+  "entries_written": 9,
+  "entities_seeded": 4,
+  "date_range": [
+    "20260101",
+    "20260101"
+  ],
+  "target_day": "20260101"
+}

think/importers/gemini.py

···
                 "title": prompt[:80] or f"Activity {valid_idx + 1}",
                 "first_ts": activity_messages[0]["create_time"],
                 "preview": prompt,
+                "message_count": len(activity_messages),
             }
             valid_idx += 1
 
···
                     "date": first_dt.strftime("%Y%m%d"),
                     "type": "conversation",
                     "preview": meta["preview"],
-                    "meta": {},
+                    "meta": {"message_count": meta.get("message_count", 0)},
                     "segments": [
                         {"day": day, "key": key}
                         for day, key in sorted(

think/importers/ics.py

···
                 entry["create_ts"], tz=dt.timezone.utc
             )
             meta: dict[str, Any] = {}
+            if entry.get("ts") and entry.get("end_ts"):
+                try:
+                    start_dt = dt.datetime.fromisoformat(entry["ts"])
+                    end_dt = dt.datetime.fromisoformat(entry["end_ts"])
+                    meta["time_range"] = (
+                        f"{start_dt.strftime('%I:%M %p').lstrip('0')}"
+                        f"–{end_dt.strftime('%I:%M %p').lstrip('0')}"
+                    )
+                except ValueError:
+                    pass
             if entry.get("location"):
                 meta["location"] = entry["location"]
             if entry.get("duration_minutes") is not None:
                 meta["duration_minutes"] = entry["duration_minutes"]
             if entry.get("attendees"):
                 meta["attendee_count"] = len(entry["attendees"])
+                meta["attendee_names"] = [
+                    a.get("name") or a.get("email", "")
+                    for a in entry["attendees"][:5]
+                ]
+            if entry.get("recurrence"):
+                meta["recurrence"] = entry["recurrence"]
+
+            # Build plain-text preview from structured data if description is empty
+            preview = entry.get("content", "").strip()
+            if not preview:
+                parts: list[str] = []
+                if meta.get("time_range"):
+                    parts.append(meta["time_range"])
+                if entry.get("location"):
+                    parts.append(entry["location"])
+                if meta.get("attendee_names"):
+                    parts.append(", ".join(meta["attendee_names"]))
+                if entry.get("recurrence"):
+                    parts.append(entry["recurrence"])
+                preview = " · ".join(parts)
             manifest_entries.append(
                 {
                     "id": f"event-{i}",
                     "title": entry.get("title", "Untitled event"),
                     "date": create_dt.strftime("%Y%m%d"),
                     "type": "event",
-                    "preview": entry.get("content", "")[:200],
+                    "preview": preview[:200],
                     "meta": meta,
                     "segments": [],
                 }

think/importers/obsidian.py

···
                 meta["tags"] = note["tags"]
             if note.get("is_daily"):
                 meta["is_daily"] = True
+            raw_preview = _strip_frontmatter(note.get("content", "")).strip()[:300]
+            # Strip markdown syntax for clean plain-text preview
+            clean_preview = re.sub(r"^#{1,6}\s+", "", raw_preview, flags=re.MULTILINE)
+            clean_preview = re.sub(r"\*\*([^*]+)\*\*", r"\1", clean_preview)
+            clean_preview = re.sub(r"\*([^*]+)\*", r"\1", clean_preview)
+            clean_preview = re.sub(r"^[-*+]\s+", "", clean_preview, flags=re.MULTILINE)
+            clean_preview = re.sub(
+                r"\[\[([^\]|]+)(?:\|([^\]]+))?\]\]",
+                lambda m: m.group(2) or m.group(1),
+                clean_preview,
+            )
+            clean_preview = re.sub(r"\[([^\]]+)\]\([^)]+\)", r"\1", clean_preview)
+            clean_preview = re.sub(r"`([^`]+)`", r"\1", clean_preview)
+            clean_preview = re.sub(r"^>\s+", "", clean_preview, flags=re.MULTILINE)
+            clean_preview = " ".join(clean_preview.split())[:200]
             note_manifest.append(
                 {
                     "id": f"note-{i}",
                     "title": note["title"],
                     "date": note_dt.strftime("%Y%m%d"),
                     "type": "note",
-                    "preview": _strip_frontmatter(note.get("content", "")).strip()[
-                        :200
-                    ],
+                    "preview": clean_preview,
                     "meta": meta,
                     "segments": [],
                 }