Add document analysis: page count, text extraction & language detection

Introduces a document analyzer service that extracts page count (via pdf-parse),
text preview, and detected language (via franc) from uploaded files. Analysis runs
automatically on upload (configurable via SystemSettings) and can be triggered
retroactively for existing files. Results are displayed as badges in the FileViewer
and fed to AI screening for language-based filtering criteria.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

src/server/services/anonymization.ts

@@ -83,6 +83,8 @@ export interface AnonymizedFileInfo {
   file_type: string // FileType enum value
   page_count: number | null // Number of pages if known
   size_kb: number // File size in KB
+  detected_lang?: string | null // ISO 639-3 language code (e.g. 'eng', 'fra')
+  lang_confidence?: number | null // 0.0–1.0 confidence score
   round_name?: string | null // Which round the file was submitted for
   is_current_round?: boolean // Whether this file belongs to the current filtering/evaluation round
   text_content?: string // Extracted text content (when aiParseFiles is enabled)
@@ -309,6 +311,8 @@ export function anonymizeProjectForAI(
       file_type: f.fileType ?? 'OTHER',
       page_count: f.pageCount ?? null,
       size_kb: Math.round((f.size ?? 0) / 1024),
+      ...(f.detectedLang ? { detected_lang: f.detectedLang } : {}),
+      ...(f.langConfidence != null ? { lang_confidence: f.langConfidence } : {}),
       ...(f.roundName ? { round_name: f.roundName } : {}),
       ...(f.isCurrentRound !== undefined ? { is_current_round: f.isCurrentRound } : {}),
       ...(f.textContent ? { text_content: f.textContent } : {}),