Akiba So
e939759458
All checks were successful
CI / Windows build (push) Successful in 15m50s
Search now covers handwriting, the PDF text layer, AND scanned (rasterized) PDFs. - PdfTextIndexer runs at import: sums the embedded text layer across pages; if present it stores that as the document body, otherwise the PDF is rasterized and its rendered pages are OCR'd in the background. The result lands in the sidecar `pageText` field (distinct from `ocrText`, the handwriting OCR). Idempotent (skips a sidecar that already has pageText); degrades gracefully with no OCR engine. - pdfrx_page_text_source abstracts text/render so it's testable. - VaultSearchIndex now harvests title + typed text + handwriting OCR + PDF pageText, so search finds notes, typed PDFs and scanned PDFs. analyze clean, 409 tests green.
57 lines
2.3 KiB
Dart
57 lines
2.3 KiB
Dart
import 'package:flutter_riverpod/flutter_riverpod.dart';
|
|
|
|
import '../services/ocr_service.dart';
|
|
import '../services/pdf_text_indexer.dart';
|
|
import '../services/pdfrx_page_text_source.dart';
|
|
|
|
enum OcrStatus { none, processing, done, failed }
|
|
|
|
final ocrServiceProvider = Provider<OcrService>((ref) => OcrService());
|
|
|
|
/// The import-time PDF document-body indexer, wired to the pdfrx-backed embedded
|
|
/// text + page-render OCR sources (see [PdfrxPageTextSource]). The import flow
|
|
/// fires [PdfTextIndexer.indexPdf] (fire-and-forget) so a scanned PDF's text
|
|
/// becomes searchable in the background without blocking the editor opening.
|
|
final pdfTextIndexerProvider = Provider<PdfTextIndexer>(
|
|
(ref) => PdfTextIndexer(
|
|
loadEmbeddedText: PdfrxPageTextSource.loadEmbeddedText,
|
|
ocrPages: PdfrxPageTextSource.ocrPages,
|
|
),
|
|
);
|
|
|
|
/// Tracks local OCR processing status per note ID.
|
|
///
|
|
/// This map only ever holds an entry per note that has had OCR triggered in
|
|
/// the current session. To keep it from growing without bound over a long
|
|
/// session, prune terminal/stale entries via [OcrStatusX] (e.g. remove an
|
|
/// entry once its result has been surfaced, or call [OcrStatusX.pruneOcr]
|
|
/// after a sweep). Kept as a [StateProvider] so existing call sites that
|
|
/// assign `ocrStatusProvider.notifier.state` continue to work.
|
|
final ocrStatusProvider = StateProvider<Map<String, OcrStatus>>((ref) => {});
|
|
|
|
/// Pruning helpers for [ocrStatusProvider] that keep its backing map bounded.
|
|
extension OcrStatusX on Ref {
|
|
/// Removes the tracked status for [noteId] (e.g. when its note is deleted
|
|
/// or its result has been consumed by the UI).
|
|
void clearOcr(String noteId) {
|
|
final current = read(ocrStatusProvider);
|
|
if (!current.containsKey(noteId)) return;
|
|
read(ocrStatusProvider.notifier).state = Map<String, OcrStatus>.from(
|
|
current,
|
|
)..remove(noteId);
|
|
}
|
|
|
|
/// Drops all completed/failed entries, keeping only in-flight work so the
|
|
/// map stays bounded.
|
|
void pruneOcr() {
|
|
final current = read(ocrStatusProvider);
|
|
final next = <String, OcrStatus>{
|
|
for (final entry in current.entries)
|
|
if (entry.value == OcrStatus.processing) entry.key: entry.value,
|
|
};
|
|
if (next.length != current.length) {
|
|
read(ocrStatusProvider.notifier).state = next;
|
|
}
|
|
}
|
|
}
|