Compare commits

...
Author SHA1 Message Date
Trenton HolmesandClaude Sonnet 5 af7a1fcdba Simplify: dedupe row-copy/migration-test helpers, type fixture params
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-26 21:25:25 -07:00
Trenton HolmesandClaude Sonnet 5 3e6579a4ec Perf: point-delete vector store chunks by id instead of a document_id scan
vec0 only gets an efficient lookup on a metadata column inside a KNN
query; a plain document_id-filtered DELETE is a full table scan
regardless of index size. Add a document_chunks side table (plain
SQLite, real index) to look up chunk ids for a document, then delete
each by its id primary key instead. Includes a v1 -> v2 migration to
backfill document_chunks for indexes created before this existed.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-26 20:31:15 -07:00
GitHub Actions f4e7a2e9e4 Auto translate strings 2026-07-26 22:38:16 +00:00
shamoonandGitHub 317e534aa0 Fix: ensure preview reload on live changes (#13321) 2026-07-26 22:36:03 +00:00
Trenton HandGitHub 3ba6d0325a Fix: clamp out-of-range MailRule.order and ApplicationConfiguration DPI/page fields before smallint migration (#13316) 2026-07-26 22:11:27 +00:00
Trenton HandGitHub 318520a0f1 Fix: add docstring to DocumentClassifierSchema for cleaner LLM tool description (#13315) 2026-07-26 22:00:06 +00:00
Trenton HandGitHub ea1f3653a9 Fix: guard build_document_node against stale FK on deleted correspondent/doc type (#13318) 2026-07-26 21:45:59 +00:00
shamoonandGitHub 423d4b9f2d Fix: prevent search filter loss when closing document with Escape key (#13317) 2026-07-26 14:17:26 -07:00
shamoonandGitHub 8bc4e5ee94 Fix: fix frontend permissions display for stats/system perms (#13305) 2026-07-26 02:56:06 -07:00
19 changed files with 722 additions and 292 deletions
@@ -1,4 +1,4 @@
import { expect, test } from '@playwright/test'
import { expect, test, type WebSocketRoute } from '@playwright/test'
import path from 'node:path'
const REQUESTS_HAR = path.join(__dirname, 'requests/api-document-detail.har')
@@ -95,3 +95,60 @@ test('should support quick filters', async ({ page }) => {
.click()
await expect(page).toHaveURL(/tags__id__all=4&sort=created&reverse=1&page=1/)
})
test('should finish reloading the preview after a remote document update', async ({
page,
}) => {
let resolveStatusSocket: (socket: WebSocketRoute) => void
const statusSocketReady = new Promise<WebSocketRoute>((resolve) => {
resolveStatusSocket = resolve
})
await page.routeWebSocket(/\/ws\/status\/$/, (socket) => {
resolveStatusSocket(socket)
})
await page.routeFromHAR(REQUESTS_HAR, { notFound: 'fallback' })
let previewRequestCount = 0
page.on('request', (request) => {
if (request.url().includes('/api/documents/175/preview/')) {
previewRequestCount++
}
})
await page.goto('/documents/175/details')
await page.locator('pngx-document-detail').waitFor()
await expect(page.getByTitle('Storage path', { exact: true })).toHaveText(
/\w+/
)
const previewWasLoaded = await page.evaluate(() => {
const detail = document.querySelector('pngx-document-detail')
const component = (window as any).ng.getComponent(detail)
component.pdfPreviewLoaded({ numPages: 1 })
return component.previewLoaded()
})
expect(previewWasLoaded).toBe(true)
const previewRequestsBeforeReload = previewRequestCount
const statusSocket = await statusSocketReady
const documentReloaded = page.waitForResponse(
(response) =>
response.url().includes('/api/documents/175/?full_perms=true') &&
response.request().method() === 'GET'
)
statusSocket.send(
JSON.stringify({
type: 'document_updated',
data: {
document_id: 175,
modified: '2026-07-26T20:00:00Z',
},
})
)
await documentReloaded
await expect(
page.getByText('Document reloaded with latest changes.').first()
).toBeVisible()
await expect
.poll(() => previewRequestCount)
.toBeGreaterThan(previewRequestsBeforeReload + 1)
})
+49 -49
View File
@@ -1277,7 +1277,7 @@
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1787</context>
<context context-type="linenumber">1791</context>
</context-group>
</trans-unit>
<trans-unit id="1577733187050997705" datatype="html">
@@ -2184,7 +2184,7 @@
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">653</context>
<context context-type="linenumber">657</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-version-dropdown/document-version-dropdown.component.html</context>
@@ -3087,11 +3087,11 @@
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1405</context>
<context context-type="linenumber">1409</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1788</context>
<context context-type="linenumber">1792</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-list/bulk-editor/bulk-editor.component.ts</context>
@@ -3688,7 +3688,7 @@
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1358</context>
<context context-type="linenumber">1362</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-list/bulk-editor/bulk-editor.component.ts</context>
@@ -3800,7 +3800,7 @@
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1841</context>
<context context-type="linenumber">1845</context>
</context-group>
</trans-unit>
<trans-unit id="6661109599266152398" datatype="html">
@@ -3811,7 +3811,7 @@
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1842</context>
<context context-type="linenumber">1846</context>
</context-group>
</trans-unit>
<trans-unit id="5162686434580248853" datatype="html">
@@ -3822,7 +3822,7 @@
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1843</context>
<context context-type="linenumber">1847</context>
</context-group>
</trans-unit>
<trans-unit id="8157388568390631653" datatype="html">
@@ -5724,7 +5724,7 @@
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1362</context>
<context context-type="linenumber">1366</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-list/bulk-editor/bulk-editor.component.ts</context>
@@ -7966,88 +7966,88 @@
<source>Enter Password</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.html</context>
<context context-type="linenumber">512</context>
<context context-type="linenumber">513</context>
</context-group>
</trans-unit>
<trans-unit id="5758784066858623886" datatype="html">
<source>Error retrieving metadata</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">423</context>
<context context-type="linenumber">424</context>
</context-group>
</trans-unit>
<trans-unit id="2218903673684131427" datatype="html">
<source>An error occurred loading content: <x id="PH" equiv-text="err.message ?? err.toString()"/></source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">525,527</context>
<context context-type="linenumber">526,528</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">982,984</context>
<context context-type="linenumber">986,988</context>
</context-group>
</trans-unit>
<trans-unit id="6357361810318120957" datatype="html">
<source>Document was updated</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">648</context>
<context context-type="linenumber">652</context>
</context-group>
</trans-unit>
<trans-unit id="5154064822428631306" datatype="html">
<source>Document was updated at <x id="PH" equiv-text="formattedModified"/>.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">649</context>
<context context-type="linenumber">653</context>
</context-group>
</trans-unit>
<trans-unit id="8462497568316256794" datatype="html">
<source>Reload to discard your local unsaved edits and load the latest remote version.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">650</context>
<context context-type="linenumber">654</context>
</context-group>
</trans-unit>
<trans-unit id="7967484035994732534" datatype="html">
<source>Reload</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">652</context>
<context context-type="linenumber">656</context>
</context-group>
</trans-unit>
<trans-unit id="2907037627372942104" datatype="html">
<source>Document reloaded with latest changes.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">708</context>
<context context-type="linenumber">712</context>
</context-group>
</trans-unit>
<trans-unit id="6435639868943916539" datatype="html">
<source>Document reloaded.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">719</context>
<context context-type="linenumber">723</context>
</context-group>
</trans-unit>
<trans-unit id="6142395741265832184" datatype="html">
<source>Next document</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">821</context>
<context context-type="linenumber">825</context>
</context-group>
</trans-unit>
<trans-unit id="651985345816518480" datatype="html">
<source>Previous document</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">831</context>
<context context-type="linenumber">835</context>
</context-group>
</trans-unit>
<trans-unit id="2885986061416655600" datatype="html">
<source>Close document</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">839</context>
<context context-type="linenumber">843</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/services/open-documents.service.ts</context>
@@ -8058,67 +8058,67 @@
<source>Save document</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">846</context>
<context context-type="linenumber">850</context>
</context-group>
</trans-unit>
<trans-unit id="1784543155727940353" datatype="html">
<source>Save and close / next</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">855</context>
<context context-type="linenumber">859</context>
</context-group>
</trans-unit>
<trans-unit id="7427704425579737895" datatype="html">
<source>Error retrieving version content</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">965</context>
<context context-type="linenumber">969</context>
</context-group>
</trans-unit>
<trans-unit id="3456881259945295697" datatype="html">
<source>Error retrieving suggestions.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1026</context>
<context context-type="linenumber">1030</context>
</context-group>
</trans-unit>
<trans-unit id="2194092841814123758" datatype="html">
<source>Document &quot;<x id="PH" equiv-text="newValues.title"/>&quot; saved successfully.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1238</context>
<context context-type="linenumber">1242</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1265</context>
<context context-type="linenumber">1269</context>
</context-group>
</trans-unit>
<trans-unit id="6626387786259219838" datatype="html">
<source>Error saving document &quot;<x id="PH" equiv-text="this.document().title"/>&quot;</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1271</context>
<context context-type="linenumber">1275</context>
</context-group>
</trans-unit>
<trans-unit id="448882439049417053" datatype="html">
<source>Error saving document</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1326</context>
<context context-type="linenumber">1330</context>
</context-group>
</trans-unit>
<trans-unit id="8410796510716511826" datatype="html">
<source>Do you really want to move the document &quot;<x id="PH" equiv-text="this.document().title"/>&quot; to the trash?</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1359</context>
<context context-type="linenumber">1363</context>
</context-group>
</trans-unit>
<trans-unit id="282586936710748252" datatype="html">
<source>Documents can be restored prior to permanent deletion.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1360</context>
<context context-type="linenumber">1364</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-list/bulk-editor/bulk-editor.component.ts</context>
@@ -8129,14 +8129,14 @@
<source>Error deleting document</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1381</context>
<context context-type="linenumber">1385</context>
</context-group>
</trans-unit>
<trans-unit id="619486176823357521" datatype="html">
<source>Reprocess confirm</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1401</context>
<context context-type="linenumber">1405</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-list/bulk-editor/bulk-editor.component.ts</context>
@@ -8147,102 +8147,102 @@
<source>This operation will permanently recreate the archive file for this document.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1402</context>
<context context-type="linenumber">1406</context>
</context-group>
</trans-unit>
<trans-unit id="302054111564709516" datatype="html">
<source>The archive file will be re-generated with the current settings.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1403</context>
<context context-type="linenumber">1407</context>
</context-group>
</trans-unit>
<trans-unit id="4700389117298802932" datatype="html">
<source>Reprocess operation for &quot;<x id="PH" equiv-text="this.document().title"/>&quot; will begin in the background.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1413</context>
<context context-type="linenumber">1417</context>
</context-group>
</trans-unit>
<trans-unit id="4409560272830824468" datatype="html">
<source>Error executing operation</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1424</context>
<context context-type="linenumber">1428</context>
</context-group>
</trans-unit>
<trans-unit id="6030453331794586802" datatype="html">
<source>Error downloading document</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1487</context>
<context context-type="linenumber">1491</context>
</context-group>
</trans-unit>
<trans-unit id="4458954481601077369" datatype="html">
<source>Page Fit</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1565</context>
<context context-type="linenumber">1569</context>
</context-group>
</trans-unit>
<trans-unit id="4663705961777238777" datatype="html">
<source>PDF edit operation for &quot;<x id="PH" equiv-text="this.document().title"/>&quot; will begin in the background.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1808</context>
<context context-type="linenumber">1812</context>
</context-group>
</trans-unit>
<trans-unit id="9043972994040261999" datatype="html">
<source>Error executing PDF edit operation</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1820</context>
<context context-type="linenumber">1824</context>
</context-group>
</trans-unit>
<trans-unit id="6172690334763056188" datatype="html">
<source>Please enter the current password before attempting to remove it.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1831</context>
<context context-type="linenumber">1835</context>
</context-group>
</trans-unit>
<trans-unit id="968660764814228922" datatype="html">
<source>Password removal operation for &quot;<x id="PH" equiv-text="this.document().title"/>&quot; will begin in the background.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1865</context>
<context context-type="linenumber">1869</context>
</context-group>
</trans-unit>
<trans-unit id="2282118435712883014" datatype="html">
<source>Error executing password removal operation</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1879</context>
<context context-type="linenumber">1883</context>
</context-group>
</trans-unit>
<trans-unit id="3740891324955700797" datatype="html">
<source>Print failed.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1925</context>
<context context-type="linenumber">1929</context>
</context-group>
</trans-unit>
<trans-unit id="6457245677384603573" datatype="html">
<source>Error loading document for printing.</source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">1938</context>
<context context-type="linenumber">1942</context>
</context-group>
</trans-unit>
<trans-unit id="6085793215710522488" datatype="html">
<source>An error occurred loading tiff: <x id="PH" equiv-text="err.toString()"/></source>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">2008</context>
<context context-type="linenumber">2012</context>
</context-group>
<context-group purpose="location">
<context context-type="sourcefile">src/app/components/document-detail/document-detail.component.ts</context>
<context context-type="linenumber">2014</context>
<context context-type="linenumber">2018</context>
</context-group>
</trans-unit>
<trans-unit id="4958946940233632319" datatype="html">
@@ -283,6 +283,22 @@ describe('PngxPdfViewerComponent', () => {
expect(mockViewer.currentPageNumber).toBe(1)
})
it('reloads when the source revision changes', () => {
const resetSpy = jest.spyOn(component as any, 'resetViewerState')
const loadSpy = jest
.spyOn(component as any, 'loadDocument')
.mockImplementation(() => {})
component.src = 'test.pdf'
component.sourceRevision = 1
component.ngOnChanges({
sourceRevision: new SimpleChange(0, 1, false),
})
expect(resetSpy).toHaveBeenCalled()
expect(loadSpy).toHaveBeenCalled()
})
it('applies viewer state after view init when already loaded', () => {
const applySpy = jest.spyOn(component as any, 'applyViewerState')
;(component as any).hasLoaded = true
@@ -43,6 +43,7 @@ export class PngxPdfViewerComponent
private readonly document = inject<Document>(DOCUMENT)
@Input() src!: string
@Input() sourceRevision = 0
@Input() password?: string
@Input() page?: number
@Output() pageChange = new EventEmitter<number>()
@@ -93,7 +94,7 @@ export class PngxPdfViewerComponent
}
ngOnChanges(changes: SimpleChanges): void {
if (changes['src'] || changes['password']) {
if (changes['src'] || changes['sourceRevision'] || changes['password']) {
this.resetViewerState()
if (this.src) {
this.loadDocument()
@@ -472,6 +472,7 @@
<div class="preview-sticky pdf-viewer-container">
<pngx-pdf-viewer
[src]="pdfSource()"
[sourceRevision]="previewRevision()"
[password]="pdfPassword()"
[renderMode]="PdfRenderMode.All"
[page]="previewCurrentPage()" (pageChange)="previewCurrentPage.set($event)"
@@ -1602,6 +1602,7 @@ describe('DocumentDetailComponent', () => {
expect(openDoc.__changedFields).toEqual([])
expect(setDirtySpy).toHaveBeenCalledWith(openDoc, false)
expect(saveSpy).toHaveBeenCalled()
expect(component.previewRevision()).toBe(1)
})
it('should ignore incoming update for a different document id', () => {
@@ -249,6 +249,7 @@ export class DocumentDetailComponent
titleSubject: Subject<string> = new Subject()
readonly previewUrl = signal<string>(undefined)
readonly pdfSource = signal<string>(undefined)
readonly previewRevision = signal(0)
readonly pdfPassword = signal<string>(undefined)
readonly thumbUrl = signal<string>(undefined)
readonly previewText = signal<string>(undefined)
@@ -609,6 +610,9 @@ export class DocumentDetailComponent
.subscribe()
}
this.updateComponent(useDoc)
if (forceRemote) {
this.previewRevision.update((revision) => revision + 1)
}
this.titleSubject
.pipe(
debounceTime(1000),
@@ -25,7 +25,7 @@
<input #textFilterInput class="form-control form-control-sm" type="text"
[disabled]="textFilterModifierIsNull"
[(ngModel)]="textFilter"
(keyup)="textFilterKeyup($event)"
(keydown)="textFilterKeydown($event)"
[ngbTypeahead]="searchAutoComplete"
(selectItem)="itemSelected($event)"
[readonly]="textFilterTarget === 'fulltext-morelike'">
@@ -2180,17 +2180,17 @@ describe('FilterEditorComponent', () => {
it('should support Enter / Esc key on text field', () => {
component.textFilterInput.nativeElement.value = 'foo'
component.textFilterInput.nativeElement.dispatchEvent(
new KeyboardEvent('keyup', { key: 'Enter' })
new KeyboardEvent('keydown', { key: 'Enter' })
)
expect(component.textFilter).toEqual('foo')
component.textFilterInput.nativeElement.value = 'foo bar'
component.textFilterInput.nativeElement.dispatchEvent(
new KeyboardEvent('keyup', { key: 'Escape' })
new KeyboardEvent('keydown', { key: 'Escape' })
)
expect(component.textFilter).toEqual('')
const blurSpy = jest.spyOn(component.textFilterInput.nativeElement, 'blur')
component.textFilterInput.nativeElement.dispatchEvent(
new KeyboardEvent('keyup', { key: 'Escape' })
new KeyboardEvent('keydown', { key: 'Escape' })
)
expect(blurSpy).toHaveBeenCalled()
})
@@ -1312,7 +1312,7 @@ export class FilterEditorComponent
}
}
textFilterKeyup(event: KeyboardEvent) {
textFilterKeydown(event: KeyboardEvent) {
if (event.key == 'Enter') {
const filterString = (
this.textFilterInput.nativeElement as HTMLInputElement
@@ -108,6 +108,18 @@ describe('PermissionsService', () => {
actionKey: 'View', // PermissionAction.View
typeKey: 'Document', // PermissionType.Document
})
expect(
permissionsService.getPermissionKeys('view_global_statistics')
).toEqual({
actionKey: 'View', // PermissionAction.View
typeKey: 'GlobalStatistics', // PermissionType.GlobalStatistics
})
expect(
permissionsService.getPermissionKeys('view_system_monitoring')
).toEqual({
actionKey: 'View', // PermissionAction.View
typeKey: 'SystemMonitoring', // PermissionType.SystemMonitoring
})
})
it('correctly checks explicit global permissions', () => {
+7 -10
View File
@@ -110,19 +110,16 @@ export class PermissionsService {
actionKey: string
typeKey: string
} {
const matches = permissionStr.match(/(.+)_/)
let typeKey
let actionKey
if (matches?.length > 0) {
const action = matches[1]
const actionIndex = Object.values(PermissionAction).indexOf(
action as PermissionAction
)
if (actionIndex > -1) {
actionKey = Object.keys(PermissionAction)[actionIndex]
}
const actionIndex = Object.values(PermissionAction).findIndex((action) =>
permissionStr.startsWith(`${action}_`)
)
if (actionIndex > -1) {
const action = Object.values(PermissionAction)[actionIndex]
actionKey = Object.keys(PermissionAction)[actionIndex]
const typeIndex = Object.values(PermissionType).indexOf(
permissionStr.replace(action, '%s') as PermissionType
permissionStr.replace(`${action}_`, '%s_') as PermissionType
)
if (typeIndex > -1) {
typeKey = Object.keys(PermissionType)[typeIndex]
@@ -5,12 +5,30 @@ from django.db import migrations
from django.db import models
def clamp_applicationconfiguration_integer_fields(apps, schema_editor):
# Clamp barcode_dpi, barcode_max_pages, image_dpi and pages because of
# PositiveIntegerField --> PositiveSmallIntegerField
ApplicationConfiguration = apps.get_model("paperless", "ApplicationConfiguration")
ApplicationConfiguration.objects.filter(barcode_dpi__gt=32767).update(
barcode_dpi=32767,
)
ApplicationConfiguration.objects.filter(barcode_max_pages__gt=32767).update(
barcode_max_pages=32767,
)
ApplicationConfiguration.objects.filter(image_dpi__gt=32767).update(image_dpi=32767)
ApplicationConfiguration.objects.filter(pages__gt=32767).update(pages=32767)
class Migration(migrations.Migration):
dependencies = [
("paperless", "0006_applicationconfiguration_barcode_tag_split"),
]
operations = [
migrations.RunPython(
clamp_applicationconfiguration_integer_fields,
migrations.RunPython.noop,
),
migrations.AlterField(
model_name="applicationconfiguration",
name="barcode_dpi",
+2
View File
@@ -2,6 +2,8 @@ from pydantic import BaseModel
class DocumentClassifierSchema(BaseModel):
"""Schema for document classification suggestions."""
title: str
tags: list[str]
correspondents: list[str]
+17 -7
View File
@@ -5,6 +5,7 @@ from datetime import timedelta
from typing import TYPE_CHECKING
from django.conf import settings
from django.core.exceptions import ObjectDoesNotExist
from django.utils import timezone
from filelock import FileLock
from filelock import ReadWriteLock
@@ -167,6 +168,19 @@ def write_store(embed_model_name: str | None = None):
yield store
def _safe_related_name(document: Document, field: str) -> str | None:
"""
Returns the ``name`` of a related object (correspondent, document_type,
storage_path), or None if the FK is unset or points at a row that has
since been deleted (e.g. concurrently with this call).
"""
try:
related = getattr(document, field)
except ObjectDoesNotExist:
return None
return related.name if related else None
def build_document_node(
document: Document,
*,
@@ -180,14 +194,10 @@ def build_document_node(
"document_id": str(document.id),
"title": document.title,
"tags": [t.name for t in document.tags.all()],
"correspondent": document.correspondent.name
if document.correspondent
else None,
"document_type": document.document_type.name
if document.document_type
else None,
"correspondent": _safe_related_name(document, "correspondent"),
"document_type": _safe_related_name(document, "document_type"),
"filename": document.filename,
"storage_path": document.storage_path.name if document.storage_path else None,
"storage_path": _safe_related_name(document, "storage_path"),
"archive_serial_number": document.archive_serial_number,
"created": document.created.isoformat() if document.created else None,
"added": document.added.isoformat() if document.added else None,
@@ -8,7 +8,9 @@ from django.test import override_settings
from django.utils import timezone
from llama_index.core.schema import MetadataMode
from documents.models import Correspondent
from documents.models import Document
from documents.models import DocumentType
from documents.models import PaperlessTask
from documents.signals import document_consumption_finished
from documents.signals import document_updated
@@ -95,6 +97,38 @@ def test_build_document_node_structured_fields_in_metadata(
assert "modified" in node.metadata
@pytest.mark.django_db
def test_build_document_node_survives_concurrently_deleted_correspondent(
real_document: Document,
) -> None:
"""Regression test for #13314.
If a document's correspondent (or document type) is deleted after the
in-memory Document instance was loaded but before build_document_node
resolves the relation, accessing the FK must not raise -- it should
behave like an unset FK and produce None in the metadata instead of
aborting the whole indexing pass.
"""
correspondent = Correspondent.objects.create(name="Stale Correspondent")
document_type = DocumentType.objects.create(name="Stale Type")
real_document.correspondent = correspondent
real_document.document_type = document_type
real_document.save()
# Re-fetch to get an instance whose correspondent/document_type relations
# are unresolved (not yet cached), mirroring a task that loaded the
# document before the concurrent deletion below.
stale_document = Document.objects.get(pk=real_document.pk)
correspondent.delete()
document_type.delete()
nodes = indexing.build_document_node(stale_document)
assert len(nodes) > 0
assert nodes[0].metadata["correspondent"] is None
assert nodes[0].metadata["document_type"] is None
@pytest.mark.django_db
def test_build_document_node_excludes_document_id_from_llm_context(
real_document: Document,
+325 -178
View File
@@ -1,5 +1,7 @@
import sqlite3
from collections.abc import Generator
from collections.abc import Iterator
from contextlib import contextmanager
from pathlib import Path
import pytest
@@ -9,6 +11,7 @@ from llama_index.core.vector_stores.types import MetadataFilter
from llama_index.core.vector_stores.types import MetadataFilters
from llama_index.core.vector_stores.types import VectorStoreQuery
from paperless_ai import vector_store as vs_mod
from paperless_ai.vector_store import DB_FILENAME
from paperless_ai.vector_store import DEFAULT_TABLE_NAME
from paperless_ai.vector_store import MIGRATIONS
@@ -89,8 +92,87 @@ def _ne_filter(document_id: str):
)
def _chunk_index_rows(
store: PaperlessSqliteVecVectorStore,
document_id: str | None = None,
) -> list[tuple[str, str]]:
"""The document_chunks side table, as (chunk_id, document_id) pairs."""
sql = "SELECT chunk_id, document_id FROM document_chunks"
params: list[str] = []
if document_id is not None:
sql += " WHERE document_id = ?"
params.append(document_id)
rows = store.client.execute(sql + " ORDER BY chunk_id", params).fetchall()
return [(r["chunk_id"], r["document_id"]) for r in rows]
@contextmanager
def _pending_migrations(*migrations: Migration, schema_version: int) -> Iterator[None]:
"""Register ``migrations`` and raise the module's SCHEMA_VERSION for the
duration of the block, so check_and_run_migrations() sees them as pending.
"""
original = vs_mod.SCHEMA_VERSION
MIGRATIONS.extend(migrations)
vs_mod.SCHEMA_VERSION = schema_version
try:
yield
finally:
for migration in migrations:
MIGRATIONS.remove(migration)
vs_mod.SCHEMA_VERSION = original
def _copying_apply(src: sqlite3.Connection, dst: sqlite3.Connection, dim: int) -> None:
"""A structural migration apply() that just copies every row across.
Deliberately spells its SQL out rather than reusing the production
helpers, so these tests exercise the migration machinery against a
realistic third-party apply() instead of co-drifting with it.
"""
dst.execute( # nosemgrep
f"CREATE VIRTUAL TABLE {DEFAULT_TABLE_NAME} USING vec0("
"id TEXT PRIMARY KEY, document_id TEXT, modified TEXT,"
f" +node_content TEXT, embedding float[{dim}] distance_metric=cosine"
")",
)
dst.execute(
"INSERT INTO index_meta (key, value) VALUES ('dim', ?) "
"ON CONFLICT(key) DO UPDATE SET value = excluded.value",
(str(dim),),
)
rows = src.execute(
"SELECT id, document_id, modified, node_content, embedding "
f"FROM {DEFAULT_TABLE_NAME}",
).fetchall()
dst.execute("BEGIN IMMEDIATE")
dst.executemany(
f"INSERT INTO {DEFAULT_TABLE_NAME} "
"(id, document_id, modified, node_content, embedding) "
"VALUES (?, ?, ?, ?, ?)",
[
(
r["id"],
r["document_id"],
r["modified"],
r["node_content"],
bytes(r["embedding"]),
)
for r in rows
],
)
dst.execute(
"INSERT INTO index_meta (key, value) VALUES ('total_inserts', ?) "
"ON CONFLICT(key) DO UPDATE SET value = excluded.value",
(str(len(rows)),),
)
dst.execute("COMMIT")
class TestCrud:
def test_add_then_query_returns_node(self, store) -> None:
def test_add_then_query_returns_node(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
node = make_node("n1", "1")
assert store.add([node]) == ["n1"]
result = _query(store, node.embedding, top_k=1)
@@ -99,21 +181,30 @@ class TestCrud:
# cosine distance of the identical vector is 0 -> similarity 1
assert result.similarities[0] == pytest.approx(1.0)
def test_query_empty_store_returns_empty_no_raise(self, store) -> None:
def test_query_empty_store_returns_empty_no_raise(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
result = _query(store, [0.0] * DIM)
assert result.ids == [] and result.nodes == [] and result.similarities == []
def test_add_empty_list_is_noop(self, store) -> None:
def test_add_empty_list_is_noop(self, store: PaperlessSqliteVecVectorStore) -> None:
assert store.add([]) == []
assert not store.table_exists()
def test_delete_removes_all_chunks_of_document(self, store) -> None:
def test_delete_removes_all_chunks_of_document(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1"), make_node("a2", "1"), make_node("b1", "2")])
store.delete("1")
result = _query(store, [0.0] * DIM, top_k=10)
assert result.ids == ["b1"]
def test_query_with_in_filter_scopes_results(self, store) -> None:
def test_query_with_in_filter_scopes_results(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add(
[
make_node("a1", "1", seed=0.0),
@@ -124,7 +215,10 @@ class TestCrud:
result = _query(store, [0.0] * DIM, top_k=10, filters=_in_filter(["2", "3"]))
assert sorted(result.ids) == ["b1", "c1"]
def test_query_respects_top_k_with_filter(self, store) -> None:
def test_query_respects_top_k_with_filter(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
# k semantics: global top-k even with IN filters (document_id is a
# metadata column, not a partition key -- see design doc).
store.add(
@@ -139,7 +233,10 @@ class TestCrud:
assert len(result.ids) == 3
assert result.similarities == sorted(result.similarities, reverse=True)
def test_get_nodes_filter_and_empty_paths(self, store) -> None:
def test_get_nodes_filter_and_empty_paths(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
assert store.get_nodes(filters=_in_filter(["1"])) == [] # no table yet
store.add([make_node("a1", "1"), make_node("b1", "2")])
nodes = store.get_nodes(filters=_in_filter(["1"]))
@@ -147,7 +244,10 @@ class TestCrud:
assert nodes[0].embedding is not None
assert store.get_nodes(filters=_in_filter(["999"])) == []
def test_query_with_eq_filter_scopes_results(self, store) -> None:
def test_query_with_eq_filter_scopes_results(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add(
[
make_node("a1", "1", seed=0.0),
@@ -163,18 +263,25 @@ class TestCrud:
)
assert result.ids == ["b1"]
def test_get_nodes_node_ids_not_implemented(self, store) -> None:
def test_get_nodes_node_ids_not_implemented(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
with pytest.raises(NotImplementedError):
store.get_nodes(node_ids=["x"])
def test_fresh_instance_sees_existing_table(self, store, tmp_path: Path) -> None:
def test_fresh_instance_sees_existing_table(
self,
store: PaperlessSqliteVecVectorStore,
tmp_path: Path,
) -> None:
store.add([make_node("a1", "1")])
with PaperlessSqliteVecVectorStore(uri=str(tmp_path)) as reopened:
assert reopened.table_exists()
assert reopened.vector_dim() == DIM
assert _query(reopened, [0.0] * DIM, top_k=1).ids == ["a1"]
def test_table_exists_and_drop(self, store) -> None:
def test_table_exists_and_drop(self, store: PaperlessSqliteVecVectorStore) -> None:
assert not store.table_exists()
store.add([make_node("a1", "1")])
assert store.table_exists()
@@ -189,7 +296,10 @@ class TestBuildWhere:
assert where == "(document_id != ?)"
assert params == ["1"]
def test_query_with_ne_filter_excludes_matching_document(self, store) -> None:
def test_query_with_ne_filter_excludes_matching_document(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1"), make_node("b1", "2")])
assert sorted(
_query(store, [0.0] * DIM, top_k=5, filters=_ne_filter("1")).ids,
@@ -214,7 +324,10 @@ class TestBuildWhere:
assert where == "1 = 0"
assert params == []
def test_query_with_untranslatable_filter_returns_no_rows(self, store) -> None:
def test_query_with_untranslatable_filter_returns_no_rows(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1"), make_node("b1", "2")])
nested = MetadataFilters(
filters=[
@@ -232,7 +345,10 @@ class TestBuildWhere:
class TestUpsert:
def test_upsert_replaces_and_prunes_stale_chunks(self, store) -> None:
def test_upsert_replaces_and_prunes_stale_chunks(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add(
[make_node("d1c1", "1"), make_node("d1c2", "1"), make_node("d2c1", "2")],
)
@@ -240,18 +356,24 @@ class TestUpsert:
result = _query(store, [0.0] * DIM, top_k=10)
assert sorted(result.ids) == ["d1new", "d2c1"]
def test_upsert_creates_table_when_missing(self, store) -> None:
def test_upsert_creates_table_when_missing(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.upsert_document("1", [make_node("a1", "1")])
assert _query(store, [0.0] * DIM, top_k=1).ids == ["a1"]
def test_upsert_empty_nodes_removes_document(self, store) -> None:
def test_upsert_empty_nodes_removes_document(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1"), make_node("b1", "2")])
store.upsert_document("1", [])
assert _query(store, [0.0] * DIM, top_k=10).ids == ["b1"]
def test_upsert_is_atomic_for_concurrent_readers(
self,
store,
store: PaperlessSqliteVecVectorStore,
tmp_path: Path,
) -> None:
"""A second connection must never observe document 1 half-replaced."""
@@ -262,8 +384,52 @@ class TestUpsert:
assert ids == ["a3"]
class TestDocumentChunksIndex:
"""document_chunks lets delete()/upsert_document() find a document's chunk
ids without a vec0 full table scan on document_id -- see
PaperlessSqliteVecVectorStore._delete_chunks_by_document_id."""
def test_add_populates_document_chunks(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1"), make_node("a2", "1"), make_node("b1", "2")])
assert _chunk_index_rows(store) == [
("a1", "1"),
("a2", "1"),
("b1", "2"),
]
def test_delete_clears_document_chunks(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1"), make_node("a2", "1"), make_node("b1", "2")])
store.delete("1")
assert _chunk_index_rows(store) == [("b1", "2")]
def test_upsert_replaces_document_chunks(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1"), make_node("b1", "2")])
store.upsert_document("1", [make_node("a2", "1")])
assert _chunk_index_rows(store) == [("a2", "1"), ("b1", "2")]
def test_drop_table_clears_document_chunks(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1")])
store.drop_table()
assert _chunk_index_rows(store) == []
class TestMetadataCoercion:
def test_none_metadata_values_become_empty_strings(self, store) -> None:
def test_none_metadata_values_become_empty_strings(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
node = make_node("a1", "1")
node.metadata["modified"] = None
store.add([node]) # must not raise (vec0 rejects NULL metadata)
@@ -308,10 +474,16 @@ class TestModelNameTracking:
class TestGetModifiedTimes:
def test_empty_store_returns_empty_dict(self, store) -> None:
def test_empty_store_returns_empty_dict(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
assert store.get_modified_times() == {}
def test_returns_one_entry_per_document(self, store) -> None:
def test_returns_one_entry_per_document(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add(
[
make_node("a1", "1", modified="2026-01-01T00:00:00"),
@@ -326,7 +498,7 @@ class TestGetModifiedTimes:
class TestCompact:
def _bloat_ratio(self, store) -> float:
def _bloat_ratio(self, store: PaperlessSqliteVecVectorStore) -> float:
live = store.client.execute(
"SELECT count(*) FROM documents",
).fetchone()[0]
@@ -338,19 +510,25 @@ class TestCompact:
total = int(row["value"]) if row else live
return total / max(live, 1)
def _churn(self, store, cycles: int) -> None:
def _churn(self, store: PaperlessSqliteVecVectorStore, cycles: int) -> None:
for i in range(cycles):
store.upsert_document(
"1",
[make_node(f"gen{i}-{j}", "1", seed=float(j)) for j in range(20)],
)
def test_compact_noop_below_threshold(self, store) -> None:
def test_compact_noop_below_threshold(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1")])
store.compact()
assert _query(store, [0.0] * DIM, top_k=1).ids == ["a1"]
def test_force_compact_preserves_rows_and_metadata(self, store) -> None:
def test_force_compact_preserves_rows_and_metadata(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1"), make_node("b1", "2", seed=3.0)])
self._churn(store, 5)
before = {
@@ -370,22 +548,44 @@ class TestCompact:
store.upsert_document("3", [make_node("c1", "3", seed=100.0)])
assert "c1" in _query(store, [100.0] * DIM, top_k=1).ids
def test_auto_compact_triggers_on_churn(self, store) -> None:
def test_auto_compact_triggers_on_churn(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node(f"s{j}", "1", seed=float(j)) for j in range(20)])
self._churn(store, 5)
assert self._bloat_ratio(store) > 2
store.compact()
assert self._bloat_ratio(store) == pytest.approx(1.0)
def test_compact_on_missing_table_is_noop(self, store) -> None:
def test_compact_on_missing_table_is_noop(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.compact()
store.compact(force=True)
def test_compact_preserves_document_chunks(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
"""document_chunks must survive the file-swap rebuild, or delete()
would silently stop finding chunk ids for anything indexed before a
compaction ran."""
store.add([make_node("a1", "1"), make_node("b1", "2")])
self._churn(store, 5)
store.compact(force=True)
assert _chunk_index_rows(store, "2") == [("b1", "2")]
store.delete("2")
assert "b1" not in _query(store, [0.0] * DIM, top_k=10).ids
def test_failed_compact_removes_temp_wal_and_shm(
self,
store,
store: PaperlessSqliteVecVectorStore,
tmp_path: Path,
monkeypatch,
monkeypatch: pytest.MonkeyPatch,
) -> None:
"""A compact() that raises mid-rebuild must leave no .compact* files.
@@ -420,8 +620,8 @@ class TestCompact:
def test_force_compact_streams_rows_across_batches(
self,
store,
monkeypatch,
store: PaperlessSqliteVecVectorStore,
monkeypatch: pytest.MonkeyPatch,
) -> None:
"""Rebuild must preserve every row when rows span multiple batches.
@@ -437,11 +637,15 @@ class TestCompact:
class TestDbFile:
def test_single_db_file_in_index_dir(self, store, tmp_path: Path) -> None:
def test_single_db_file_in_index_dir(
self,
store: PaperlessSqliteVecVectorStore,
tmp_path: Path,
) -> None:
store.add([make_node("a1", "1")])
assert (tmp_path / DB_FILENAME).exists()
def test_wal_mode_enabled(self, store) -> None:
def test_wal_mode_enabled(self, store: PaperlessSqliteVecVectorStore) -> None:
assert (
store.client.execute("PRAGMA journal_mode").fetchone()[0].lower() == "wal"
)
@@ -456,193 +660,136 @@ class TestMigrations:
).fetchone()
return int(row[0]) if row else None
def test_new_table_records_schema_version(self, store) -> None:
def test_new_table_records_schema_version(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1")])
assert self._schema_version(store) == SCHEMA_VERSION
def test_check_migrations_no_table_returns_false(self, store) -> None:
def test_check_migrations_no_table_returns_false(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
assert store.check_and_run_migrations() is False
def test_check_migrations_current_version_returns_false(self, store) -> None:
def test_check_migrations_current_version_returns_false(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1")])
assert store.check_and_run_migrations() is False
def test_reembed_migration_returns_true(self, store, tmp_path: Path) -> None:
def test_reembed_migration_returns_true(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1")])
# store was just created at the real (current) SCHEMA_VERSION, so the
# simulated pending migration must target one version past that --
# not a hardcoded 1 -> 2, which would collide with the real
# document_chunks migration already registered in MIGRATIONS.
migration = Migration(
from_version=1,
to_version=2,
from_version=SCHEMA_VERSION,
to_version=SCHEMA_VERSION + 1,
kind="re-embed",
description="test re-embed",
)
MIGRATIONS.append(migration)
try:
from paperless_ai import vector_store as vs_mod
original = vs_mod.SCHEMA_VERSION
vs_mod.SCHEMA_VERSION = 2
result = store.check_and_run_migrations()
finally:
MIGRATIONS.remove(migration)
vs_mod.SCHEMA_VERSION = original
assert result is True
with _pending_migrations(migration, schema_version=SCHEMA_VERSION + 1):
assert store.check_and_run_migrations() is True
def test_structural_migration_copies_rows_and_updates_version(
self,
store,
tmp_path: Path,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1"), make_node("b1", "2")])
def apply(
src: sqlite3.Connection,
dst: sqlite3.Connection,
dim: int,
) -> None:
dst.execute( # nosemgrep
f"CREATE VIRTUAL TABLE {DEFAULT_TABLE_NAME} USING vec0("
"id TEXT PRIMARY KEY, document_id TEXT, modified TEXT,"
f" +node_content TEXT, embedding float[{dim}] distance_metric=cosine"
")",
)
dst.execute(
"INSERT INTO index_meta (key, value) VALUES ('dim', ?) "
"ON CONFLICT(key) DO UPDATE SET value = excluded.value",
(str(dim),),
)
rows = src.execute(
"SELECT id, document_id, modified, node_content, embedding "
f"FROM {DEFAULT_TABLE_NAME}",
).fetchall()
dst.execute("BEGIN IMMEDIATE")
dst.executemany(
f"INSERT INTO {DEFAULT_TABLE_NAME} "
"(id, document_id, modified, node_content, embedding) "
"VALUES (?, ?, ?, ?, ?)",
[
(
r["id"],
r["document_id"],
r["modified"],
r["node_content"],
bytes(r["embedding"]),
)
for r in rows
],
)
dst.execute(
"INSERT INTO index_meta (key, value) VALUES ('total_inserts', ?) "
"ON CONFLICT(key) DO UPDATE SET value = excluded.value",
(str(len(rows)),),
)
dst.execute("COMMIT")
migration = Migration(
from_version=1,
to_version=2,
from_version=SCHEMA_VERSION,
to_version=SCHEMA_VERSION + 1,
kind="structural",
description="test structural",
apply=apply,
apply=_copying_apply,
)
MIGRATIONS.append(migration)
try:
from paperless_ai import vector_store as vs_mod
with _pending_migrations(migration, schema_version=SCHEMA_VERSION + 1):
assert store.check_and_run_migrations() is False
original = vs_mod.SCHEMA_VERSION
vs_mod.SCHEMA_VERSION = 2
result = store.check_and_run_migrations()
finally:
MIGRATIONS.remove(migration)
vs_mod.SCHEMA_VERSION = original
assert result is False
assert self._schema_version(store) == 2
assert self._schema_version(store) == SCHEMA_VERSION + 1
ids = {n.node_id for n in store.get_nodes()}
assert ids == {"a1", "b1"}
def test_compact_preserves_schema_version(self, store) -> None:
def test_compact_preserves_schema_version(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
store.add([make_node("a1", "1")])
assert self._schema_version(store) == SCHEMA_VERSION
store.compact(force=True)
assert self._schema_version(store) == SCHEMA_VERSION
def test_stop_at_reembed_boundary(self, store) -> None:
# Registry: structural v2, re-embed v3, structural v4.
# Only v2 should apply; the re-embed boundary must stop execution
# before v4 runs, and the stored version must stay at 2.
def test_v1_to_v2_migration_backfills_document_chunks(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
"""The real v1 -> v2 migration (registered in MIGRATIONS) must
backfill document_chunks for rows written before it existed, or
delete()/upsert_document() would find zero chunk ids for them and
leave their vec0 rows orphaned forever."""
store.add([make_node("a1", "1"), make_node("a2", "1"), make_node("b1", "2")])
# Simulate a pre-migration (schema v1) store: rows exist in the vec0
# table, but document_chunks (added by the v1 -> v2 migration) has no
# entries for them, mirroring a real on-disk index created before
# this migration existed.
store.client.execute("DELETE FROM document_chunks")
store.client.execute(
"UPDATE index_meta SET value = '1' WHERE key = 'schema_version'",
)
assert store.check_and_run_migrations() is False # structural, not re-embed
assert self._schema_version(store) == SCHEMA_VERSION
assert _chunk_index_rows(store) == [
("a1", "1"),
("a2", "1"),
("b1", "2"),
]
# And delete() now actually removes rows for a document that
# predates the migration, instead of silently no-op'ing.
store.delete("1")
assert sorted(_query(store, [0.0] * DIM, top_k=10).ids) == ["b1"]
def test_stop_at_reembed_boundary(
self,
store: PaperlessSqliteVecVectorStore,
) -> None:
# Registry, relative to the store's current (real) SCHEMA_VERSION N:
# structural N+1, re-embed N+2, structural N+3. Only N+1 should apply;
# the re-embed boundary must stop execution before N+3 runs, and the
# stored version must stay at N+1.
store.add([make_node("a1", "1"), make_node("b1", "2")])
def copy_apply(
src: sqlite3.Connection,
dst: sqlite3.Connection,
dim: int,
) -> None:
dst.execute( # nosemgrep
f"CREATE VIRTUAL TABLE {DEFAULT_TABLE_NAME} USING vec0("
"id TEXT PRIMARY KEY, document_id TEXT, modified TEXT,"
f" +node_content TEXT, embedding float[{dim}] distance_metric=cosine"
")",
)
dst.execute(
"INSERT INTO index_meta (key, value) VALUES ('dim', ?) "
"ON CONFLICT(key) DO UPDATE SET value = excluded.value",
(str(dim),),
)
rows = src.execute(
"SELECT id, document_id, modified, node_content, embedding "
f"FROM {DEFAULT_TABLE_NAME}",
).fetchall()
dst.execute("BEGIN IMMEDIATE")
dst.executemany(
f"INSERT INTO {DEFAULT_TABLE_NAME} "
"(id, document_id, modified, node_content, embedding) "
"VALUES (?, ?, ?, ?, ?)",
[
(
r["id"],
r["document_id"],
r["modified"],
r["node_content"],
bytes(r["embedding"]),
)
for r in rows
],
)
dst.execute("COMMIT")
migrations = [
Migration(
from_version=1,
to_version=2,
from_version=SCHEMA_VERSION,
to_version=SCHEMA_VERSION + 1,
kind="structural",
description="v2 structural",
apply=copy_apply,
description="structural",
apply=_copying_apply,
),
Migration(
from_version=2,
to_version=3,
from_version=SCHEMA_VERSION + 1,
to_version=SCHEMA_VERSION + 2,
kind="re-embed",
description="v3 re-embed boundary",
description="re-embed boundary",
),
Migration(
from_version=3,
to_version=4,
from_version=SCHEMA_VERSION + 2,
to_version=SCHEMA_VERSION + 3,
kind="structural",
description="v4 structural - must not run",
apply=copy_apply,
description="structural - must not run",
apply=_copying_apply,
),
]
MIGRATIONS.extend(migrations)
try:
from paperless_ai import vector_store as vs_mod
with _pending_migrations(*migrations, schema_version=SCHEMA_VERSION + 3):
assert store.check_and_run_migrations() is True
original = vs_mod.SCHEMA_VERSION
vs_mod.SCHEMA_VERSION = 4
result = store.check_and_run_migrations()
finally:
for m in migrations:
MIGRATIONS.remove(m)
vs_mod.SCHEMA_VERSION = original
assert result is True
assert self._schema_version(store) == 2
assert self._schema_version(store) == SCHEMA_VERSION + 1
+149 -41
View File
@@ -31,10 +31,20 @@ logger = logging.getLogger("paperless_ai.vector_store")
DB_FILENAME = "llmindex.db"
DEFAULT_TABLE_NAME = "documents"
_INSERT = (
"INSERT INTO "
+ DEFAULT_TABLE_NAME
+ " (id, document_id, modified, node_content, embedding) VALUES (?, ?, ?, ?, ?)"
)
_INSERT_CHUNK_INDEX = (
"INSERT INTO document_chunks (chunk_id, document_id) VALUES (?, ?)"
)
# Current schema version. Written to index_meta at table creation and bumped
# whenever a Migration is added to MIGRATIONS. check_and_run_migrations() uses
# this to decide which migrations to run on an existing store.
SCHEMA_VERSION = 1
SCHEMA_VERSION = 2
# compact(): rebuild when the cumulative rowid count exceeds this multiple of
# the live row count. DELETEs on vec0 tables never reclaim space (upstream
@@ -80,8 +90,10 @@ class Migration:
)
# Registry of all schema migrations in order. Empty at v1 -- this is the
# baseline. Add entries here (and bump SCHEMA_VERSION) when the schema changes.
# Registry of all schema migrations in order. Populated after the class body
# below, since v1 -> v2's apply() needs PaperlessSqliteVecVectorStore's own
# static methods. Add entries here (and bump SCHEMA_VERSION) when the schema
# changes.
MIGRATIONS: list[Migration] = []
@@ -93,6 +105,42 @@ def _unpack(blob: bytes) -> list[float]:
return list(struct.unpack(f"{len(blob) // 4}f", blob))
def _copy_rows(src_conn: sqlite3.Connection, dst_conn: sqlite3.Connection) -> int:
"""Copy every live vec0 row from ``src_conn`` into ``dst_conn``, recording
each one in ``dst_conn``'s document_chunks side table. Returns the number
of rows copied. The caller owns ``dst_conn``'s transaction.
Rows are streamed from the source cursor in batches instead of being
materialized all at once, so a large index does not cause an OOM during a
routine compaction or migration.
"""
src_cursor = src_conn.execute(
"SELECT id, document_id, modified, node_content, embedding FROM "
+ DEFAULT_TABLE_NAME,
)
copied = 0
while batch := src_cursor.fetchmany(COMPACT_BATCH_SIZE):
dst_conn.executemany(
_INSERT,
[
(
r["id"],
r["document_id"],
r["modified"],
r["node_content"],
bytes(r["embedding"]),
)
for r in batch
],
)
dst_conn.executemany(
_INSERT_CHUNK_INDEX,
[(r["id"], r["document_id"]) for r in batch],
)
copied += len(batch)
return copied
def _build_where(filters: MetadataFilters | None) -> tuple[str, list[str]]:
"""Translate the EQ / IN / NE filters we use into a parameterized SQL clause
on vec0 metadata columns. Returns ("", []) when there is nothing to filter.
@@ -189,6 +237,18 @@ class PaperlessSqliteVecVectorStore(BasePydanticVectorStore):
conn.execute(
"CREATE TABLE IF NOT EXISTS index_meta (key TEXT PRIMARY KEY, value TEXT)",
)
# vec0 metadata columns only get an efficient lookup path inside a KNN
# (MATCH) query; a plain `WHERE document_id = ?` is a full table scan
# regardless of index size. This plain, indexed table is how delete()/
# upsert_document() find a document's chunk ids without that scan.
conn.execute(
"CREATE TABLE IF NOT EXISTS document_chunks "
"(chunk_id TEXT PRIMARY KEY, document_id TEXT NOT NULL)",
)
conn.execute(
"CREATE INDEX IF NOT EXISTS idx_document_chunks_document_id "
"ON document_chunks (document_id)",
)
return conn
@property
@@ -223,13 +283,17 @@ class PaperlessSqliteVecVectorStore(BasePydanticVectorStore):
else:
self._conn.execute("COMMIT")
def _meta_get(self, key: str) -> str | None:
row = self._conn.execute(
@staticmethod
def _meta_get_on(conn: sqlite3.Connection, key: str) -> str | None:
row = conn.execute(
"SELECT value FROM index_meta WHERE key = ?",
(key,),
).fetchone()
return row["value"] if row else None
def _meta_get(self, key: str) -> str | None:
return self._meta_get_on(self._conn, key)
@staticmethod
def _meta_set_on(conn: sqlite3.Connection, key: str, value: str) -> None:
conn.execute(
@@ -259,6 +323,7 @@ class PaperlessSqliteVecVectorStore(BasePydanticVectorStore):
def drop_table(self) -> None:
self._conn.execute("DROP TABLE IF EXISTS " + DEFAULT_TABLE_NAME)
self._conn.execute("DELETE FROM index_meta")
self._conn.execute("DELETE FROM document_chunks")
def stored_model_name(self) -> str | None:
"""Return the embedding model name recorded at table creation, or None."""
@@ -325,11 +390,37 @@ class PaperlessSqliteVecVectorStore(BasePydanticVectorStore):
_pack(node.get_embedding()),
)
_INSERT = (
"INSERT INTO "
+ DEFAULT_TABLE_NAME
+ " (id, document_id, modified, node_content, embedding) VALUES (?, ?, ?, ?, ?)"
)
def _index_chunks(self, rows: list[tuple[str, str, str, str, bytes]]) -> None:
"""Record each row's (chunk_id, document_id) in the document_chunks
side table, kept in lockstep with every insert into the vec0 table."""
self._conn.executemany(
_INSERT_CHUNK_INDEX,
[(chunk_id, document_id) for chunk_id, document_id, *_ in rows],
)
def _delete_chunks_by_document_id(self, document_id: str) -> None:
"""Delete all of a document's chunks via point-deletes on `id`.
vec0 has no efficient lookup on the document_id metadata column
outside a KNN query (see _open_connection), so a plain
`DELETE ... WHERE document_id = ?` is a full table scan regardless of
index size. Looking the chunk ids up in document_chunks first (a real
indexed lookup) and deleting each by its `id` primary key instead
turns that scan into a handful of O(1) point deletes.
"""
doc_id = str(document_id)
chunk_rows = self._conn.execute(
"SELECT chunk_id FROM document_chunks WHERE document_id = ?",
(doc_id,),
).fetchall()
self._conn.executemany(
"DELETE FROM " + DEFAULT_TABLE_NAME + " WHERE id = ?",
[(row["chunk_id"],) for row in chunk_rows],
)
self._conn.execute(
"DELETE FROM document_chunks WHERE document_id = ?",
(doc_id,),
)
def _increment_total_inserts(self, count: int) -> None:
"""Increment the cumulative insert counter stored in index_meta.
@@ -348,7 +439,8 @@ class PaperlessSqliteVecVectorStore(BasePydanticVectorStore):
rows = [self._row(node) for node in nodes]
with self._transaction():
self._ensure_table(len(nodes[0].get_embedding()))
self._conn.executemany(self._INSERT, rows)
self._conn.executemany(_INSERT, rows)
self._index_chunks(rows)
self._increment_total_inserts(len(rows))
return [node.node_id for node in nodes]
@@ -365,22 +457,17 @@ class PaperlessSqliteVecVectorStore(BasePydanticVectorStore):
if nodes:
self._ensure_table(len(nodes[0].get_embedding()))
if self.table_exists():
self._conn.execute(
"DELETE FROM " + DEFAULT_TABLE_NAME + " WHERE document_id = ?",
(str(document_id),),
)
self._delete_chunks_by_document_id(document_id)
if rows:
self._conn.executemany(self._INSERT, rows)
self._conn.executemany(_INSERT, rows)
self._index_chunks(rows)
self._increment_total_inserts(len(rows))
return [node.node_id for node in nodes]
def delete(self, ref_doc_id: str, **delete_kwargs: Any) -> None:
if self.table_exists():
with self._transaction():
self._conn.execute(
"DELETE FROM " + DEFAULT_TABLE_NAME + " WHERE document_id = ?",
(str(ref_doc_id),),
)
self._delete_chunks_by_document_id(ref_doc_id)
def _rows_to_nodes(self, rows: list[sqlite3.Row]) -> list[BaseNode]:
nodes: list[BaseNode] = []
@@ -508,28 +595,8 @@ class PaperlessSqliteVecVectorStore(BasePydanticVectorStore):
value = self._meta_get(key)
if value is not None:
self._meta_set_on(new_conn, key, value)
src_cursor = self._conn.execute(
"SELECT id, document_id, modified, node_content, embedding "
"FROM " + DEFAULT_TABLE_NAME,
)
new_conn.execute("BEGIN IMMEDIATE")
# Stream rows from the source cursor in batches instead of
# materializing the whole table in memory, so a large index does
# not cause an OOM during routine maintenance compactions.
while batch := src_cursor.fetchmany(COMPACT_BATCH_SIZE):
new_conn.executemany(
self._INSERT,
[
(
r["id"],
r["document_id"],
r["modified"],
r["node_content"],
bytes(r["embedding"]),
)
for r in batch
],
)
_copy_rows(self._conn, new_conn)
# Reset the cumulative counter: after compact, total_inserts == live.
self._meta_set_on(new_conn, "total_inserts", str(live))
new_conn.execute("COMMIT")
@@ -614,3 +681,44 @@ class PaperlessSqliteVecVectorStore(BasePydanticVectorStore):
raise
new_conn.close()
self._swap_in_compact(compact_path, db_path)
def _migrate_v1_to_v2_add_document_chunks(
src_conn: sqlite3.Connection,
dst_conn: sqlite3.Connection,
dim: int,
) -> None:
"""v1 -> v2: backfill the document_chunks side table.
document_chunks (see PaperlessSqliteVecVectorStore._open_connection) lets
delete()/upsert_document() find a document's chunk ids without a vec0
full table scan on the document_id metadata column. Every row written
before this migration predates that table, so without backfilling,
deleting a pre-migration document would find zero chunk ids and leave its
vec0 rows permanently orphaned. Backfilling is a plain row copy into the
new-schema file (``_copy_rows``, the same helper compact() uses), which
records every copied row in document_chunks as it goes.
"""
PaperlessSqliteVecVectorStore._create_vec_table(dst_conn, dim)
PaperlessSqliteVecVectorStore._meta_set_on(dst_conn, "dim", str(dim))
embed_model = PaperlessSqliteVecVectorStore._meta_get_on(src_conn, "embed_model")
if embed_model is not None:
PaperlessSqliteVecVectorStore._meta_set_on(dst_conn, "embed_model", embed_model)
dst_conn.execute("BEGIN IMMEDIATE")
live = _copy_rows(src_conn, dst_conn)
# This migration only ever copies live rows (like compact()), so the
# cumulative counter resets to match -- the new file has no bloat yet.
PaperlessSqliteVecVectorStore._meta_set_on(dst_conn, "total_inserts", str(live))
dst_conn.execute("COMMIT")
MIGRATIONS.append(
Migration(
from_version=1,
to_version=2,
kind="structural",
description="add document_chunks side table for O(1) per-document deletes",
apply=_migrate_v1_to_v2_add_document_chunks,
),
)
@@ -10,6 +10,24 @@ def clamp_mailrule_maximum_age(apps, schema_editor):
MailRule.objects.filter(maximum_age__gt=32767).update(maximum_age=32767)
def clamp_mailrule_order(apps, schema_editor):
# The order field of MailRule is only used for relative sorting
# (account.rules.order_by("order")), so out-of-range values must be
# renumbered by rank rather than clamped to a single value, which
# would collapse distinct rules to the same order.
MailRule = apps.get_model("paperless_mail", "MailRule")
if (
MailRule.objects.filter(order__gt=32767).exists()
or MailRule.objects.filter(
order__lt=-32768,
).exists()
): # pragma: no cover
for index, rule_id in enumerate(
MailRule.objects.order_by("order", "pk").values_list("pk", flat=True),
):
MailRule.objects.filter(pk=rule_id).update(order=index)
class Migration(migrations.Migration):
# The data update must commit before PostgreSQL rewrites the table, otherwise
# pending foreign-key trigger events can block the subsequent ALTER TABLE.
@@ -135,6 +153,10 @@ class Migration(migrations.Migration):
verbose_name="maximum age",
),
),
migrations.RunPython(
clamp_mailrule_order,
migrations.RunPython.noop,
),
migrations.AlterField(
model_name="mailrule",
name="order",