Feature: OCR Templates (#13043)

[skip ci]

Signed-off-by: dependabot[bot] <support@github.com>
Co-Authored-By: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-Authored-By: stumpylog <797416+stumpylog@users.noreply.github.com>
Co-Authored-By: GitHub Actions <41898282+github-actions[bot]@users.noreply.github.com>
Co-Authored-By: shamoon <4887959+shamoon@users.noreply.github.com>
This commit is contained in:
Christoph Schlaepfer
2026-06-23 07:32:58 -07:00
committed by shamoon
co-authored by dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> stumpylog GitHub Actions <41898282+github-actions[bot]@users.noreply.github.com> shamoon
parent bf70e597ee
commit bf73b5b1d1
26 changed files with 4545 additions and 1 deletions
+26
View File
@@ -13,6 +13,8 @@ import { DocumentDetailComponent } from './components/document-detail/document-d
import { DocumentListComponent } from './components/document-list/document-list.component'
import { DocumentAttributesComponent } from './components/manage/document-attributes/document-attributes.component'
import { MailComponent } from './components/manage/mail/mail.component'
import { OcrTemplateEditorComponent } from './components/manage/ocr-templates/ocr-template-editor/ocr-template-editor.component'
import { OcrTemplatesComponent } from './components/manage/ocr-templates/ocr-templates.component'
import { SavedViewsComponent } from './components/manage/saved-views/saved-views.component'
import { WorkflowsComponent } from './components/manage/workflows/workflows.component'
import { NotFoundComponent } from './components/not-found/not-found.component'
@@ -274,6 +276,30 @@ export const routes: Routes = [
componentName: 'WorkflowsComponent',
},
},
{
path: 'ocr-templates',
component: OcrTemplatesComponent,
canActivate: [PermissionsGuard],
data: {
requiredPermission: {
action: PermissionAction.View,
type: PermissionType.OcrTemplate,
},
componentName: 'OcrTemplatesComponent',
},
},
{
path: 'ocr-templates/:id',
component: OcrTemplateEditorComponent,
canActivate: [PermissionsGuard],
data: {
requiredPermission: {
action: PermissionAction.Change,
type: PermissionType.OcrTemplate,
},
componentName: 'OcrTemplateEditorComponent',
},
},
{
path: 'mail',
component: MailComponent,
@@ -243,6 +243,14 @@
<i-bs class="me-2" name="boxes"></i-bs><span><ng-container i18n>Workflows</ng-container></span>
</a>
</li>
<li class="nav-item app-link"
*pngxIfPermissions="{ action: PermissionAction.View, type: PermissionType.OcrTemplate }">
<a class="nav-link" routerLink="ocr-templates" routerLinkActive="active" (click)="closeMenu()"
ngbPopover="OCR Templates" i18n-ngbPopover [disablePopover]="!slimSidebarEnabled" placement="end"
container="body" triggers="mouseenter:mouseleave" popoverClass="popover-slim">
<i-bs class="me-2" name="file-earmark-break"></i-bs><span><ng-container i18n>OCR Templates</ng-container></span>
</a>
</li>
<li class="nav-item app-link" *pngxIfPermissions="{ action: PermissionAction.View, type: PermissionType.MailAccount }"
tourAnchor="tour.mail">
<a class="nav-link" routerLink="mail" routerLinkActive="active" (click)="closeMenu()" ngbPopover="Mail"
@@ -82,6 +82,14 @@
<i-bs name="pencil" class="me-1"></i-bs><ng-container i18n>PDF Editor</ng-container>
</button>
<button ngbDropdownItem (click)="runZoneOcr()" [disabled]="!userCanEdit || !document?.document_type">
<i-bs width="1em" height="1em" name="file-earmark-ruled" class="me-1"></i-bs><span i18n>Run Zone OCR</span>
</button>
<button ngbDropdownItem (click)="createOcrTemplate()">
<i-bs width="1em" height="1em" name="file-earmark-medical" class="me-1"></i-bs><span i18n>Create OCR Template</span>
</button>
@if (userIsOwner && (requiresPassword || password)) {
<button ngbDropdownItem (click)="removePassword()" [disabled]="!password">
<i-bs name="unlock" class="me-1"></i-bs><ng-container i18n>Remove Password</ng-container>
@@ -1405,6 +1405,48 @@ export class DocumentDetailComponent
})
}
runZoneOcr() {
this.documentsService.runZoneOcr(this.document.id).subscribe({
next: (res) => {
const results = res.results ?? []
if (results.length) {
const failed = results.filter(
(r) =>
r.value === null ||
r.value === undefined ||
`${r.value}`.trim() === ''
)
const filled = results.length - failed.length
let msg = $localize`Filled ${filled} of ${results.length} fields`
if (failed.length) {
const names = failed.map((r) => r.zone).join(', ')
msg = `${msg}. ${$localize`Failed to match zones: ${names}`}`
}
this.toastService.showInfo(msg)
} else {
this.toastService.showInfo(
$localize`Zone OCR ran but no results extracted.`
)
}
this.documentsService
.get(this.documentId)
.subscribe((doc) => this.updateComponent(doc))
},
error: (error) => {
this.toastService.showError($localize`Zone OCR failed`, error)
},
})
}
createOcrTemplate() {
this.router.navigate(['/ocr-templates', 'new'], {
queryParams: {
document_type: this.document.document_type,
sample_document: this.document.id,
},
})
}
private getSelectedNonLatestVersionId(): number | null {
const versions = this.document?.versions ?? []
if (!versions.length || !this.selectedVersionId) {
@@ -95,6 +95,9 @@
<button ngbDropdownItem (click)="mergeSelected()" [disabled]="!userCanAdd || list.allSelected || list.selectedCount < 2">
<i-bs name="journals" class="me-1"></i-bs><ng-container i18n>Merge</ng-container>
</button>
<button ngbDropdownItem (click)="runZoneOcrSelected()" [disabled]="!userCanEditAll || list.allSelected">
<i-bs name="file-earmark-ruled" class="me-1"></i-bs><ng-container i18n>Run Zone OCR</ng-container>
</button>
</div>
</div>
</div>
@@ -12,7 +12,15 @@ import {
} from '@ng-bootstrap/ng-bootstrap'
import { saveAs } from 'file-saver'
import { NgxBootstrapIconsModule } from 'ngx-bootstrap-icons'
import { first, map, Observable, Subject, switchMap, takeUntil } from 'rxjs'
import {
first,
forkJoin,
map,
Observable,
Subject,
switchMap,
takeUntil,
} from 'rxjs'
import { ConfirmDialogComponent } from 'src/app/components/common/confirm-dialog/confirm-dialog.component'
import { CustomField } from 'src/app/data/custom-field'
import { MatchingModel } from 'src/app/data/matching-model'
@@ -908,6 +916,27 @@ export class BulkEditorComponent
})
}
runZoneOcrSelected() {
const ids = Array.from(this.list.selected)
if (!ids.length) return
const modal = this.modalService.open(ConfirmDialogComponent, {
backdrop: 'static',
})
modal.componentInstance.title = $localize`Run Zone OCR`
modal.componentInstance.messageBold = $localize`Run zone OCR on ${this.getSelectionSize()} selected document(s)?`
modal.componentInstance.message = $localize`Each document's type template (if it has one) is applied, overwriting the mapped fields.`
modal.componentInstance.btnCaption = $localize`Proceed`
modal.componentInstance.confirmClicked
.pipe(takeUntil(this.unsubscribeNotifier))
.subscribe(() => {
modal.componentInstance.buttonsEnabled = false
this.executeDocumentAction(
modal,
forkJoin(ids.map((id) => this.documentService.runZoneOcr(id)))
)
})
}
setPermissions() {
let modal = this.modalService.open(PermissionsDialogComponent, {
backdrop: 'static',
@@ -0,0 +1,414 @@
<pngx-page-header [title]="pageTitle" [id]="template.id">
<div class="input-group input-group-sm me-5 align-items-center">
<div class="input-group-text">
<i-bs name="file-text"></i-bs>
</div>
<input
type="text"
class="form-control"
[(ngModel)]="previewDocModel"
[ngbTypeahead]="searchDocuments"
[inputFormatter]="documentFormatter"
[resultFormatter]="documentFormatter"
(selectItem)="onPreviewDocSelected($event)"
[editable]="false"
placeholder="Search documents by title..."
i18n-placeholder
/>
</div>
<div class="d-flex align-items-center flex-wrap gap-2">
<div class="input-group input-group-sm ms-2 d-none d-md-flex">
<div class="input-group-text" i18n>Page</div>
<input class="form-control flex-grow-0 w-auto" type="number" min="1" [max]="previewPageCount" [(ngModel)]="previewPageDisplay" />
<div class="input-group-text" i18n>of {{previewPageCount}}</div>
</div>
<button type="button" class="btn btn-sm btn-outline-secondary" i18n-title title="Previous" (click)="prevPage()" [disabled]="!pageImageUrl || previewPage <= 0">
<i-bs width="1.2em" height="1.2em" name="arrow-left"></i-bs>
</button>
<button type="button" class="btn btn-sm btn-outline-secondary" i18n-title title="Next" (click)="nextPage()" [disabled]="!pageImageUrl || previewPage >= (previewPageCount ?? 1) - 1">
<i-bs width="1.2em" height="1.2em" name="arrow-right"></i-bs>
</button>
<div class="input-group input-group-sm">
<button class="btn btn-outline-secondary" (click)="zoomOut()" i18n>-</button>
<span class="input-group-text">{{ zoom * 100 | number: '1.0-0' }}%</span>
<button class="btn btn-outline-secondary" (click)="zoomIn()" i18n>+</button>
</div>
</div>
</pngx-page-header>
<div class="row">
<div class="col-md-4">
<div class="btn-toolbar mb-1 border-bottom">
<div class="btn-group pb-3">
<a routerLink="/ocr-templates" class="btn btn-sm btn-outline-secondary">
<i-bs width="1.2em" height="1.2em" name="x"></i-bs>
<span class="ms-1" i18n>Close</span>
</a>
</div>
<div class="btn-group ms-auto pb-3">
<button class="btn btn-sm btn-primary" (click)="save()" [disabled]="saving">
@if (saving) {
<span class="spinner-border spinner-border-sm me-1"></span>
}
<span i18n>Save</span>
</button>
</div>
</div>
<ul ngbNav #nav="ngbNav" [(activeId)]="activeTab" class="nav-underline flex-nowrap flex-md-wrap overflow-auto">
<li ngbNavItem="settings">
<a ngbNavLink i18n>Settings</a>
<ng-template ngbNavContent>
<div class="row mb-3">
<div class="col-9">
<pngx-input-text [(ngModel)]="template.name" title="Template name" i18n-title></pngx-input-text>
</div>
<div class="col-3">
<pngx-input-switch [(ngModel)]="template.enabled" title="Enabled" i18n-title></pngx-input-switch>
</div>
</div>
<pngx-input-select [(ngModel)]="template.document_type" [items]="documentTypes" bindLabel="name" bindValue="id" title="Document type" i18n-title></pngx-input-select>
<small class="text-muted" i18n>
Draw rectangles on the preview to define extraction zones. Use the
page controls above the preview to add zones on different pages.
</small>
</ng-template>
</li>
<li ngbNavItem="zones">
<a ngbNavLink><ng-container i18n>Zones</ng-container> <span class="badge bg-primary ms-2">{{ template.zones.length }}</span></a>
<ng-template ngbNavContent>
@if (template.zones.length === 0) {
<p class="text-muted" i18n>
No zones defined. Load a document preview and draw rectangles to add zones.
</p>
}
<div class="list-group">
@for (zone of template.zones; track $index; let i = $index) {
<div
class="list-group-item list-group-item-action d-flex justify-content-between align-items-center"
[style.box-shadow]="selectedZoneIndex === i ? 'inset 3px 0 0 0 var(--bs-primary)' : null"
>
<div class="flex-grow-1" role="button" style="cursor: pointer;" (click)="selectZone(i)">
<div><strong [class.text-primary]="selectedZoneIndex === i">{{ zone.name }}</strong></div>
<div class="small text-muted">
{{ getZoneTargetName(zone) }} - {{ zone.width }}x{{ zone.height }}px <ng-container i18n>p.</ng-container>{{ zonePage(zone) }}
</div>
</div>
<div class="btn-group">
<button class="btn btn-sm btn-outline-secondary" type="button" (click)="selectZone(i)" title="Edit" i18n-title>
<i-bs name="pencil"></i-bs>
</button>
<button class="btn btn-sm btn-outline-danger" type="button" (click)="removeZone(i)" title="Delete" i18n-title>
<i-bs name="trash"></i-bs>
</button>
</div>
</div>
}
</div>
</ng-template>
</li>
<li ngbNavItem="zone">
<a ngbNavLink i18n>Zone</a>
<ng-template ngbNavContent>
@if (selectedZone; as zone) {
<div class="d-flex justify-content-between align-items-center mb-3">
<strong>{{ zone.name }}</strong>
<div class="d-flex gap-2">
<button class="btn btn-sm btn-primary" (click)="save()" [disabled]="saving">
@if (saving) {
<span class="spinner-border spinner-border-sm me-1"></span>
}
<span i18n>Save</span>
</button>
<button class="btn btn-sm btn-outline-danger" (click)="deleteSelectedZone()">
<i-bs name="trash" class="me-1"></i-bs><ng-container i18n>Delete zone</ng-container>
</button>
</div>
</div>
<div class="mb-3">
<label class="form-label" i18n>Zone Name</label>
<input
type="text"
class="form-control"
[(ngModel)]="zone.name"
(ngModelChange)="redrawCanvas()"
/>
</div>
<div class="mb-3">
<label class="form-label" i18n>Page</label>
<input
type="number"
class="form-control"
[(ngModel)]="zone.page"
min="-1"
(ngModelChange)="redrawCanvas()"
/>
<small class="text-muted" i18n>Page this zone is on. Use -1 for the last page. Set automatically when you draw it.</small>
</div>
<div class="mb-3">
<label class="form-label" i18n>Field</label>
<div class="input-group">
<select class="form-select" [ngModel]="zoneFieldValue(zone)" (ngModelChange)="setZoneField(zone, $event)">
<optgroup label="Built-in fields" i18n-label>
@for (t of builtinTargets; track t.id) {
<option [ngValue]="t.id">{{ t.name }}</option>
}
</optgroup>
<optgroup label="Custom fields" i18n-label>
@for (cf of customFields; track cf.id) {
<option [ngValue]="cf.id">{{ cf.name }} ({{ cf.data_type }})</option>
}
</optgroup>
</select>
<button
class="btn btn-outline-secondary"
type="button"
(click)="openQuickCreate(selectedZoneIndex)"
title="Create new custom field"
i18n-title
>
<i-bs name="plus"></i-bs>
</button>
</div>
<small class="text-muted" i18n>Write the extracted value to a custom field, or to a built-in field (Title, ASN, Date created).</small>
</div>
@if (isFieldShared(zone)) {
<div class="card mb-3 border-info">
<div class="card-body">
<h6 class="card-title d-flex align-items-center gap-2">
<i-bs name="braces"></i-bs>
<span i18n>Combine zones into this field</span>
</h6>
<p class="small text-muted mb-2" i18n>
More than one zone writes to this field. Build the combined
value below: click a zone to insert its token, and type any
separators or literal text between tokens.
</p>
<div class="d-flex flex-wrap gap-1 mb-2">
@for (z of zonesForField(zone); track $index) {
<button
type="button"
class="btn btn-sm btn-outline-info"
(click)="insertCombineToken(zone, z)"
title="Insert token"
i18n-title
>
+ {{ z.name || 'Zone' }}
</button>
}
</div>
<input
type="text"
class="form-control font-monospace"
[ngModel]="getCombineFormat(zone)"
(ngModelChange)="setCombineFormat(zone, $event)"
placeholder="{Zone 1} - {Zone 2}"
/>
<small class="text-muted" i18n>
Tokens are matched by zone name. An empty zone leaves its
token blank and the stray separator is trimmed. Leave empty
to just join the zones in order with a space.
</small>
</div>
</div>
}
@if (showQuickCreate) {
<div class="card mb-3 border-primary">
<div class="card-body">
<h6 class="card-title" i18n>Create Custom Field</h6>
<div class="mb-2">
<label class="form-label small" i18n>Field Name</label>
<input type="text" class="form-control form-control-sm"
[(ngModel)]="quickCreateName" placeholder="e.g. Invoice Number" />
</div>
<div class="mb-2">
<label class="form-label small" i18n>Field Type</label>
<select class="form-select form-select-sm" [(ngModel)]="quickCreateType">
@for (t of quickCreateTypes; track t.id) {
<option [ngValue]="t.id">{{ t.name }}</option>
}
</select>
</div>
<div class="d-flex gap-2">
<button class="btn btn-primary btn-sm" (click)="submitQuickCreate()"
[disabled]="!quickCreateName.trim()" i18n>
Create & Assign
</button>
<button class="btn btn-outline-secondary btn-sm" (click)="cancelQuickCreate()" i18n>
Cancel
</button>
</div>
</div>
</div>
}
<div class="mb-3">
<label class="form-label" i18n>OCR Language</label>
<ng-select
[items]="ocrLanguageOptions"
bindLabel="name"
bindValue="id"
[multiple]="true"
[closeOnSelect]="false"
[ngModel]="ocrLanguageArray(zone)"
(ngModelChange)="setOcrLanguages(zone, $event)"
placeholder="Select languages"
i18n-placeholder
></ng-select>
</div>
<div class="mb-3">
<label class="form-label" i18n>Transform</label>
<select class="form-select" [(ngModel)]="zone.transform">
@for (opt of transformOptions; track opt.id) {
<option [ngValue]="opt.id">{{ opt.name }}</option>
}
</select>
</div>
@if (zone.transform === 'date') {
<div class="mb-3">
<label class="form-label" i18n>Date format</label>
<select class="form-select" [ngModel]="dateFormatChoice(zone)" (ngModelChange)="setDateFormatChoice(zone, $event)">
@for (opt of dateFormatOptions; track opt.id) {
<option [ngValue]="opt.id">{{ opt.name }}</option>
}
<option [ngValue]="'custom'" i18n>Custom...</option>
</select>
@if (dateFormatCustom) {
<div class="input-group mt-2">
<input type="text" class="form-control font-monospace" [(ngModel)]="zone.date_format" placeholder="%d.%m.%Y" />
<button class="btn btn-outline-secondary" type="button" [ngbPopover]="dateFmtHelp" [autoClose]="true" title="Date format help" i18n-title>
<i-bs name="question-circle"></i-bs>
</button>
</div>
<ng-template #dateFmtHelp>
<p class="mb-1" i18n>Python date codes:</p>
<ul class="mb-1 ps-3">
<li><code>%d</code> <ng-container i18n>day (01-31)</ng-container></li>
<li><code>%m</code> <ng-container i18n>month (01-12)</ng-container></li>
<li><code>%Y</code> <ng-container i18n>year, 4-digit</ng-container></li>
<li><code>%y</code> <ng-container i18n>year, 2-digit</ng-container></li>
<li><code>%b</code> <ng-container i18n>month name (Jan)</ng-container></li>
</ul>
<span i18n>Example:</span> <code>%d.%m.%Y</code> -> 03.03.2026
</ng-template>
}
</div>
}
<div class="mb-3">
<label class="form-label" i18n>Validation Regex</label>
<input
type="text"
class="form-control font-monospace"
[(ngModel)]="zone.validation_regex"
placeholder="e.g. \d{2}\.\d{2}\.\d{4}"
>
</div>
<div class="text-muted small">
{{ zone.x }}, {{ zone.y }} - {{ zone.width }}x{{ zone.height }}px
</div>
<hr class="my-3" />
<h6 i18n>Test</h6>
@if (!previewDocId) {
<p class="text-muted small mb-0" i18n>
Load a document in the Settings tab to test this zone.
</p>
} @else {
<button class="btn btn-sm btn-outline-secondary" (click)="testZone()" [disabled]="zoneTesting">
@if (zoneTesting) {
<span class="spinner-border spinner-border-sm me-1"></span>
}
<span i18n>Test this zone</span>
</button>
@if (zoneTestResult) {
@if (zoneTestResult.error) {
<div class="alert alert-warning py-2 mt-2 mb-0 small">{{ zoneTestResult.error }}</div>
} @else {
<dl class="row small mt-2 mb-0">
<dt class="col-sm-4" i18n>OCR text</dt>
<dd class="col-sm-8"><code>{{ zoneTestResult.raw_text || '(nothing detected)' }}</code></dd>
<dt class="col-sm-4" i18n>Value</dt>
<dd class="col-sm-8"><code>{{ zoneTestResult.value || '(empty)' }}</code></dd>
@if (zoneTestResult.regex) {
<dt class="col-sm-4" i18n>Validation</dt>
<dd class="col-sm-8">
@if (zoneTestResult.regex_match) {
<span class="badge bg-success" i18n>Regex matches</span>
} @else {
<span class="badge bg-danger" i18n>Regex does not match</span>
}
</dd>
}
</dl>
}
}
}
} @else {
<p class="text-muted" i18n>
Select a zone from the Zones tab, or draw a rectangle on the document to create one.
</p>
}
</ng-template>
</li>
</ul>
<div [ngbNavOutlet]="nav" class="mt-3"></div>
</div>
<!-- Right column: Document preview with zone overlay -->
<div class="col-md-8">
@if (pageImageUrl) {
<div class="border" style="overflow: auto; max-height: 78vh;">
<div class="position-relative d-inline-block" [style.width.%]="zoom * 100">
<img
#pageImage
[src]="pageImageUrl"
(load)="onImageLoad()"
style="width: 100%; display: block;"
[style.visibility]="imageLoaded ? 'visible' : 'hidden'"
crossorigin="use-credentials"
/>
@if (imageLoaded) {
<canvas
#zoneCanvas
class="position-absolute top-0 start-0"
style="width: 100%; height: 100%; cursor: crosshair;"
(mousedown)="onCanvasMouseDown($event)"
(mousemove)="onCanvasMouseMove($event)"
(mouseup)="onCanvasMouseUp($event)"
></canvas>
}
@if (!imageLoaded) {
<div class="d-flex justify-content-center p-5">
<div class="spinner-border" role="status">
<span class="visually-hidden" i18n>Loading page...</span>
</div>
</div>
}
</div>
</div>
} @else {
<div class="border rounded p-5 text-center text-muted">
<i-bs name="file-earmark-image" width="48" height="48"></i-bs>
<p class="mt-3" i18n>
Enter a document ID and click "Load" to preview a page and draw extraction zones.
</p>
</div>
}
</div>
</div>
@@ -0,0 +1,3 @@
:host {
display: block;
}
@@ -0,0 +1,997 @@
import { CommonModule } from '@angular/common'
import {
AfterViewInit,
Component,
ElementRef,
inject,
OnDestroy,
OnInit,
ViewChild,
} from '@angular/core'
import { FormsModule } from '@angular/forms'
import { ActivatedRoute, Router, RouterModule } from '@angular/router'
import {
NgbNavModule,
NgbPopoverModule,
NgbTypeaheadModule,
NgbTypeaheadSelectItemEvent,
} from '@ng-bootstrap/ng-bootstrap'
import { NgSelectModule } from '@ng-select/ng-select'
import { NgxBootstrapIconsModule } from 'ngx-bootstrap-icons'
import {
catchError,
debounceTime,
distinctUntilChanged,
map,
Observable,
of,
Subject,
switchMap,
takeUntil,
} from 'rxjs'
import { SelectComponent } from 'src/app/components/common/input/select/select.component'
import { SwitchComponent } from 'src/app/components/common/input/switch/switch.component'
import { TextComponent } from 'src/app/components/common/input/text/text.component'
import { PageHeaderComponent } from 'src/app/components/common/page-header/page-header.component'
import { CustomField } from 'src/app/data/custom-field'
import { Document } from 'src/app/data/document'
import { DocumentType } from 'src/app/data/document-type'
import {
DATE_FORMAT_OPTIONS,
OCR_BUILTIN_TARGETS,
OCR_LANGUAGE_OPTIONS,
OcrTemplate,
OcrTemplateZone,
OcrZoneTestResult,
TRANSFORM_OPTIONS,
ZoneTestRequest,
} from 'src/app/data/ocr-template'
import { CorrespondentService } from 'src/app/services/rest/correspondent.service'
import { CustomFieldsService } from 'src/app/services/rest/custom-fields.service'
import { DocumentTypeService } from 'src/app/services/rest/document-type.service'
import { DocumentService } from 'src/app/services/rest/document.service'
import { OcrTemplateService } from 'src/app/services/rest/ocr-template.service'
import { ToastService } from 'src/app/services/toast.service'
interface DrawingRect {
startX: number
startY: number
endX: number
endY: number
}
type ResizeHandle = 'n' | 's' | 'e' | 'w' | 'ne' | 'nw' | 'se' | 'sw'
type ActiveTab = 'settings' | 'zones' | 'zone'
@Component({
selector: 'pngx-ocr-template-editor',
standalone: true,
imports: [
PageHeaderComponent,
TextComponent,
SelectComponent,
SwitchComponent,
CommonModule,
FormsModule,
RouterModule,
NgbNavModule,
NgbPopoverModule,
NgbTypeaheadModule,
NgSelectModule,
NgxBootstrapIconsModule,
],
templateUrl: './ocr-template-editor.component.html',
styleUrls: ['./ocr-template-editor.component.scss'],
})
export class OcrTemplateEditorComponent
implements OnInit, OnDestroy, AfterViewInit
{
private readonly route = inject(ActivatedRoute)
private readonly router = inject(Router)
private readonly templateService = inject(OcrTemplateService)
private readonly customFieldsService = inject(CustomFieldsService)
private readonly documentTypeService = inject(DocumentTypeService)
private readonly correspondentService = inject(CorrespondentService)
private readonly documentService = inject(DocumentService)
private readonly toastService = inject(ToastService)
private readonly destroy$ = new Subject<void>()
@ViewChild('zoneCanvas') canvasRef: ElementRef<HTMLCanvasElement>
@ViewChild('pageImage') imageRef: ElementRef<HTMLImageElement>
template: OcrTemplate = {
id: null,
name: '',
document_type: null,
sample_document: null,
source_width: 0,
source_height: 0,
enabled: true,
combine_formats: {},
zones: [],
}
customFields: CustomField[] = []
documentTypes: DocumentType[] = []
transformOptions = TRANSFORM_OPTIONS
builtinTargets = OCR_BUILTIN_TARGETS
dateFormatOptions = DATE_FORMAT_OPTIONS
ocrLanguageOptions = OCR_LANGUAGE_OPTIONS
dateFormatCustom = false
isNew = true
saving = false
previewDocId: number | null = null
previewPage = 0
previewPageCount: number | null = null
private pageCountForDoc: number | null = null
pageImageUrl: string | null = null
imageLoaded = false
zoom = 1
previewDocModel: Document | string = ''
private correspondentNames = new Map<number, string>()
public get previewPageDisplay(): number {
return this.previewPage + 1
}
public set previewPageDisplay(value: number) {
this.previewPage = Math.max(0, value) - 1
}
activeTab: ActiveTab = 'settings'
isDrawing = false
currentRect: DrawingRect | null = null
selectedZoneIndex: number | null = null
isResizing = false
resizeHandle: ResizeHandle | null = null
resizeZoneIndex: number | null = null
private readonly HANDLE_SIZE = 8
isMoving = false
moveZoneIndex: number | null = null
private moveStart = { mouseX: 0, mouseY: 0, zoneX: 0, zoneY: 0 }
zoneTestResult: OcrZoneTestResult | null = null
zoneTesting = false
showQuickCreate = false
quickCreateName = ''
quickCreateType = 'string'
quickCreateForZoneIndex: number | null = null
quickCreateTypes = [
{ id: 'string', name: $localize`String` },
{ id: 'integer', name: $localize`Integer` },
{ id: 'float', name: $localize`Float` },
{ id: 'date', name: $localize`Date` },
{ id: 'monetary', name: $localize`Monetary` },
{ id: 'boolean', name: $localize`Boolean` },
{ id: 'url', name: $localize`URL` },
{ id: 'longtext', name: $localize`Long Text` },
]
get selectedZone(): OcrTemplateZone | null {
return this.selectedZoneIndex !== null
? (this.template.zones[this.selectedZoneIndex] ?? null)
: null
}
get pageTitle(): string {
return this.isNew
? $localize`New OCR Template`
: $localize`Edit OCR Template`
}
ngOnInit() {
this.customFieldsService
.listAll()
.pipe(takeUntil(this.destroy$))
.subscribe((r) => (this.customFields = r.results))
this.documentTypeService
.listAll()
.pipe(takeUntil(this.destroy$))
.subscribe((r) => (this.documentTypes = r.results))
this.correspondentService
.listAll()
.pipe(takeUntil(this.destroy$))
.subscribe((r) => {
this.correspondentNames = new Map(r.results.map((c) => [c.id, c.name]))
})
const id = this.route.snapshot.paramMap.get('id')
if (id && id !== 'new') {
this.isNew = false
this.templateService
.get(parseInt(id))
.pipe(takeUntil(this.destroy$))
.subscribe((t) => {
this.template = t
this.template.combine_formats ??= {}
if (t.sample_document) {
this.previewDocId = t.sample_document
this.loadPreview()
}
})
} else {
const qp = this.route.snapshot.queryParams
if (qp['document_type']) {
this.template.document_type = parseInt(qp['document_type'])
}
if (qp['sample_document']) {
const docId = parseInt(qp['sample_document'])
this.template.sample_document = docId
this.previewDocId = docId
this.loadPreview()
}
}
}
ngAfterViewInit() {}
searchDocuments = (text$: Observable<string>): Observable<Document[]> =>
text$.pipe(
debounceTime(250),
distinctUntilChanged(),
switchMap((term) => {
if (!term || term.trim().length < 2) return of([])
const params: { title__icontains: string; document_type__id?: number } =
{ title__icontains: term.trim() }
if (this.template.document_type) {
params['document_type__id'] = this.template.document_type
}
return this.documentService.list(1, 10, 'created', true, params).pipe(
map((r) => r.results),
catchError(() => of([]))
)
})
)
documentFormatter = (doc: Document | string): string => {
if (typeof doc === 'string') return doc
const corr = doc.correspondent
? this.correspondentNames.get(doc.correspondent)
: null
return corr
? `#${doc.id} ${doc.title} (${corr})`
: `#${doc.id} ${doc.title}`
}
onPreviewDocSelected(event: NgbTypeaheadSelectItemEvent<Document>) {
event.preventDefault()
const doc: Document = event.item
this.previewDocModel = doc
this.previewDocId = doc.id
if (!this.template.document_type && doc.document_type) {
this.template.document_type = doc.document_type
}
this.previewPage = 0
this.loadPreview()
}
clearPreviewDoc() {
this.previewDocModel = ''
this.previewDocId = null
this.previewPageCount = null
this.pageCountForDoc = null
this.previewPage = 0
this.pageImageUrl = null
this.imageLoaded = false
}
loadPreview() {
if (!this.previewDocId) return
if (this.pageCountForDoc !== this.previewDocId) {
this.pageCountForDoc = this.previewDocId
this.previewPageCount = null
this.documentService
.get(this.previewDocId)
.pipe(takeUntil(this.destroy$))
.subscribe({
next: (doc) => {
this.previewPageCount = doc?.page_count ?? null
if (doc && !this.previewDocModel) this.previewDocModel = doc
},
error: () => (this.previewPageCount = null),
})
}
this.pageImageUrl = this.templateService.getPageImageUrl(
this.previewDocId,
this.previewPage
)
this.imageLoaded = false
}
goToPage(page: number) {
const max = this.previewPageCount ? this.previewPageCount - 1 : page
const clamped = Math.max(0, Math.min(page, max))
if (clamped === this.previewPage) return
this.previewPage = clamped
this.loadPreview()
}
prevPage() {
this.goToPage(this.previewPage - 1)
}
nextPage() {
this.goToPage(this.previewPage + 1)
}
zoomIn() {
this.zoom = Math.min(4, Math.round((this.zoom + 0.25) * 100) / 100)
this.afterZoom()
}
zoomOut() {
this.zoom = Math.max(0.5, Math.round((this.zoom - 0.25) * 100) / 100)
this.afterZoom()
}
resetZoom() {
this.zoom = 1
this.afterZoom()
}
private afterZoom() {
// Defer so the wrapper reflows to the new width before the canvas resizes.
setTimeout(() => this.redrawCanvas())
}
zonePage(zone: OcrTemplateZone): number {
const v = zone.page ?? 1
if (v === -1) return this.previewPageCount ?? this.previewPage + 1
return v >= 1 ? v : 1
}
private isOnCurrentPage(zone: OcrTemplateZone): boolean {
return this.zonePage(zone) === this.previewPage + 1
}
onImageLoad() {
this.imageLoaded = true
const img = this.imageRef.nativeElement
this.template.source_width = img.naturalWidth
this.template.source_height = img.naturalHeight
// The canvas only exists after @if(imageLoaded) renders, so defer the draw.
setTimeout(() => this.redrawCanvas())
}
onCanvasMouseDown(event: MouseEvent) {
const rect = this.canvasRef.nativeElement.getBoundingClientRect()
const x = event.clientX - rect.left
const y = event.clientY - rect.top
if (this.selectedZoneIndex !== null) {
const handle = this.findHandleAt(x, y, this.selectedZoneIndex)
if (handle) {
this.isResizing = true
this.resizeHandle = handle
this.resizeZoneIndex = this.selectedZoneIndex
return
}
}
const clickedIdx = this.findZoneAt(x, y)
if (clickedIdx !== null && !event.shiftKey) {
this.selectZone(clickedIdx)
const zone = this.template.zones[clickedIdx]
this.isMoving = true
this.moveZoneIndex = clickedIdx
this.moveStart = { mouseX: x, mouseY: y, zoneX: zone.x, zoneY: zone.y }
return
}
// Shift+click or click on empty area starts a new zone.
this.isDrawing = true
this.currentRect = { startX: x, startY: y, endX: x, endY: y }
this.selectedZoneIndex = null
}
onCanvasMouseMove(event: MouseEvent) {
const rect = this.canvasRef.nativeElement.getBoundingClientRect()
const mx = event.clientX - rect.left
const my = event.clientY - rect.top
if (this.isResizing && this.resizeZoneIndex !== null && this.resizeHandle) {
this.applyResize(mx, my)
this.redrawCanvas()
return
}
if (this.isMoving && this.moveZoneIndex !== null) {
const zone = this.template.zones[this.moveZoneIndex]
const canvas = this.canvasRef.nativeElement
const img = this.imageRef.nativeElement
const srcW = zone.zone_source_width || img.naturalWidth
const srcH = zone.zone_source_height || img.naturalHeight
const scaleX = srcW / canvas.width
const scaleY = srcH / canvas.height
const dx = Math.round((mx - this.moveStart.mouseX) * scaleX)
const dy = Math.round((my - this.moveStart.mouseY) * scaleY)
zone.x = Math.max(
0,
Math.min(this.moveStart.zoneX + dx, srcW - zone.width)
)
zone.y = Math.max(
0,
Math.min(this.moveStart.zoneY + dy, srcH - zone.height)
)
this.redrawCanvas()
return
}
if (this.isDrawing && this.currentRect) {
this.currentRect.endX = mx
this.currentRect.endY = my
this.redrawCanvas()
return
}
// Cursor feedback: resize handle > move (over a zone) > crosshair.
const canvas = this.canvasRef.nativeElement
if (this.selectedZoneIndex !== null) {
const handle = this.findHandleAt(mx, my, this.selectedZoneIndex)
if (handle) {
const cursorMap: Record<ResizeHandle, string> = {
nw: 'nw-resize',
ne: 'ne-resize',
sw: 'sw-resize',
se: 'se-resize',
n: 'n-resize',
s: 's-resize',
w: 'w-resize',
e: 'e-resize',
}
canvas.style.cursor = cursorMap[handle] || 'crosshair'
return
}
}
canvas.style.cursor =
this.findZoneAt(mx, my) !== null ? 'move' : 'crosshair'
}
onCanvasMouseUp(event: MouseEvent) {
if (this.isMoving) {
this.isMoving = false
this.moveZoneIndex = null
return
}
if (this.isResizing) {
this.isResizing = false
this.resizeHandle = null
this.resizeZoneIndex = null
return
}
if (!this.isDrawing || !this.currentRect) return
this.isDrawing = false
const canvas = this.canvasRef.nativeElement
const img = this.imageRef.nativeElement
const scaleX = img.naturalWidth / canvas.width
const scaleY = img.naturalHeight / canvas.height
const x = Math.round(
Math.min(this.currentRect.startX, this.currentRect.endX) * scaleX
)
const y = Math.round(
Math.min(this.currentRect.startY, this.currentRect.endY) * scaleY
)
const w = Math.round(
Math.abs(this.currentRect.endX - this.currentRect.startX) * scaleX
)
const h = Math.round(
Math.abs(this.currentRect.endY - this.currentRect.startY) * scaleY
)
// Ignore tiny accidental clicks.
if (w < 10 || h < 10) {
this.currentRect = null
this.redrawCanvas()
return
}
const zone: OcrTemplateZone = {
name: `Zone ${this.template.zones.length + 1}`,
target: 'custom_field',
custom_field:
this.customFields.length > 0 ? this.customFields[0].id : null,
x,
y,
width: w,
height: h,
page: this.previewPage + 1,
ocr_language: 'deu+eng',
transform: 'strip',
date_format: '',
validation_regex: '',
order: this.template.zones.length,
zone_source_width: img.naturalWidth,
zone_source_height: img.naturalHeight,
}
this.template.zones.push(zone)
this.currentRect = null
this.selectZone(this.template.zones.length - 1)
}
private getZoneDisplayRect(
zoneIdx: number
): { x: number; y: number; w: number; h: number } | null {
const canvas = this.canvasRef?.nativeElement
const img = this.imageRef?.nativeElement
if (!canvas || !img || !img.naturalWidth) return null
const zone = this.template.zones[zoneIdx]
if (!zone) return null
if (!this.isOnCurrentPage(zone)) return null
const srcW = zone.zone_source_width || img.naturalWidth
const srcH = zone.zone_source_height || img.naturalHeight
const scaleX = canvas.width / srcW
const scaleY = canvas.height / srcH
return {
x: zone.x * scaleX,
y: zone.y * scaleY,
w: zone.width * scaleX,
h: zone.height * scaleY,
}
}
private findHandleAt(
mx: number,
my: number,
zoneIdx: number
): ResizeHandle | null {
const r = this.getZoneDisplayRect(zoneIdx)
if (!r) return null
const hs = this.HANDLE_SIZE
const handles: [ResizeHandle, number, number][] = [
['nw', r.x, r.y],
['n', r.x + r.w / 2, r.y],
['ne', r.x + r.w, r.y],
['w', r.x, r.y + r.h / 2],
['e', r.x + r.w, r.y + r.h / 2],
['sw', r.x, r.y + r.h],
['s', r.x + r.w / 2, r.y + r.h],
['se', r.x + r.w, r.y + r.h],
]
for (const [name, hx, hy] of handles) {
if (Math.abs(mx - hx) <= hs && Math.abs(my - hy) <= hs) return name
}
return null
}
private applyResize(mx: number, my: number) {
const canvas = this.canvasRef.nativeElement
const img = this.imageRef.nativeElement
const zone = this.template.zones[this.resizeZoneIndex]
if (!zone) return
const srcW = zone.zone_source_width || img.naturalWidth
const srcH = zone.zone_source_height || img.naturalHeight
const scaleX = srcW / canvas.width
const scaleY = srcH / canvas.height
const imgX = Math.round(mx * scaleX)
const imgY = Math.round(my * scaleY)
const handle = this.resizeHandle
if (handle.includes('w')) {
const right = zone.x + zone.width
zone.x = Math.max(0, Math.min(imgX, right - 10))
zone.width = right - zone.x
}
if (handle.includes('e')) {
zone.width = Math.max(10, imgX - zone.x)
}
if (handle.includes('n')) {
const bottom = zone.y + zone.height
zone.y = Math.max(0, Math.min(imgY, bottom - 10))
zone.height = bottom - zone.y
}
if (handle.includes('s')) {
zone.height = Math.max(10, imgY - zone.y)
}
}
private findZoneAt(displayX: number, displayY: number): number | null {
const canvas = this.canvasRef.nativeElement
const img = this.imageRef.nativeElement
if (!img.naturalWidth) return null
for (let i = this.template.zones.length - 1; i >= 0; i--) {
const z = this.template.zones[i]
if (!this.isOnCurrentPage(z)) continue
const srcW = z.zone_source_width || img.naturalWidth
const srcH = z.zone_source_height || img.naturalHeight
const scaleX = canvas.width / srcW
const scaleY = canvas.height / srcH
const zx = z.x * scaleX
const zy = z.y * scaleY
const zw = z.width * scaleX
const zh = z.height * scaleY
if (
displayX >= zx &&
displayX <= zx + zw &&
displayY >= zy &&
displayY <= zy + zh
) {
return i
}
}
return null
}
redrawCanvas() {
if (!this.canvasRef || !this.imageRef) return
const canvas = this.canvasRef.nativeElement
const img = this.imageRef.nativeElement
const ctx = canvas.getContext('2d')
canvas.width = img.clientWidth
canvas.height = img.clientHeight
ctx.clearRect(0, 0, canvas.width, canvas.height)
const colors = [
'#4f8ff7',
'#ff6b6b',
'#51cf66',
'#ffd43b',
'#cc5de8',
'#ff922b',
'#20c997',
'#e599f7',
]
this.template.zones.forEach((zone, idx) => {
if (!this.isOnCurrentPage(zone)) return
const color = colors[idx % colors.length]
const srcW = zone.zone_source_width || img.naturalWidth
const srcH = zone.zone_source_height || img.naturalHeight
const scaleX = canvas.width / srcW
const scaleY = canvas.height / srcH
const x = zone.x * scaleX
const y = zone.y * scaleY
const w = zone.width * scaleX
const h = zone.height * scaleY
ctx.strokeStyle = color
ctx.lineWidth = idx === this.selectedZoneIndex ? 3 : 2
ctx.strokeRect(x, y, w, h)
ctx.fillStyle = color + '20'
ctx.fillRect(x, y, w, h)
const label = zone.name || `Zone ${idx + 1}`
ctx.font = '12px sans-serif'
ctx.textBaseline = 'middle'
const padX = 6
const pillH = 17
const pillW = ctx.measureText(label).width + padX * 2
const pillX = x
const pillY = Math.max(0, y - pillH - 2)
const r = 4
ctx.fillStyle = color
ctx.beginPath()
ctx.moveTo(pillX + r, pillY)
ctx.arcTo(pillX + pillW, pillY, pillX + pillW, pillY + pillH, r)
ctx.arcTo(pillX + pillW, pillY + pillH, pillX, pillY + pillH, r)
ctx.arcTo(pillX, pillY + pillH, pillX, pillY, r)
ctx.arcTo(pillX, pillY, pillX + pillW, pillY, r)
ctx.closePath()
ctx.fill()
ctx.fillStyle = '#ffffff'
ctx.fillText(label, pillX + padX, pillY + pillH / 2 + 0.5)
ctx.textBaseline = 'alphabetic'
if (idx === this.selectedZoneIndex) {
const hs = this.HANDLE_SIZE
ctx.fillStyle = color
const handles = [
[x, y],
[x + w / 2, y],
[x + w, y],
[x, y + h / 2],
[x + w, y + h / 2],
[x, y + h],
[x + w / 2, y + h],
[x + w, y + h],
]
for (const [hx, hy] of handles) {
ctx.fillRect(hx - hs / 2, hy - hs / 2, hs, hs)
}
}
})
if (this.currentRect) {
const cw = this.currentRect.endX - this.currentRect.startX
const ch = this.currentRect.endY - this.currentRect.startY
ctx.fillStyle = 'rgba(105, 219, 124, 0.25)'
ctx.fillRect(this.currentRect.startX, this.currentRect.startY, cw, ch)
ctx.strokeStyle = '#69db7c'
ctx.lineWidth = 2
ctx.setLineDash([5, 5])
ctx.strokeRect(this.currentRect.startX, this.currentRect.startY, cw, ch)
ctx.setLineDash([])
}
}
removeZone(index: number) {
this.template.zones.splice(index, 1)
if (this.selectedZoneIndex === index) {
this.selectedZoneIndex = null
} else if (this.selectedZoneIndex > index) {
this.selectedZoneIndex--
}
this.redrawCanvas()
}
selectZone(index: number) {
this.selectedZoneIndex = index
this.activeTab = 'zone'
this.zoneTestResult = null
const zone = this.template.zones[index]
if (zone) {
this.dateFormatCustom =
!!zone.date_format &&
!this.dateFormatOptions.some((o) => o.id === zone.date_format)
this.seedCombineDefault(zone)
this.goToPage(this.zonePage(zone) - 1)
}
this.redrawCanvas()
}
testZone() {
const zone = this.selectedZone
if (!zone || !this.previewDocId) return
this.zoneTesting = true
this.zoneTestResult = null
const payload: ZoneTestRequest = {
name: zone.name,
x: zone.x,
y: zone.y,
width: zone.width,
height: zone.height,
page: zone.page ?? 1,
ocr_language: zone.ocr_language,
transform: zone.transform,
date_format: zone.date_format,
validation_regex: zone.validation_regex,
zone_source_width: zone.zone_source_width,
zone_source_height: zone.zone_source_height,
}
this.templateService
.testZone(this.previewDocId, payload)
.pipe(takeUntil(this.destroy$))
.subscribe({
next: (res) => {
this.zoneTestResult = res
this.zoneTesting = false
},
error: (err) => {
this.zoneTestResult = {
error: err.error?.error || $localize`Test failed`,
}
this.zoneTesting = false
},
})
}
deleteSelectedZone() {
if (this.selectedZoneIndex === null) return
this.removeZone(this.selectedZoneIndex)
this.activeTab = 'zones'
}
save() {
this.saving = true
this.pruneCombineFormats()
this.template.sample_document = this.previewDocId
const obs = this.isNew
? this.templateService.create(this.template)
: this.templateService.update(this.template)
obs.pipe(takeUntil(this.destroy$)).subscribe({
next: (saved) => {
const idx = this.selectedZoneIndex
this.template = saved
this.isNew = false
this.selectedZoneIndex = idx
this.saving = false
this.toastService.showInfo($localize`OCR template saved.`)
this.redrawCanvas()
},
error: (e) => {
this.saving = false
this.toastService.showError($localize`Error saving OCR template.`, e)
},
})
}
private ocrLangCache = new WeakMap<
OcrTemplateZone,
{ src: string; arr: string[] }
>()
ocrLanguageArray(zone: OcrTemplateZone): string[] {
const src = zone.ocr_language || ''
const cached = this.ocrLangCache.get(zone)
if (cached && cached.src === src) return cached.arr
const arr = src ? src.split('+').filter(Boolean) : []
this.ocrLangCache.set(zone, { src, arr })
return arr
}
setOcrLanguages(zone: OcrTemplateZone, langs: string[]) {
zone.ocr_language = (langs || []).join('+')
this.ocrLangCache.set(zone, {
src: zone.ocr_language,
arr: langs ? [...langs] : [],
})
}
getCustomFieldName(id: number): string {
const cf = this.customFields.find((f) => f.id === id)
return cf ? cf.name : `Field #${id}`
}
/** Value bound to the field select: a built-in id string or a custom-field id. */
zoneFieldValue(zone: OcrTemplateZone): number | string | null {
const target = zone.target || 'custom_field'
return target === 'custom_field' ? zone.custom_field : target
}
setZoneField(zone: OcrTemplateZone, value: number | string) {
if (value === 'title' || value === 'asn' || value === 'created') {
zone.target = value
zone.custom_field = null
} else {
zone.target = 'custom_field'
zone.custom_field = typeof value === 'number' ? value : null
}
this.seedCombineDefault(zone)
}
fieldKeyFor(zone: OcrTemplateZone): string | null {
const v = this.zoneFieldValue(zone)
return v === null || v === undefined || v === '' ? null : String(v)
}
zonesForField(zone: OcrTemplateZone): OcrTemplateZone[] {
const key = this.fieldKeyFor(zone)
if (!key) return []
return this.template.zones.filter((z) => this.fieldKeyFor(z) === key)
}
isFieldShared(zone: OcrTemplateZone): boolean {
return this.zonesForField(zone).length > 1
}
getCombineFormat(zone: OcrTemplateZone): string {
const key = this.fieldKeyFor(zone)
return (key && this.template.combine_formats?.[key]) || ''
}
setCombineFormat(zone: OcrTemplateZone, value: string) {
const key = this.fieldKeyFor(zone)
if (!key) return
this.template.combine_formats ??= {}
this.template.combine_formats[key] = value
}
insertCombineToken(zone: OcrTemplateZone, tokenZone: OcrTemplateZone) {
const token = `{${tokenZone.name}}`
const current = this.getCombineFormat(zone)
const sep = current && !current.endsWith(' ') ? ' ' : ''
this.setCombineFormat(zone, `${current}${sep}${token}`)
}
private seedCombineDefault(zone: OcrTemplateZone) {
const key = this.fieldKeyFor(zone)
if (!key) return
const shared = this.zonesForField(zone)
if (shared.length <= 1) return
this.template.combine_formats ??= {}
if (!this.template.combine_formats[key]) {
this.template.combine_formats[key] = shared
.map((z) => `{${z.name}}`)
.join(' ')
}
}
private pruneCombineFormats() {
const formats = this.template.combine_formats
if (!formats) return
const counts = new Map<string, number>()
for (const z of this.template.zones) {
const key = this.fieldKeyFor(z)
if (key) counts.set(key, (counts.get(key) ?? 0) + 1)
}
for (const key of Object.keys(formats)) {
if ((counts.get(key) ?? 0) <= 1) delete formats[key]
}
}
/** Value bound to the date-format select: a preset, '' (auto), or 'custom'. */
dateFormatChoice(zone: OcrTemplateZone): string {
if (this.dateFormatCustom) return 'custom'
return zone.date_format || ''
}
setDateFormatChoice(zone: OcrTemplateZone, value: string) {
if (value === 'custom') {
this.dateFormatCustom = true
} else {
this.dateFormatCustom = false
zone.date_format = value
}
}
getZoneTargetName(zone: OcrTemplateZone): string {
const target = zone.target || 'custom_field'
if (target === 'custom_field') {
return zone.custom_field
? this.getCustomFieldName(zone.custom_field)
: $localize`(no field)`
}
return this.builtinTargets.find((t) => t.id === target)?.name ?? target
}
getDocumentTypeName(id: number): string {
const dt = this.documentTypes.find((d) => d.id === id)
return dt ? dt.name : `Type #${id}`
}
openQuickCreate(zoneIndex: number | null) {
if (zoneIndex === null) return
this.quickCreateForZoneIndex = zoneIndex
this.quickCreateName = this.template.zones[zoneIndex]?.name || ''
this.quickCreateType = 'string'
this.showQuickCreate = true
}
cancelQuickCreate() {
this.showQuickCreate = false
this.quickCreateForZoneIndex = null
}
submitQuickCreate() {
if (!this.quickCreateName.trim()) return
this.templateService
.quickCreateField(this.quickCreateName.trim(), this.quickCreateType)
.pipe(takeUntil(this.destroy$))
.subscribe({
next: (result) => {
this.customFieldsService.clearCache()
this.customFieldsService
.listAll()
.pipe(takeUntil(this.destroy$))
.subscribe((r) => {
this.customFields = r.results
if (this.quickCreateForZoneIndex !== null) {
this.template.zones[this.quickCreateForZoneIndex].custom_field =
result.id
this.template.zones[this.quickCreateForZoneIndex].target =
'custom_field'
}
this.showQuickCreate = false
this.quickCreateForZoneIndex = null
})
},
error: (err) => {
alert(err.error?.error || 'Failed to create custom field')
},
})
}
ngOnDestroy() {
this.destroy$.next()
this.destroy$.complete()
}
}
@@ -0,0 +1,75 @@
<pngx-page-header
title="OCR Templates"
i18n-title
info="Define extraction zones on document types to automatically populate custom fields via OCR."
i18n-info
>
<button type="button" class="btn btn-sm btn-outline-primary" (click)="createTemplate()" *pngxIfPermissions="{ action: PermissionAction.Add, type: PermissionType.OcrTemplate }">
<i-bs name="plus-circle" class="me-1"></i-bs><ng-container i18n>Create Template</ng-container>
</button>
</pngx-page-header>
<ul class="list-group">
<li class="list-group-item">
<div class="row">
<div class="col" i18n>Name</div>
<div class="col d-none d-sm-flex" i18n>Document Type</div>
<div class="col d-none d-sm-flex" i18n>Zones</div>
<div class="col" i18n>Status</div>
<div class="col" i18n>Actions</div>
</div>
</li>
@if (loading && templates.length === 0) {
<li class="list-group-item">
<div class="spinner-border spinner-border-sm me-2" role="status"></div>
<ng-container i18n>Loading...</ng-container>
</li>
}
@for (t of templates; track t.id) {
<li class="list-group-item">
<div class="row fade" [class.show]="show">
<div class="col d-flex align-items-center"><button class="btn btn-link p-0 text-start" type="button" (click)="editTemplate(t)" [disabled]="!permissionsService.currentUserCan(PermissionAction.Change, PermissionType.OcrTemplate)">{{t.name}}</button></div>
<div class="col d-flex align-items-center d-none d-sm-flex">{{getDocumentTypeName(t)}}</div>
<div class="col d-flex align-items-center d-none d-sm-flex"><code>{{t.zones?.length || 0}}</code></div>
<div class="col d-flex align-items-center">
<div class="form-check form-switch mb-0">
<input type="checkbox" class="form-check-input cursor-pointer" [id]="t.id+'_enable'" [(ngModel)]="t.enabled" (change)="toggleTemplate(t)" *pngxIfPermissions="{ action: PermissionAction.Change, type: PermissionType.OcrTemplate }">
<label class="form-check-label cursor-pointer" [for]="t.id+'_enable'">
<code> @if(t.enabled) { <ng-container i18n>Enabled</ng-container> } @else { <span i18n class="text-muted">Disabled</span> }</code>
</label>
</div>
</div>
<div class="col">
<div class="btn-group d-block d-sm-none">
<div ngbDropdown container="body" class="d-inline-block">
<button type="button" class="btn btn-link" id="actionsMenuMobile{{t.id}}" (click)="$event.stopPropagation()" ngbDropdownToggle>
<i-bs name="three-dots-vertical"></i-bs>
</button>
<div ngbDropdownMenu aria-labelledby="actionsMenuMobile{{t.id}}">
<button (click)="editTemplate(t)" *pngxIfPermissions="{ action: PermissionAction.Change, type: PermissionType.OcrTemplate }" ngbDropdownItem i18n>Edit</button>
<button (click)="deleteTemplate(t)" *pngxIfPermissions="{ action: PermissionAction.Delete, type: PermissionType.OcrTemplate }" ngbDropdownItem i18n>Delete</button>
</div>
</div>
</div>
<div class="btn-toolbar d-none d-sm-flex gap-2" role="toolbar">
<div class="btn-group">
<button *pngxIfPermissions="{ action: PermissionAction.Change, type: PermissionType.OcrTemplate }" class="btn btn-sm btn-outline-secondary" type="button" (click)="editTemplate(t)">
<i-bs width="1em" height="1em" name="pencil" class="me-1"></i-bs><ng-container i18n>Edit</ng-container>
</button>
<button *pngxIfPermissions="{ action: PermissionAction.Delete, type: PermissionType.OcrTemplate }" class="btn btn-sm btn-outline-danger" type="button" (click)="deleteTemplate(t)">
<i-bs width="1em" height="1em" name="trash" class="me-1"></i-bs><ng-container i18n>Delete</ng-container>
</button>
</div>
</div>
</div>
</div>
</li>
}
@if (!loading && templates.length === 0) {
<li class="list-group-item" [class.show]="show" i18n>No OCR templates defined.</li>
}
</ul>
@@ -0,0 +1,98 @@
import { Component, OnInit, inject } from '@angular/core'
import { FormsModule } from '@angular/forms'
import { Router } from '@angular/router'
import { NgbDropdownModule, NgbModal } from '@ng-bootstrap/ng-bootstrap'
import { NgxBootstrapIconsModule } from 'ngx-bootstrap-icons'
import { delay, takeUntil, tap } from 'rxjs'
import { OcrTemplate } from 'src/app/data/ocr-template'
import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
import { PermissionsService } from 'src/app/services/permissions.service'
import { DocumentTypeService } from 'src/app/services/rest/document-type.service'
import { OcrTemplateService } from 'src/app/services/rest/ocr-template.service'
import { ConfirmDialogComponent } from '../../common/confirm-dialog/confirm-dialog.component'
import { PageHeaderComponent } from '../../common/page-header/page-header.component'
import { LoadingComponentWithPermissions } from '../../loading-component/loading.component'
@Component({
selector: 'pngx-ocr-templates',
templateUrl: './ocr-templates.component.html',
imports: [
PageHeaderComponent,
IfPermissionsDirective,
FormsModule,
NgbDropdownModule,
NgxBootstrapIconsModule,
],
})
export class OcrTemplatesComponent
extends LoadingComponentWithPermissions
implements OnInit
{
private readonly service = inject(OcrTemplateService)
private readonly documentTypeService = inject(DocumentTypeService)
private readonly router = inject(Router)
private readonly modalService = inject(NgbModal)
permissionsService = inject(PermissionsService)
public templates: OcrTemplate[] = []
private documentTypeNames: Map<number, string> = new Map()
ngOnInit() {
this.documentTypeService
.listAll()
.pipe(takeUntil(this.unsubscribeNotifier))
.subscribe((r) => {
this.documentTypeNames = new Map(
r.results.map((dt) => [dt.id, dt.name])
)
})
this.reload()
}
reload() {
this.loading = true
this.service
.listAll()
.pipe(
takeUntil(this.unsubscribeNotifier),
tap((r) => (this.templates = r.results)),
delay(100)
)
.subscribe(() => {
this.show = true
this.loading = false
})
}
getDocumentTypeName(t: OcrTemplate): string {
return (
this.documentTypeNames.get(t.document_type) ?? `${t.document_type ?? ''}`
)
}
createTemplate() {
this.router.navigate(['/ocr-templates', 'new'])
}
editTemplate(t: OcrTemplate) {
this.router.navigate(['/ocr-templates', t.id])
}
toggleTemplate(t: OcrTemplate) {
// ngModel has already flipped t.enabled — just persist it.
this.service.patch(t).subscribe()
}
deleteTemplate(t: OcrTemplate) {
const modal = this.modalService.open(ConfirmDialogComponent)
modal.componentInstance.title = $localize`Delete OCR Template`
modal.componentInstance.messageBoldPart = t.name
modal.componentInstance.message = $localize`Do you really want to delete this OCR template?`
modal.componentInstance.btnClass = 'btn-danger'
modal.componentInstance.btnCaption = $localize`Delete`
modal.componentInstance.confirmClicked.subscribe(() => {
modal.close()
this.service.delete(t).subscribe(() => this.reload())
})
}
}
+102
View File
@@ -0,0 +1,102 @@
import { ObjectWithId } from './object-with-id'
export type OcrZoneTarget = 'custom_field' | 'title' | 'asn' | 'created'
export const OCR_BUILTIN_TARGETS = [
{ id: 'title', name: $localize`Title` },
{ id: 'asn', name: $localize`Archive serial number` },
{ id: 'created', name: $localize`Date created` },
]
export interface OcrTemplateZone {
id?: number
name: string
target?: OcrZoneTarget
custom_field: number | null
page?: number
x: number
y: number
width: number
height: number
ocr_language: string
transform: string
date_format?: string
validation_regex: string
order: number
zone_source_width?: number
zone_source_height?: number
}
export const TRANSFORM_OPTIONS = [
{ id: 'none', name: $localize`None` },
{ id: 'strip', name: $localize`Strip whitespace` },
{ id: 'uppercase', name: $localize`Uppercase` },
{ id: 'lowercase', name: $localize`Lowercase` },
{ id: 'numeric', name: $localize`Numeric only` },
{
id: 'strip_punctuation',
name: $localize`Remove leading/trailing punctuation`,
},
{ id: 'date', name: $localize`Parse date` },
{ id: 'qr_code', name: $localize`Read QR/barcode` },
]
export const OCR_LANGUAGE_OPTIONS = [
{ id: 'eng', name: $localize`English` },
{ id: 'deu', name: $localize`German` },
{ id: 'fra', name: $localize`French` },
{ id: 'ita', name: $localize`Italian` },
{ id: 'spa', name: $localize`Spanish` },
{ id: 'por', name: $localize`Portuguese` },
{ id: 'nld', name: $localize`Dutch` },
]
export const DATE_FORMAT_OPTIONS = [
{ id: '', name: $localize`Auto-detect` },
{ id: '%d.%m.%Y', name: 'DD.MM.YYYY' },
{ id: '%Y/%m/%d', name: 'YYYY/MM/DD' },
{ id: '%d/%m/%Y', name: 'DD/MM/YYYY' },
]
export interface OcrTemplate extends ObjectWithId {
name: string
document_type: number
sample_document: number | null
source_width: number
source_height: number
enabled: boolean
combine_formats?: Record<string, string>
created?: string
updated?: string
zones: OcrTemplateZone[]
}
export interface ZoneTestRequest {
name: string
x: number
y: number
width: number
height: number
page: number
ocr_language: string
transform: string
date_format?: string
validation_regex: string
zone_source_width?: number
zone_source_height?: number
}
export interface OcrZoneTestResult {
raw_text?: string | null
value?: string | null
regex?: string
regex_match?: boolean | null
error?: string
}
export interface OcrZoneRunResult {
template: string
zone: string
custom_field: string
value: string | number | null
}
@@ -28,6 +28,7 @@ export enum PermissionType {
ShareLink = '%s_sharelink',
CustomField = '%s_customfield',
Workflow = '%s_workflow',
OcrTemplate = '%s_ocrtemplate',
ProcessedMail = '%s_processedmail',
GlobalStatistics = '%s_global_statistics',
SystemMonitoring = '%s_system_monitoring',
@@ -12,6 +12,7 @@ import {
import { DocumentMetadata } from 'src/app/data/document-metadata'
import { DocumentSuggestions } from 'src/app/data/document-suggestions'
import { FilterRule } from 'src/app/data/filter-rule'
import { OcrZoneRunResult } from 'src/app/data/ocr-template'
import { Results, SelectionData } from 'src/app/data/results'
import { SETTINGS_KEYS } from 'src/app/data/ui-settings'
import { queryParamsFromFilterRules } from '../../utils/query-params'
@@ -355,6 +356,13 @@ export class DocumentService extends AbstractPaperlessService<Document> {
})
}
runZoneOcr(id: number): Observable<{ results: OcrZoneRunResult[] }> {
return this.http.post<{ results: OcrZoneRunResult[] }>(
this.getResourceUrl(id, 'run-zone-ocr'),
{}
)
}
rotateDocuments(
selection: DocumentSelectionQuery,
degrees: number,
@@ -0,0 +1,47 @@
import { Injectable } from '@angular/core'
import { Observable } from 'rxjs'
import {
OcrTemplate,
OcrZoneTestResult,
ZoneTestRequest,
} from '../../data/ocr-template'
import { AbstractPaperlessService } from './abstract-paperless-service'
export interface QuickCreateFieldResult {
id: number
name: string
data_type: string
created: boolean
}
@Injectable({ providedIn: 'root' })
export class OcrTemplateService extends AbstractPaperlessService<OcrTemplate> {
constructor() {
super()
this.resourceName = 'ocr_templates'
}
getPageImageUrl(docId: number, page: number): string {
return `${this.baseUrl}${this.resourceName}/document-page-image/${docId}/${page}/`
}
testZone(
docId: number,
zone: ZoneTestRequest
): Observable<OcrZoneTestResult> {
return this.http.post<OcrZoneTestResult>(
`${this.baseUrl}${this.resourceName}/test-zone/`,
{ document: docId, zone }
)
}
quickCreateField(
name: string,
dataType: string
): Observable<QuickCreateFieldResult> {
return this.http.post<QuickCreateFieldResult>(
`${this.baseUrl}${this.resourceName}/quick-create-field/`,
{ name, data_type: dataType }
)
}
}
+6
View File
@@ -79,13 +79,16 @@ import {
exclamationTriangleFill,
eye,
fileEarmark,
fileEarmarkBreak,
fileEarmarkCheck,
fileEarmarkDiff,
fileEarmarkFill,
fileEarmarkLock,
fileEarmarkMedical,
fileEarmarkMinus,
fileEarmarkPlus,
fileEarmarkRichtext,
fileEarmarkRuled,
fileText,
files,
filter,
@@ -302,13 +305,16 @@ const icons = {
exclamationTriangleFill,
eye,
fileEarmark,
fileEarmarkBreak,
fileEarmarkCheck,
fileEarmarkDiff,
fileEarmarkFill,
fileEarmarkLock,
fileEarmarkMedical,
fileEarmarkMinus,
fileEarmarkPlus,
fileEarmarkRichtext,
fileEarmarkRuled,
files,
fileText,
filter,