Feature/derived problems

backend/problem/management/commands/import_fracas.py

@@ -43,6 +43,9 @@ def _annotate_section_subsections(tree: ET.ElementTree) -> None:
 
         root = tree.getroot()
 
+        if root is None:
+            raise ValueError("The XML file is empty or malformed.")
+
         for element in root:
             if element.tag == "comment" and element.attrib.get("class") == "section":
                 current_section = element.text.strip() if element.text else None
@@ -107,15 +110,15 @@ def import_fracas_problems(self, fracas_path: str) -> None:
                 )
                 skipped += 1
                 continue
-            hypothesis = Sentence.objects.create(
+            hypothesis = Sentence.objects.update_or_create(
                 text=FracasData._text_from_element(hypothesis_node)
-            )
+            )[0]
 
             premises = []
             premise_nodes = problem.findall("p")
             for node in premise_nodes:
                 if node.text:
-                    premises.append(Sentence.objects.create(text=node.text.strip()))
+                    premises.append(Sentence.objects.update_or_create(text=node.text.strip())[0])
 
 
             entailment_label = self.ENTAILMENT_LABELS.get(

backend/problem/management/commands/import_sick.py

@@ -57,9 +57,8 @@ def import_sick_problems(self, sick_path: str) -> None:
 
                 extra_data = SickData.import_data(problem)
 
-                premise = Sentence.objects.create(text=problem["sentence_A"])
-
-                hypothesis = Sentence.objects.create(text=problem["sentence_B"])
+                premise = Sentence.objects.update_or_create(text=problem["sentence_A"])[0]
+                hypothesis = Sentence.objects.update_or_create(text=problem["sentence_B"])[0]
 
                 problem = Problem.objects.create(
                     dataset=Problem.Dataset.SICK,

backend/problem/management/commands/import_snli.py

@@ -1,4 +1,5 @@
 import csv
+from typing import Literal
 
 from django.core.management.base import BaseCommand
 from tqdm import tqdm
@@ -53,7 +54,9 @@ def handle(self, *args, **options):
 
         self.import_snli_problems(snli_paths)
 
-    def import_snli_problems(self, snli_paths: list[tuple[str, str]]) -> None:
+    def import_snli_problems(
+        self, snli_paths: list[tuple[Literal["dev", "train", "test"], str]]
+    ) -> None:
         """
         Import SNLI 1.0 problems from a list of SNLI TSV files and enter them into the database.
         """
@@ -99,9 +102,13 @@ def import_snli_problems(self, snli_paths: list[tuple[str, str]]) -> None:
 
                         extra_data = SNLIData.import_data(problem, subset)
 
-                        premise = Sentence.objects.create(text=problem["sentence1"])
+                        premise = Sentence.objects.update_or_create(
+                            text=problem["sentence1"]
+                        )[0]
 
-                        hypothesis = Sentence.objects.create(text=problem["sentence2"])
+                        hypothesis = Sentence.objects.update_or_create(
+                            text=problem["sentence2"]
+                        )[0]
 
                         new_problem = Problem.objects.create(
                             dataset=Problem.Dataset.SNLI,

backend/problem/migrations/0005_problem_base.py

@@ -0,0 +1,26 @@
+# Generated by Django 4.2.20 on 2025-09-22 11:47
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("problem", "0004_sentence_remove_problem_premises_knowledgebase_and_more"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="problem",
+            name="base",
+            field=models.ForeignKey(
+                blank=True,
+                help_text="The base problem from which this problem was derived, if any.",
+                null=True,
+                on_delete=django.db.models.deletion.SET_NULL,
+                related_name="derived_problems",
+                to="problem.problem",
+            ),
+        ),
+    ]

backend/problem/models.py

@@ -29,6 +29,15 @@ class EntailmentLabel(models.TextChoices):
         default=Dataset.USER,
     )
 
+    base = models.ForeignKey(
+        "self",
+        on_delete=models.SET_NULL,
+        null=True,
+        blank=True,
+        related_name="derived_problems",
+        help_text="The base problem from which this problem was derived, if any.",
+    )
+
     premises = models.ManyToManyField(
         Sentence,
         related_name="premise_problems",
@@ -77,6 +86,7 @@ def serialize(self) -> dict:
 
         return {
             "id": self.pk,
+            "base": self.base.pk if self.base else None,
             "dataset": self.dataset,
             "premises": [premise.text for premise in self.premises.all()],
             "hypothesis": self.hypothesis.text,

backend/problem/problem_details.py

@@ -1,4 +1,3 @@
-from typing import Optional
 from dataclasses import dataclass
 
 from django.http import QueryDict

backend/problem/serializers.py

@@ -40,6 +40,8 @@ class ProblemInputSerializer(serializers.Serializer):
         many=True, allow_empty=True, help_text="List of knowledge base items"
     )
 
+    base = serializers.IntegerField(required=False, allow_null=True)
+
     def validate_id(self, value):
         """
         Validate that the problem ID exists and belongs to a user problem.
@@ -53,3 +55,12 @@ def validate_id(self, value):
                     f"Problem with ID {value} does not exist or is not a user problem."
                 )
         return value
+
+    def validate_base(self, value):
+        """Validate that the base problem ID exists if provided."""
+        if value is not None:
+            if not Problem.objects.filter(id=value).exists():
+                raise serializers.ValidationError(
+                    f"Base problem with ID {value} does not exist."
+                )
+        return value

backend/problem/views/problem.py

@@ -158,6 +158,7 @@ def create_problem_from_input(parse_input: dict) -> Problem:
     )[0]
 
     problem = Problem.objects.create(
+        base_id=parse_input["base"],
         hypothesis=hypothesis_sentence,
         dataset=Problem.Dataset.USER,
         # TODO: Determine entailment label based on LangPro parser output.
@@ -207,6 +208,16 @@ def update_problem_from_input(parse_input: dict) -> Problem:
     problem.hypothesis = Sentence.objects.get_or_create(
         text=parse_input["hypothesis"],
     )[0]
+
+    if parse_input["base"] is None:
+        problem.base = None
+    else:
+        try:
+            base_problem = Problem.objects.get(id=parse_input["base"])
+        except Problem.DoesNotExist:
+            raise ValueError(f"Cannot find base Problem with ID: {parse_input['base']}")
+        problem.base = base_problem  # type: ignore
+
     problem.save()
 
     premises: list[Sentence] = []

frontend/src/app/annotate/annotate.component.html

@@ -4,52 +4,39 @@
 @let adding = appMode === 'add';
 @let editing = appMode === 'edit';
 
-
 <div class="row mb-4">
     <div class="d-flex flex-column col-4 gap-2">
         @if (browsing) {
         <div class="card text-bg-light p-3">
             <la-navigator />
             <la-search />
         </div>
-        } @else if (adding) {
+        } @else if (adding || editing) {
         <section class="d-flex flex-column gap-2">
+            @if (adding) {
             <p class="text-muted" i18n>
                 You are currently adding a new problem.
             </p>
-            <p class="text-muted" i18n>
-                Hit 'Start Parse' to send your problem data to the parser and
-                inspect the results.
-            </p>
-            <p class="text-muted" i18n>
-                If you are happy with the changes, save the problem by clicking
-                the 'Save' button.
-            </p>
-            <p class="text-muted" i18n>
-                This will add the problem to the database.
-            </p>
-        </section>
-        } @else if (editing) {
-        <section>
+            } @else {
             <p class="text-muted" i18n>
                 You are currently editing an existing problem.
             </p>
+            }
             <p class="text-muted" i18n>
                 Hit 'Start Parse' to send your problem data to the parser and
                 inspect the results.
             </p>
             <p class="text-muted" i18n>
-                If you are happy with the changes, save the changes by clicking
+                If you are happy with the changes, save the problem by clicking
                 the 'Save problem' button.
             </p>
-            @if (isUserProblem$ | async) {
+            @if (adding) {
             <p class="text-muted" i18n>
-                This will update this problem in the database.
+                This will add the problem to the database.
             </p>
             } @else {
             <p class="text-muted" i18n>
-                This will create a new problem in the database, leaving the
-                original problem unchanged.
+                This will update this problem in the database.
             </p>
             }
         </section>

frontend/src/app/annotate/annotate.component.ts

@@ -5,7 +5,7 @@ import { AnnotationInputComponent } from "./annotation-input/annotation-input.co
 import { SearchComponent } from "./search/search.component";
 import { FontAwesomeModule } from "@fortawesome/angular-fontawesome";
 import { faBinoculars, faPlus } from "@fortawesome/free-solid-svg-icons";
-import { ActivatedRoute, Router } from "@angular/router";
+import { ActivatedRoute, Router, ParamMap } from "@angular/router";
 import { ProblemService } from "@/services/problem.service";
 import { combineLatest, distinctUntilChanged, map } from "rxjs";
 import { CommonModule } from "@angular/common";
@@ -56,6 +56,7 @@ export class AnnotateComponent implements OnInit {
             editParam$
         ])
             .pipe(
+                distinctUntilChanged((oldParams, newParams) => this.areParamsEqual(oldParams, newParams)),
                 takeUntilDestroyed(this.destroyRef)
             )
             .subscribe(([params, queryParams, edit]) => {
@@ -70,4 +71,22 @@ export class AnnotateComponent implements OnInit {
     public addProblem(): void {
         this.router.navigate(["/", "annotate", "new"]);
     }
+
+    private areParamsEqual(
+        [oldParams, oldQueryParams, oldEditParam]: [ParamMap, ParamMap, boolean],
+        [newParams, newQueryParams, newEditParam]: [ParamMap, ParamMap, boolean]
+    ): boolean {
+        const compareMaps = (map1: ParamMap, map2: ParamMap) => {
+            if (map1.keys.length !== map2.keys.length) {
+                return false;
+            }
+            return map1.keys.every((key: string) => map1.get(key) === map2.get(key));
+        };
+
+        return (
+            compareMaps(oldParams, newParams) &&
+            compareMaps(oldQueryParams, newQueryParams) &&
+            oldEditParam === newEditParam
+        );
+    }
 }

frontend/src/app/annotate/annotation-input/annotation-input.component.html

@@ -19,7 +19,16 @@
                 (click)="startParse()"
                 i18n-label
             />
-            @if ((viewMode === 'edit' || viewMode === 'add') && userProblem) {
+            @if (appMode === 'browse') {
+                <la-icon-button
+                    [icon]="faCopy"
+                    buttonStyle="secondary"
+                    label="Copy problem"
+                    (click)="copyProblem()"
+                    i18n-label
+                />
+            }
+            @if ((appMode === 'edit' || appMode === 'add') && userProblem) {
             <la-icon-button
                 [icon]="faFloppyDisk"
                 label="Save problem"

frontend/src/app/annotate/annotation-input/annotation-input.component.spec.ts

@@ -49,6 +49,7 @@ describe("AnnotationInputComponent", () => {
         it('should build form with correct structure and values from problem data', () => {
             const mockProblem: Problem = {
                 id: 123,
+                base: null,
                 premises: ["First premise", "Second premise"],
                 hypothesis: "Test hypothesis",
                 entailmentLabel: EntailmentLabel.ENTAILMENT,
@@ -104,6 +105,7 @@ describe("AnnotationInputComponent", () => {
         it('should handle empty premises and kbItems arrays', () => {
             const mockProblem: Problem = {
                 id: 123,
+                base: null,
                 premises: [],
                 hypothesis: "Empty test hypothesis",
                 entailmentLabel: EntailmentLabel.NEUTRAL,
@@ -127,6 +129,7 @@ describe("AnnotationInputComponent", () => {
         it('should create form controls with required validators', () => {
             const mockProblem: Problem = {
                 id: 1,
+                base: null,
                 premises: ["Test premise"],
                 hypothesis: "Test hypothesis",
                 entailmentLabel: EntailmentLabel.CONTRADICTION,
@@ -148,6 +151,7 @@ describe("AnnotationInputComponent", () => {
         it('should navigate when problem ID is different from current route', () => {
             const mockProblem: Problem = {
                 id: 12,
+                base: null,
                 premises: [],
                 hypothesis: "",
                 entailmentLabel: EntailmentLabel.UNKNOWN,
@@ -167,6 +171,7 @@ describe("AnnotationInputComponent", () => {
         it('should not navigate when problem ID matches current route', () => {
             const mockProblem: Problem = {
                 id: 17,
+                base: null,
                 premises: [],
                 hypothesis: "",
                 entailmentLabel: EntailmentLabel.UNKNOWN,