add new endpoints for data pipeline

Author: shenghann

src/api/migrations/0024_rename_speech_speech__adebe1_gin_api_speech_speech__c0087e_gin_and_more.py

@@ -0,0 +1,45 @@
+# Generated by Django 5.1.4 on 2025-01-26 08:41
+
+from django.db import migrations
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("api", "0023_alter_speech_speaker"),
+    ]
+
+    operations = [
+        migrations.RenameIndex(
+            model_name="speech",
+            new_name="api_speech_speech__c0087e_gin",
+            old_name="speech_speech__adebe1_gin",
+        ),
+        migrations.AlterModelTable(
+            name="area",
+            table="api_area",
+        ),
+        migrations.AlterModelTable(
+            name="attendance",
+            table="api_attendance",
+        ),
+        migrations.AlterModelTable(
+            name="author",
+            table="api_author",
+        ),
+        migrations.AlterModelTable(
+            name="authorhistory",
+            table="api_author_history",
+        ),
+        migrations.AlterModelTable(
+            name="parliamentarycycle",
+            table="api_parliamentary_cycle",
+        ),
+        migrations.AlterModelTable(
+            name="sitting",
+            table="api_sitting",
+        ),
+        migrations.AlterModelTable(
+            name="speech",
+            table="api_speech",
+        ),
+    ]

src/api/migrations/0025_alter_author_birth_year_alter_authorhistory_area_and_more.py

@@ -0,0 +1,33 @@
+# Generated by Django 5.1.4 on 2025-02-25 16:18
+
+import django.db.models.deletion
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        (
+            "api",
+            "0024_rename_speech_speech__adebe1_gin_api_speech_speech__c0087e_gin_and_more",
+        ),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name="author",
+            name="birth_year",
+            field=models.IntegerField(null=True),
+        ),
+        migrations.AlterField(
+            model_name="authorhistory",
+            name="area",
+            field=models.ForeignKey(
+                null=True, on_delete=django.db.models.deletion.CASCADE, to="api.area"
+            ),
+        ),
+        migrations.AlterField(
+            model_name="authorhistory",
+            name="party",
+            field=models.CharField(max_length=32, null=True),
+        ),
+    ]

src/api/models.py

@@ -65,7 +65,7 @@ class Author(models.Model):
 
     new_author_id = models.AutoField(primary_key=True)
     name = models.TextField()
-    birth_year = models.IntegerField()
+    birth_year = models.IntegerField(null=True)
     ethnicity = models.TextField()  # possible choices?
     sex = models.CharField(max_length=1, choices=(("m", "male"), ("f", "female")))
 
@@ -89,8 +89,8 @@ class AuthorHistory(models.Model):
 
     record_id = models.AutoField(primary_key=True)
     author = models.ForeignKey(Author, on_delete=models.CASCADE)
-    party = models.CharField(max_length=32)  # map to party table?
-    area = models.ForeignKey(Area, on_delete=models.CASCADE)
+    party = models.CharField(max_length=32, null=True)  # map to party table?
+    area = models.ForeignKey(Area, on_delete=models.CASCADE, null=True)
     # seat = models.ForeignKey(Seat, on_delete=models.CASCADE)
     exec_posts = models.TextField(null=True)
     service_posts = models.TextField(null=True)

src/api/serializers.py

@@ -1,16 +1,94 @@
 from rest_framework import serializers
-from .models import Sitting, ParliamentaryCycle
+from .models import ParliamentaryCycle, Sitting, Speech, AuthorHistory, Author
 
 
 class ParliamentaryCycleSerializer(serializers.ModelSerializer):
     class Meta:
         model = ParliamentaryCycle
-        fields = ["house", "term", "session", "meeting"]
+        fields = [
+            "cycle_id",
+            "start_date",
+            "end_date",
+            "house",
+            "term",
+            "session",
+            "meeting",
+        ]
+        read_only_fields = ["cycle_id"]
 
 
 class SittingSerializer(serializers.ModelSerializer):
-    cycle = ParliamentaryCycleSerializer()
+    cycle_id = serializers.PrimaryKeyRelatedField(
+        source="cycle", queryset=ParliamentaryCycle.objects.all(), write_only=True
+    )
+    cycle = ParliamentaryCycleSerializer(read_only=True)
 
     class Meta:
         model = Sitting
-        fields = ["cycle", "date", "filename"]
+        fields = [
+            "sitting_id",
+            "cycle",
+            "cycle_id",
+            "date",
+            "filename",
+            "has_dataset",
+            "is_final",
+            "speech_data",
+        ]
+        read_only_fields = ["sitting_id"]
+
+
+class SpeechSerializer(serializers.ModelSerializer):
+    sitting_id = serializers.PrimaryKeyRelatedField(
+        source="sitting", queryset=Sitting.objects.all(), write_only=True
+    )
+    speaker_id = serializers.PrimaryKeyRelatedField(
+        source="speaker",
+        queryset=AuthorHistory.objects.all(),
+        write_only=True,
+        required=False,
+        allow_null=True,
+    )
+
+    class Meta:
+        model = Speech
+        fields = [
+            "speech_id",
+            "sitting_id",
+            "index",
+            "speaker_id",
+            "timestamp",
+            "speech",
+            "speech_tokens",
+            "length",
+            "level_1",
+            "level_2",
+            "level_3",
+            "is_annotation",
+        ]
+        read_only_fields = ["speech_id", "speech_vector"]
+
+
+class AuthorSerializer(serializers.ModelSerializer):
+    class Meta:
+        model = Author
+        fields = ["new_author_id", "name", "birth_year", "ethnicity", "sex"]
+
+
+class AuthorHistorySerializer(serializers.ModelSerializer):
+    author = serializers.PrimaryKeyRelatedField(
+        queryset=Author.objects.all(), write_only=True
+    )
+
+    class Meta:
+        model = AuthorHistory
+        fields = [
+            "record_id",
+            "author",
+            "party",
+            "area",
+            "exec_posts",
+            "service_posts",
+            "start_date",
+            "end_date",
+        ]

src/api/urls.py

@@ -5,6 +5,9 @@
     HansardView,
     SearchPlotView,
     AttendanceView,
+    ParliamentaryCycleCreateView,
+    SpeechView,
+    AuthorHistoryView,
     health_check,
 )
 
@@ -13,6 +16,13 @@
     path("search/", SearchResultsList.as_view(), name="search"),
     path("search-plot/", SearchPlotView.as_view(), name="search-plot"),
     path("sitting/", HansardView.as_view(), name="sitting"),
+    path("speech/", SpeechView.as_view(), name="speech"),
     path("attendance/", AttendanceView.as_view(), name="attendance"),
+    path(
+        "parliamentary-cycle/",
+        ParliamentaryCycleCreateView.as_view(),
+        name="parliamentary-cycle-create",
+    ),
+    path("author-history/", AuthorHistoryView.as_view(), name="author-history"),
     path("health/", health_check, name="health"),
 ]