[DRAFT] [scielo-tools-18]: Adiciona API REST para marcar o front SPS 1.10

.envs.example/.local/.django

@@ -18,3 +18,14 @@ REFERENCE_MODEL=llama3.2:3b
 # REFERENCE_TIMEOUT=300
 # REFERENCE_BATCH_SIZE=10
 # REFERENCE_NUM_CTX=8192
+# REFERENCE_KEEP_ALIVE=-1
+
+# Front (Llama via HTTP — serviço ollama no local.yml)
+# ------------------------------------------------------------------------------
+FRONT_ENABLED=true
+FRONT_URL=http://ollama:11434
+FRONT_MODEL=llama3.2:3b
+# FRONT_TOKEN=
+# FRONT_TIMEOUT=300
+# FRONT_NUM_CTX=8192
+# FRONT_KEEP_ALIVE=-1

.envs.example/.production/.django

@@ -29,3 +29,14 @@ REFERENCE_MODEL=llama3.2:3b
 # REFERENCE_TIMEOUT=300
 # REFERENCE_BATCH_SIZE=10
 # REFERENCE_NUM_CTX=8192
+# REFERENCE_KEEP_ALIVE=-1
+
+# Front (Llama via HTTP — Ollama-compatible API)
+# ------------------------------------------------------------------------------
+FRONT_ENABLED=true
+FRONT_URL=
+FRONT_MODEL=llama3.2:3b
+# FRONT_TOKEN=
+# FRONT_TIMEOUT=300
+# FRONT_NUM_CTX=8192
+# FRONT_KEEP_ALIVE=-1

README.md

@@ -46,6 +46,7 @@ make down                  # para os containers
 | `core/` | Modelos base, Wagtail home, templates e static |
 | `core_settings/` | Configurações editáveis do site (nome, logo, favicon) |
 | `users/` | `CustomUser` (`AUTH_USER_MODEL`) |
+| `front/` | Marcação do front do artigo (API REST, Llama local, XML SPS 1.10) |
 | `compose/` | Dockerfiles e scripts de inicialização |
 | `requirements/` | Dependências Python (base, local, production) |
 

config/api_router.py

@@ -1,6 +1,7 @@
 from django.conf import settings
 from rest_framework.routers import DefaultRouter, SimpleRouter
 
+from front.api.v1.views import FrontViewSet
 from reference.api.v1.views import ReferenceViewSet
 
 if settings.DEBUG:
@@ -9,5 +10,6 @@
     router = SimpleRouter()
 
 router.register("reference", ReferenceViewSet, basename="reference")
+router.register("front", FrontViewSet, basename="front")
 
 urlpatterns = router.urls

config/settings/base.py

@@ -81,6 +81,7 @@
     "core_settings",
     "xml_manager",
     "reference",
+    "front",
 ]
 
 INSTALLED_APPS = DJANGO_APPS + THIRD_PARTY_APPS + LOCAL_APPS + WAGTAIL
@@ -325,3 +326,12 @@
 REFERENCE_TOKEN = env("REFERENCE_TOKEN", default="")
 REFERENCE_BATCH_SIZE = env.int("REFERENCE_BATCH_SIZE", default=10)
 REFERENCE_NUM_CTX = env.int("REFERENCE_NUM_CTX", default=8192)
+REFERENCE_KEEP_ALIVE = env("REFERENCE_KEEP_ALIVE", default="-1")
+
+FRONT_ENABLED = env.bool("FRONT_ENABLED", default=True)
+FRONT_URL = env("FRONT_URL", default="")
+FRONT_MODEL = env("FRONT_MODEL", default="llama3.2:3b")
+FRONT_TIMEOUT = env.int("FRONT_TIMEOUT", default=300)
+FRONT_TOKEN = env("FRONT_TOKEN", default="")
+FRONT_NUM_CTX = env.int("FRONT_NUM_CTX", default=8192)
+FRONT_KEEP_ALIVE = env("FRONT_KEEP_ALIVE", default="-1")

front/api/v1/serializers.py

@@ -0,0 +1,47 @@
+from rest_framework import serializers
+
+OUTPUT_TYPE_CHOICES = ["json", "xml"]
+
+
+class FrontMarkRequestSerializer(serializers.Serializer):
+    front = serializers.CharField(
+        allow_blank=False,
+        trim_whitespace=True,
+        help_text="Texto do front do artigo a marcar.",
+    )
+    type = serializers.ChoiceField(
+        choices=OUTPUT_TYPE_CHOICES,
+        default="json",
+        required=False,
+        help_text="Formato de saída: json ou xml.",
+    )
+    language = serializers.CharField(
+        required=False,
+        allow_blank=True,
+        help_text="Idioma principal do artigo (ex.: pt, en, es).",
+    )
+
+
+class FrontDocxRequestSerializer(serializers.Serializer):
+    file = serializers.FileField(
+        help_text="Arquivo .docx com o front do artigo.",
+    )
+    type = serializers.ChoiceField(
+        choices=OUTPUT_TYPE_CHOICES,
+        default="json",
+        required=False,
+        help_text="Formato de saída: json ou xml.",
+    )
+    language = serializers.CharField(
+        required=False,
+        allow_blank=True,
+        help_text="Idioma principal do artigo (ex.: pt, en, es).",
+    )
+
+    def validate_file(self, value):
+        name = (getattr(value, "name", "") or "").lower()
+        if not name.endswith(".docx"):
+            raise serializers.ValidationError("Only .docx files are accepted.")
+        if getattr(value, "size", None) == 0:
+            raise serializers.ValidationError("Empty file.")
+        return value

front/api/v1/views.py

@@ -0,0 +1,96 @@
+from collections.abc import Mapping
+
+from django.http import JsonResponse
+from rest_framework.decorators import action
+from rest_framework.parsers import FormParser, MultiPartParser
+from rest_framework.permissions import IsAuthenticated
+from rest_framework.response import Response
+from rest_framework.viewsets import GenericViewSet
+
+from front.api.v1.serializers import (
+    FrontDocxRequestSerializer,
+    FrontMarkRequestSerializer,
+)
+from front.data_utils import resolve_front_result
+from front.exceptions import (
+    FrontDocxError,
+    FrontLlamaDisabledError,
+    FrontLlamaMisconfiguredError,
+    FrontLlamaUnavailableError,
+)
+from front.utils import front_from_docx_upload
+
+
+class FrontViewSet(GenericViewSet):
+    serializer_class = FrontMarkRequestSerializer
+    permission_classes = [IsAuthenticated]
+    http_method_names = [
+        "get",
+        "post",
+        "head",
+        "options",
+    ]
+
+    def get_serializer_class(self):
+        if getattr(self, "action", None) == "docx":
+            return FrontDocxRequestSerializer
+        return FrontMarkRequestSerializer
+
+    def create(self, request, *args, **kwargs):
+        data = request.data
+        if not isinstance(data, Mapping):
+            return JsonResponse({"error": "Error processing"}, status=400)
+
+        serializer = self.get_serializer(data=data)
+        if not serializer.is_valid():
+            return JsonResponse(serializer.errors, status=400)
+
+        return self.mark_and_respond(
+            serializer.validated_data["front"],
+            serializer.validated_data.get("type", "json"),
+            serializer.validated_data.get("language") or None,
+        )
+
+    @action(
+        detail=False,
+        methods=["get", "post"],
+        url_path="docx",
+        parser_classes=[MultiPartParser, FormParser],
+    )
+    def docx(self, request):
+        if request.method == "GET":
+            return Response({})
+
+        serializer = self.get_serializer(data=request.data)
+        if not serializer.is_valid():
+            return JsonResponse(serializer.errors, status=400)
+
+        uploaded = serializer.validated_data["file"]
+        output_type = serializer.validated_data.get("type", "json")
+        language = serializer.validated_data.get("language") or None
+        try:
+            front_text = front_from_docx_upload(uploaded)
+        except FrontDocxError as exc:
+            return JsonResponse({"error": str(exc)}, status=400)
+        return self.mark_and_respond(front_text, output_type, language)
+
+    def mark_and_respond(self, front_text, output_type, language):
+        if not str(front_text or "").strip():
+            return JsonResponse({"error": "No front provided"}, status=400)
+        try:
+            result = resolve_front_result(
+                front_text,
+                user=self.request.user,
+                output_type=output_type,
+                language=language,
+            )
+        except (
+            FrontLlamaDisabledError,
+            FrontLlamaMisconfiguredError,
+            FrontLlamaUnavailableError,
+        ) as exc:
+            return JsonResponse(
+                {"error": f"Llama model is not available: {exc}"},
+                status=503,
+            )
+        return JsonResponse(result)

front/apps.py

@@ -0,0 +1,6 @@
+from django.apps import AppConfig
+
+
+class FrontConfig(AppConfig):
+    default_auto_field = "django.db.models.BigAutoField"
+    name = "front"