Add application file

.gitignore

@@ -0,0 +1,2 @@
+__pycache__
+.env

Dockerfile

@@ -0,0 +1,27 @@
+# Usa la imagen slim-buster para evitar problemas con dependencias C
+FROM python:3.12-slim
+
+# Configura el directorio de trabajo
+WORKDIR /app
+
+# Copia solo requirements.txt para cachear la instalación de dependencias
+COPY requirements.txt .
+
+# Instala dependencias del sistema para compatibilidad con Milvus y Transformers
+RUN apt-get update && apt-get install -y \
+    gcc \
+    libpq-dev \
+    build-essential \
+    && rm -rf /var/lib/apt/lists/*
+
+# Instala dependencias de Python
+RUN pip install --no-cache-dir --upgrade pip \
+    && pip install --no-cache-dir -r requirements.txt
+
+# Copia el resto del código de la app
+COPY . .
+
+EXPOSE 7860
+
+# Define el comando de arranque de la API
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

app.py

@@ -0,0 +1,38 @@
+from fastapi.security import HTTPBearer
+from fastapi import FastAPI, HTTPException, Depends
+from functions import retrieve_similar_sentence
+
+app = FastAPI()
+security = HTTPBearer()
+
+@app.post("/translate")
+def translate_sentence(data: dict, token: str = Depends(security)):
+	try: 
+		api_key = token.credentials
+		sentence = data["sentence"]
+		source_language = data["source_language"]
+		if not sentence or not source_language:
+			raise HTTPException(status_code=400, detail="Missing data in the request body")
+		
+		results = retrieve_similar_sentence(sentence, source_language, api_key)
+		return { 
+			"top_match": {
+				"source_sentence": results[0]["source_sentence"],
+				"target_sentence": results[0]["target_sentence"]
+			},
+			"2nd_match": {
+				"source_sentence": results[1]["source_sentence"],
+				"target_sentence": results[1]["target_sentence"]
+			},
+			"3rd_match": {
+				"source_sentence": results[2]["source_sentence"],
+				"target_sentence": results[2]["target_sentence"]
+			},
+			"4th_match": {
+				"source_sentence": results[3]["source_sentence"],
+				"target_sentence": results[3]["target_sentence"]
+			}
+		}
+	
+	except Exception as e:
+		raise HTTPException(status_code=500, detail=str(e))

functions.py

@@ -0,0 +1,38 @@
+import os
+from dotenv import load_dotenv
+from pinecone import Pinecone
+from sentence_transformers import SentenceTransformer
+
+load_dotenv()
+
+# === LOAD GNRL DATA ===
+EMBEDDING_MODEL = "sentence-transformers/all-MiniLM-L6-v2"
+model = SentenceTransformer(EMBEDDING_MODEL)
+
+
+# === RETRIEVAL FUNCTION FROM PINECONE ===
+def retrieve_similar_sentence(query_sentence, source_language, api_key):
+    pc = Pinecone(api_key = api_key)
+    index_name = "spa-quz-translation-index" if source_language == "es" else "quz-spa-translation-index"
+    index = pc.Index(index_name)
+
+    query_embedding = model.encode(query_sentence).tolist()
+
+    response = index.query(
+        vector=query_embedding,
+        top_k=4,
+        include_metadata=True
+    )
+
+    results = []
+    for match in response['matches']:
+      metadata = match['metadata']
+      score = match['score']
+
+      results.append({
+          "source_sentence": metadata["source_sentence"],
+          "target_sentence": metadata["target_sentence"],
+          "score": score
+      })
+
+    return results

requirements.txt

@@ -0,0 +1,6 @@
+fastapi==0.95.1
+pymilvus==2.4.8
+sentence_transformers==2.3.1
+uvicorn==0.22.0
+pinecone==5.3.1
+pinecone-client==5.0.1