Return XML using the API container

docker-compose-test.yml

@@ -50,6 +50,8 @@ services:
       - "5060:5060"
     networks:
       - network-pdf-layout
+    volumes:
+      - data:/app/xmls
 
   mongo-pdf-layout:
     container_name: "mongo-pdf-layout"

docker-compose.yml

@@ -51,6 +51,8 @@ services:
             - driver: nvidia
               count: 1
               capabilities: [ gpu ]
+    volumes:
+      - data:/app/xmls
 
   mongo-pdf-layout:
     container_name: "mongo-pdf-layout"

src/QueueProcessor.py

@@ -56,14 +56,13 @@ def process(self, id, message, rc, ts):
             xml_file_name = get_xml_name(task)
             extraction_data = extract_segments(task, xml_file_name)
             service_url = f"{SERVICE_HOST}:{SERVICE_PORT}"
-            get_xml_url = f"{SERVICE_HOST}:{DOCUMENT_LAYOUT_ANALYSIS_PORT}"
             extraction_message = ResultMessage(
                 tenant=extraction_data.tenant,
                 task=task.task,
                 params=task.params,
                 success=True,
                 data_url=f"{service_url}/get_paragraphs/{task.tenant}/{task.params.filename}",
-                file_url=f"{get_xml_url}/get_xml/{xml_file_name}",
+                file_url=f"{service_url}/get_xml/{xml_file_name}",
             )
 
             extraction_data_json = extraction_data.model_dump_json()

src/app.py

@@ -8,11 +8,13 @@
 from sentry_sdk.integrations.asgi import SentryAsgiMiddleware
 import sentry_sdk
 from starlette.concurrency import run_in_threadpool
+from starlette.responses import PlainTextResponse
 
 from catch_exceptions import catch_exceptions
 from configuration import MONGO_HOST, MONGO_PORT, service_logger
 from PdfFile import PdfFile
 from get_paragraphs import get_paragraphs
+from get_xml import get_xml
 from run import extract_segments_from_file
 
 
@@ -69,3 +71,9 @@ async def async_extraction(tenant, file: UploadFile = File(...)):
 @catch_exceptions
 async def get_paragraphs_endpoint(tenant: str, pdf_file_name: str):
     return await run_in_threadpool(get_paragraphs, app.mongodb_client, tenant, pdf_file_name)
+
+
+@app.get("/get_xml/{xml_file_name}", response_class=PlainTextResponse)
+@catch_exceptions
+async def get_xml_by_name(xml_file_name: str):
+    return await run_in_threadpool(get_xml, xml_file_name)

src/get_xml.py

@@ -0,0 +1,14 @@
+import os
+from os.path import join
+from pathlib import Path
+
+from configuration import DATA_PATH
+
+
+def get_xml(xml_file_name: str) -> str:
+    xml_file_path = Path(join(DATA_PATH, xml_file_name))
+
+    with open(xml_file_path, mode="r") as file:
+        content = file.read()
+        os.remove(xml_file_path)
+        return content