Make TOC endpoint async

Makefile

@@ -1,22 +1,22 @@
 install:
-	. venv/bin/activate; pip install -Ur requirements.txt
+	. .venv/bin/activate; pip install -Ur requirements.txt
 
 activate:
-	. venv/bin/activate
+	. .venv/bin/activate
 
 install_venv:
-	python3 -m venv venv
-	. venv/bin/activate; python -m pip install --upgrade pip
-	. venv/bin/activate; python -m pip install -r dev-requirements.txt
+	python3 -m venv .venv
+	. .venv/bin/activate; python -m pip install --upgrade pip
+	. .venv/bin/activate; python -m pip install -r dev-requirements.txt
 
 formatter:
-	. venv/bin/activate; command black --line-length 125 .
+	. .venv/bin/activate; command black --line-length 125 .
 
 check_format:
-	. venv/bin/activate; command black --line-length 125 . --check
+	. .venv/bin/activate; command black --line-length 125 . --check
 
 test:
-	. venv/bin/activate; command cd src; command pytest
+	. .venv/bin/activate; command cd src; command pytest
 
 remove_docker_containers:
 	docker compose ps -q | xargs docker rm
@@ -43,4 +43,4 @@ stop:
 	docker compose stop
 
 delete_queues:
-	. venv/bin/activate; python scripts/delete_queues.py
+	. .venv/bin/activate; python scripts/delete_queues.py

dev-requirements.txt

@@ -1,4 +1,4 @@
 -r requirements.txt
-pytest==7.4.2
-black==23.9.1
+pytest==8.2.0
+black==24.4.2
 mongomock==4.1.2

requirements.txt

@@ -1,14 +1,14 @@
-fastapi==0.110.1
+fastapi==0.110.2
 graypy==2.1.0
 python-multipart==0.0.9
 uvicorn==0.29.0
-gunicorn==21.2.0
+gunicorn==22.0.0
 nltk==3.8.1
 PyYAML==6.0.1
 requests==2.31.0
-pymongo==4.6.3
+pymongo==4.7.0
 PyRSMQ==0.5.0
-redis==5.0.3
+redis==5.0.4
 httpx==0.27.0
-sentry-sdk==1.44.0
+sentry-sdk==2.0.1
 git+https://github.com/huridocs/pdf-tokens-type-labeler@1c12c368887372164ab4981c3277a49e9dc43b9a

setup.py

@@ -12,7 +12,7 @@
     name=PROJECT_NAME,
     packages=["paragraph_extraction_trainer"],
     package_dir={"": "src"},
-    version="0.32",
+    version="0.33",
     url="https://github.com/huridocs/pdf_paragraphs_extraction",
     author="HURIDOCS",
     description="Service for extracting paragraphs from PDFs.",

src/app.py

@@ -135,7 +135,7 @@ async def get_xml(tenant: str, pdf_file_name: str):
 
 
 @app.post("/get_toc")
-def get_toc(file: UploadFile = File(...)):
+async def get_toc(file: UploadFile = File(...)):
     filename = "No file name!"
     try:
         filename = file.filename

src/toc/TitleFeatures.py

@@ -273,7 +273,7 @@ def get_features_toc(self) -> np.array:
             1 if self.italics else 0,
             self.first_characters_type,
             self.first_characters_special_markers_count,
-            self.bullet_points_type
+            self.bullet_points_type,
             # self.text_centered,
             # self.is_left,
             # self.indentation