refactor extractors to use S3 text extraction function and update dependencies

Author: monicasmith463

backend/app/core/extractors.py

@@ -5,30 +5,12 @@
 import textract
 import requests
 from app.models import Document
+from app.s3 import extract_text_from_s3_file
 
-
-def extract_text_from_file(s3_url: str) -> str:
-    try:
-        response = requests.get(s3_url)
-        response.raise_for_status()
-
-        with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp_file:
-            tmp_file.write(response.content)
-            tmp_path = tmp_file.name
-
-        text = textract.process(tmp_path).decode("utf-8") or ""
-
-        os.remove(tmp_path)
-
-        return text
-
-    except Exception as e:
-        raise Exception(f"Failed to extract text: {e}")
-    
 def extract_text_and_save_to_db(s3_url: str, document_id: str) -> None:
     try:
         with Session(engine) as session:
-            text = extract_text_from_file(s3_url)
+            text = extract_text_from_s3_file(s3_url)
 
             document_query = select(Document).where(Document.id == document_id)
             document = session.exec(document_query).first()

backend/app/s3.py

@@ -1,9 +1,12 @@
+import os
+import tempfile
 import uuid
 
 import boto3
 from fastapi import UploadFile
 
 from app.core.config import settings
+import textract
 
 s3 = boto3.client(
     "s3",
@@ -29,3 +32,12 @@ def upload_file_to_s3(file: UploadFile, user_id: str) -> str:
 
 def generate_s3_url(key: str) -> str:
     return f"https://{settings.S3_BUCKET_NAME}.s3.amazonaws.com/{key}"
+
+def extract_text_from_s3_file(bucket: str, key: str) -> str:
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp_file:
+        s3.download_fileobj(bucket, key, tmp_file)
+        tmp_path = tmp_file.name
+
+    text = textract.process(tmp_path).decode("utf-8") or ""
+    os.remove(tmp_path)
+    return text

backend/pyproject.toml

@@ -21,10 +21,10 @@ dependencies = [
     "boto3",
     "pydantic-settings<3.0.0,>=2.2.1",
     "sentry-sdk[fastapi]<2.0.0,>=1.40.6",
-    "pyjwt<3.0.0,>=2.8.0",
+    "pyjwt<3.0.0,>=2.8.0",    
     "textract>=1.6.5",
     "boto3-stubs>=1.40.2",
-    "requests>=2.32.3",
+    "textract>=1.6.5",
 ]
 
 [tool.uv]
@@ -35,6 +35,8 @@ dev-dependencies = [
     "pre-commit<4.0.0,>=3.6.2",
     "types-passlib<2.0.0.0,>=1.7.7.20240106",
     "coverage<8.0.0,>=7.4.3",
+    "types-requests >=2.32.4.20250611",
+    "types-boto3>=1.39.16"
 ]
 
 [build-system]