compress images on serverside

.gitignore

@@ -3,7 +3,7 @@ __pycache__/
 *$py.class
 
 **/data
-**/venv
+**/.venv
 *.env
 
 .github/workflows/main.yml

Dockerfile-api

@@ -1,8 +1,11 @@
-FROM python:3.13
+FROM python:3.13-slim
 
 
 # install requirements
 COPY requirements.txt requirements.txt
+# this is awful but the build time inprovements
+RUN sed -i 's/fastapi\[standard\]/fastapi/g' requirements.txt
+
 RUN pip install --no-cache-dir --upgrade -r requirements.txt
 
 

api.py

@@ -122,7 +122,7 @@ async def executives_image(filename):
         return 404
 
 
-    response = FileResponse(path, headers={"Accept-Encoding": "gzip"})
+    response = FileResponse(path)
     return response
 
 @app.get(
@@ -149,7 +149,7 @@ async def event_image(filename):
     if not os.path.isfile(path):
         return 404
 
-    response = FileResponse(path, headers={"Accept-Encoding": "gzip"})
+    response = FileResponse(path)
     return response
 
 if __name__ == "__main__":

requirements.txt

@@ -3,4 +3,5 @@ notion-client
 fastapi[standard]
 pydantic
 requests
-schedule
+schedule
+pillow

sdk/fetch_events.py

@@ -7,7 +7,7 @@
 from notion_client import Client
 from pydantic import BaseModel
 import requests
-from sdk.helpers import create_human_readable_date, image_filename_to_url, propTextExtractor, multiplePropTextExtractor, create_folder_if_not_existing
+from sdk.helpers import attempt_compress_image, create_human_readable_date, image_filename_to_url, propTextExtractor, multiplePropTextExtractor, create_folder_if_not_existing
 import logging
 
 from sdk.models import LCSCEvent
@@ -110,9 +110,12 @@ def updateDataFromNotion(writeLocation="data/"):
             assert file_name.split(".")[-1].lower() in ["webp", "jpg", "png", "jpeg", ".gif"]
 
             if page_updated:
-                with open(f"data/event_images/{page_id}.{file_extension}", "wb") as fi:          
+                file = f"data/event_images/{page_id}.{file_extension}"
+                with open(file, "wb") as fi:          
                     r = requests.get(file_url)
                     fi.write(r.content)
+
+                attempt_compress_image(file)
 
             event_images[page_id] = image_filename_to_url("events/images", f"{page_id}.{file_extension}")
         else:

sdk/fetch_execs.py

@@ -11,7 +11,7 @@
 from notion_client import Client
 from pydantic import BaseModel
 
-from sdk.helpers import create_human_readable_date, image_filename_to_url, propTextExtractor, multiplePropTextExtractor
+from sdk.helpers import attempt_compress_image, create_human_readable_date, image_filename_to_url, propTextExtractor, multiplePropTextExtractor
 
 logger = logging.getLogger("lcsc.execs")
 
@@ -145,11 +145,14 @@ def updateDataFromNotion(writeLocation="data/"):
 
             # don't actually request the image if we know it hasn't changed
             if not stale_data:
-                with open(f"data/exec_images/{student_id}.{file_extension}", "wb") as fi:
+                file = f"data/exec_images/{student_id}.{file_extension}"
+                with open(file, "wb") as fi:
 
                     r = requests.get(file_url)
                     fi.write(r.content)            
 
+                attempt_compress_image(file)
+
             executive_images[student_id] = image_filename_to_url("executives/images", f"{student_id}.{file_extension}")
         else:
             executive_images[student_id] = None

sdk/helpers.py

@@ -2,6 +2,8 @@
 from os import makedirs
 import os
 from os.path import exists
+from PIL import Image
+Image.MAX_IMAGE_PIXELS = None # pil is paranoid and thinks large images are a zip bomb
 
 # Iterate over Notion pages and extract data
 def propTextExtractor(property:dict) -> str | None:
@@ -91,4 +93,24 @@ def image_filename_to_url(api_route:str, filename:str):
         path += "/"
 
     return f"{path}{api_route}/{filename}"
-
+
+# take care to only call this once per image (ie right after you save it)
+def attempt_compress_image(filepath:str):
+    image = Image.open(filepath)
+
+    i_width, i_height = image.size
+
+    if i_width > 2000 or i_height > 2000:
+        # Calculate the new size while maintaining the aspect ratio
+        if i_width > i_height:
+            new_width = 2000
+            new_height = int((i_height * new_width) / i_width)
+        else:
+            new_height = 2000
+            new_width = int((i_width * new_height) / i_height)
+
+        # Resize the image
+        image = image.resize((new_width, new_height), Image.LANCZOS)
+
+    # Save the image with optimized settings
+    image.save(filepath, optimize=True, quality=95)