Add image upload and vision analysis to Ask Simba chat

Users can now attach images in the web chat for Simba to analyze using
Ollama's gemma3 vision model. Images are stored in Garage (S3-compatible)
and displayed in chat history.

Also fixes aerich migration config by extracting TORTOISE_CONFIG into a
standalone config/db.py module, removing the stale aerich_config.py, and
adding missing MODELS_STATE to migration 3.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
2026-04-04 08:03:19 -04:00
parent ac9c821ec7
commit 0415610d64
17 changed files with 501 additions and 58 deletions

View File

@@ -76,6 +76,39 @@ def describe_simba_image(input):
return result
async def analyze_user_image(file_bytes: bytes) -> str:
"""Analyze an image uploaded by a user and return a text description.
Uses Ollama vision model to describe the image contents.
Works with JPEG, PNG, WebP bytes (HEIC should be converted before calling).
"""
import tempfile
# Write to temp file since ollama client expects a file path
with tempfile.NamedTemporaryFile(suffix=".jpg", delete=False) as f:
f.write(file_bytes)
temp_path = f.name
try:
response = client.chat(
model="gemma3:4b",
messages=[
{
"role": "system",
"content": "You are a helpful image analyst. Describe what you see in the image in detail. Be thorough but concise.",
},
{
"role": "user",
"content": "Please describe this image in detail.",
"images": [temp_path],
},
],
)
return response["message"]["content"]
finally:
os.remove(temp_path)
if __name__ == "__main__":
args = parser.parse_args()
if args.filepath:

62
utils/image_upload.py Normal file
View File

@@ -0,0 +1,62 @@
import io
import logging
from PIL import Image
from pillow_heif import register_heif_opener
register_heif_opener()
logging.basicConfig(level=logging.INFO)
ALLOWED_TYPES = {"image/jpeg", "image/png", "image/webp", "image/heic", "image/heif"}
MAX_DIMENSION = 1920
class ImageValidationError(Exception):
pass
def process_image(file_bytes: bytes, content_type: str) -> tuple[bytes, str]:
"""Validate, resize, and strip EXIF from an uploaded image.
Returns processed bytes and the output content type (always image/jpeg or image/png or image/webp).
"""
if content_type not in ALLOWED_TYPES:
raise ImageValidationError(
f"Unsupported image type: {content_type}. "
f"Allowed: JPEG, PNG, WebP, HEIC"
)
img = Image.open(io.BytesIO(file_bytes))
# Resize if too large
width, height = img.size
if max(width, height) > MAX_DIMENSION:
ratio = MAX_DIMENSION / max(width, height)
new_size = (int(width * ratio), int(height * ratio))
img = img.resize(new_size, Image.LANCZOS)
logging.info(
f"Resized image from {width}x{height} to {new_size[0]}x{new_size[1]}"
)
# Strip EXIF by copying pixel data to a new image
clean_img = Image.new(img.mode, img.size)
clean_img.putdata(list(img.getdata()))
# Convert HEIC/HEIF to JPEG; otherwise keep original format
if content_type in {"image/heic", "image/heif"}:
output_format = "JPEG"
output_content_type = "image/jpeg"
elif content_type == "image/png":
output_format = "PNG"
output_content_type = "image/png"
elif content_type == "image/webp":
output_format = "WEBP"
output_content_type = "image/webp"
else:
output_format = "JPEG"
output_content_type = "image/jpeg"
buf = io.BytesIO()
clean_img.save(buf, format=output_format, quality=85)
return buf.getvalue(), output_content_type

53
utils/s3_client.py Normal file
View File

@@ -0,0 +1,53 @@
import os
import logging
import aioboto3
from dotenv import load_dotenv
load_dotenv()
logging.basicConfig(level=logging.INFO)
S3_ENDPOINT_URL = os.getenv("S3_ENDPOINT_URL")
S3_ACCESS_KEY_ID = os.getenv("S3_ACCESS_KEY_ID")
S3_SECRET_ACCESS_KEY = os.getenv("S3_SECRET_ACCESS_KEY")
S3_BUCKET_NAME = os.getenv("S3_BUCKET_NAME", "asksimba-images")
S3_REGION = os.getenv("S3_REGION", "garage")
session = aioboto3.Session()
def _get_client():
return session.client(
"s3",
endpoint_url=S3_ENDPOINT_URL,
aws_access_key_id=S3_ACCESS_KEY_ID,
aws_secret_access_key=S3_SECRET_ACCESS_KEY,
region_name=S3_REGION,
)
async def upload_image(file_bytes: bytes, key: str, content_type: str) -> str:
async with _get_client() as client:
await client.put_object(
Bucket=S3_BUCKET_NAME,
Key=key,
Body=file_bytes,
ContentType=content_type,
)
logging.info(f"Uploaded image to S3: {key}")
return key
async def get_image(key: str) -> tuple[bytes, str]:
async with _get_client() as client:
response = await client.get_object(Bucket=S3_BUCKET_NAME, Key=key)
body = await response["Body"].read()
content_type = response.get("ContentType", "image/jpeg")
return body, content_type
async def delete_image(key: str) -> None:
async with _get_client() as client:
await client.delete_object(Bucket=S3_BUCKET_NAME, Key=key)
logging.info(f"Deleted image from S3: {key}")