Merge pull request #150 from presenton/fix/embedding-issue-in-arm

fix/embedding issue in arm

nginx.conf

@@ -30,28 +30,41 @@ http {
       proxy_connect_timeout 30m;
     }
 
-    location /static {
-      proxy_pass http://localhost:8000;
-      proxy_read_timeout 30m;
-      proxy_connect_timeout 30m;
-    }
-
-    location /app_data {
-      proxy_pass http://localhost:8000;
-      proxy_read_timeout 30m;
-      proxy_connect_timeout 30m;
-    }
-
     location /docs {
       proxy_pass http://localhost:8000/docs;
       proxy_read_timeout 30m;
       proxy_connect_timeout 30m;
     }
-    
+
     location /openapi.json {
       proxy_pass http://localhost:8000/openapi.json;
       proxy_read_timeout 30m;
       proxy_connect_timeout 30m;
     }
+
+    # Static
+    location /static {
+      alias /app/servers/fastapi/static/;
+      expires 1y;
+      add_header Cache-Control "public, immutable";
+    }
+
+    location /app_data/images/ {
+      alias /app_data/images/;
+      expires 1y;
+      add_header Cache-Control "public, immutable";
+    }
+
+    location /app_data/exports/ {
+      alias /app_data/exports/;
+      expires 1y;
+      add_header Cache-Control "public, immutable";
+    }
+
+    location /app_data/uploads/ {
+      alias /app_data/uploads/;
+      expires 1y;
+      add_header Cache-Control "public, immutable";
+    }
   }
 }

servers/fastapi/api/v1/ppt/endpoints/presentation.py

@@ -22,6 +22,8 @@ from models.generate_presentation_api import (
     PresentationPathAndEditPath,
 )
 from services.get_layout_by_name import get_layout_by_name
+from services.icon_finder_service import IconFinderService
+from services.image_generation_service import ImageGenerationService
 from utils.llm_calls.generate_presentation_outlines import generate_ppt_outline
 from models.sql.slide import SlideModel
 from models.sse_response import SSECompleteResponse, SSEResponse
@@ -30,7 +32,7 @@ from services.database import get_sql_session
 from services.documents_loader import DocumentsLoader
 from models.sql.presentation import PresentationModel
 from services.pptx_presentation_creator import PptxPresentationCreator
-from utils.asset_directory_utils import get_exports_directory
+from utils.asset_directory_utils import get_exports_directory, get_images_directory
 from utils.llm_calls.generate_document_summary import generate_document_summary
 from utils.llm_calls.generate_presentation_structure import (
     generate_presentation_structure,
@@ -195,6 +197,9 @@ async def stream_presentation(presentation_id: str):
                 detail="Outlines can not be empty",
             )
 
+    image_generation_service = ImageGenerationService(get_images_directory())
+    icon_finder_service = IconFinderService()
+
     async def inner():
         structure = presentation.get_structure()
         layout = presentation.get_layout()
@@ -223,7 +228,11 @@ async def stream_presentation(presentation_id: str):
             slides.append(slide)
 
             # This will mutate slide
-            async_assets_generation_tasks.append(process_slide_and_fetch_assets(slide))
+            async_assets_generation_tasks.append(
+                process_slide_and_fetch_assets(
+                    image_generation_service, icon_finder_service, slide
+                )
+            )
 
             # Give control to the event loop
             await asyncio.sleep(0)
@@ -423,9 +432,13 @@ async def generate_presentation_api(
 
     # Process slides to fetch assets (images, icons, etc.)
     print("Processing slides to fetch assets")
+    image_generation_service = ImageGenerationService(get_images_directory())
+    icon_finder_service = IconFinderService()
     for slide in slides:
         try:
-            await process_slide_and_fetch_assets(slide)
+            await process_slide_and_fetch_assets(
+                image_generation_service, icon_finder_service, slide
+            )
             print(f"Processed slide {slide.index} successfully")
         except Exception as e:
             print(f"Error processing slide {slide.index}: {e}")

servers/fastapi/api/v1/ppt/endpoints/slide.py

@@ -4,6 +4,9 @@ from fastapi import APIRouter, Body, HTTPException
 from models.sql.presentation import PresentationModel
 from models.sql.slide import SlideModel
 from services.database import get_sql_session
+from services.icon_finder_service import IconFinderService
+from services.image_generation_service import ImageGenerationService
+from utils.asset_directory_utils import get_images_directory
 from utils.llm_calls.edit_slide import get_edited_slide_content
 from utils.llm_calls.edit_slide_html import get_edited_slide_html
 from utils.llm_calls.select_slide_type_on_edit import get_slide_layout_from_prompt
@@ -34,9 +37,15 @@ async def edit_slide(id: Annotated[str, Body()], prompt: Annotated[str, Body()])
         prompt, slide_layout, slide, presentation.language
     )
 
+    image_generation_service = ImageGenerationService(get_images_directory())
+    icon_finder_service = IconFinderService()
+
     # This will mutate edited_slide_content
     new_assets = await process_old_and_new_slides_and_fetch_assets(
-        slide.content, edited_slide_content
+        image_generation_service,
+        icon_finder_service,
+        slide.content,
+        edited_slide_content,
     )
 
     # Always assign a new unique id to the slide

servers/fastapi/qdrant/.lock

@@ -0,0 +1 @@
+tmp lock file

servers/fastapi/qdrant/meta.json

@@ -0,0 +1 @@
+{"collections": {"icons": {"vectors": {"fast-bge-small-en-v1.5": {"size": 384, "distance": "Cosine", "hnsw_config": null, "quantization_config": null, "on_disk": null, "datatype": null, "multivector_config": null}}, "shard_number": null, "sharding_method": null, "replication_factor": null, "write_consistency_factor": null, "on_disk_payload": null, "hnsw_config": null, "wal_config": null, "optimizers_config": null, "init_from": null, "quantization_config": null, "sparse_vectors": null, "strict_mode_config": null}}, "aliases": {}}

servers/fastapi/requirements.txt

@@ -5,70 +5,120 @@ annotated-types==0.7.0
 anyio==4.9.0
 async-timeout==5.0.1
 attrs==25.3.0
+backoff==2.2.1
+bcrypt==4.3.0
+build==1.2.2.post1
 cachetools==5.5.2
 certifi==2025.7.14
 cffi==1.17.1
 charset-normalizer==3.4.2
 click==8.2.1
+coloredlogs==15.0.1
 cryptography==45.0.5
 distro==1.9.0
 dnspython==2.7.0
+durationpy==0.10
 email_validator==2.2.0
 fastapi==0.116.1
 fastapi-cli==0.0.8
 fastapi-cloud-cli==0.1.4
-fastembed_vectorstore==0.1.6
+fastembed==0.7.1
+filelock==3.18.0
+flatbuffers==25.2.10
 frozenlist==1.7.0
+fsspec==2025.7.0
 google-auth==2.40.3
 google-genai==1.25.0
+googleapis-common-protos==1.70.0
 greenlet==3.2.3
+grpcio==1.74.0
 h11==0.16.0
+h2==4.2.0
+hf-xet==1.1.5
+hpack==4.1.0
 httpcore==1.0.9
 httptools==0.6.4
 httpx==0.28.1
+huggingface-hub==0.34.1
+humanfriendly==10.0
+hyperframe==6.1.0
 idna==3.10
+importlib_metadata==8.7.0
+importlib_resources==6.5.2
 iniconfig==2.1.0
 Jinja2==3.1.6
 jiter==0.10.0
+jsonschema==4.25.0
+jsonschema-specifications==2025.4.1
+kubernetes==33.1.0
+loguru==0.7.3
 lxml==6.0.0
 markdown-it-py==3.0.0
 MarkupSafe==3.0.2
 mdurl==0.1.2
+mmh3==5.1.0
+mpmath==1.3.0
 multidict==6.6.3
+numpy==2.3.2
+oauthlib==3.3.1
+onnxruntime==1.22.1
 openai==1.95.1
+opentelemetry-api==1.35.0
+opentelemetry-exporter-otlp-proto-common==1.35.0
+opentelemetry-exporter-otlp-proto-grpc==1.35.0
+opentelemetry-proto==1.35.0
+opentelemetry-sdk==1.35.0
+opentelemetry-semantic-conventions==0.56b0
+orjson==3.11.1
+overrides==7.7.0
 packaging==25.0
 pathvalidate==3.3.1
 pdfminer.six==20250506
 pdfplumber==0.11.7
 pillow==11.3.0
 pluggy==1.6.0
+portalocker==3.2.0
+posthog==5.4.0
 propcache==0.3.2
+protobuf==6.31.1
+py_rust_stemmers==0.1.5
 pyasn1==0.6.1
 pyasn1_modules==0.4.2
+pybase64==1.4.2
 pycparser==2.22
 pydantic==2.11.7
 pydantic_core==2.33.2
 Pygments==2.19.2
 pypdfium2==4.30.1
+PyPika==0.48.9
+pyproject_hooks==1.2.0
 pytest==8.4.1
+python-dateutil==2.9.0.post0
 python-docx==1.2.0
 python-dotenv==1.1.1
 python-multipart==0.0.20
 python-pptx==1.0.2
 PyYAML==6.0.2
+qdrant-client==1.15.0
 redis==6.2.0
+referencing==0.36.2
 requests==2.32.4
+requests-oauthlib==2.0.0
 rich==14.0.0
 rich-toolkit==0.14.8
 rignore==0.6.2
+rpds-py==0.26.0
 rsa==4.9.1
 sentry-sdk==2.32.0
 shellingham==1.5.4
+six==1.17.0
 sniffio==1.3.1
 SQLAlchemy==2.0.41
 sqlmodel==0.0.24
 starlette==0.47.1
+sympy==1.14.0
 tenacity==8.5.0
+tokenizers==0.21.2
 tqdm==4.67.1
 typer==0.16.0
 typing-inspection==0.4.1
@@ -77,6 +127,8 @@ urllib3==2.5.0
 uvicorn==0.35.0
 uvloop==0.21.0
 watchfiles==1.1.0
+websocket-client==1.8.0
 websockets==15.0.1
 xlsxwriter==3.2.5
 yarl==1.20.1
+zipp==3.23.0

servers/fastapi/server.py

@@ -1,7 +1,8 @@
-import os
 import uvicorn
 import argparse
 
+from services.icon_finder_service import IconFinderService
+
 
 if __name__ == "__main__":
     parser = argparse.ArgumentParser(description="Run the FastAPI server")
@@ -13,6 +14,9 @@ if __name__ == "__main__":
     )
     args = parser.parse_args()
 
+    # Initialize the Icons Collection
+    IconFinderService()
+
     uvicorn.run(
         "api.main:app",
         host="0.0.0.0",

servers/fastapi/services/icon_finder_service.py

@@ -1,35 +1,48 @@
 import asyncio
 import json
-import os
-from fastembed_vectorstore import FastembedVectorstore, FastembedEmbeddingModel
+from qdrant_client import QdrantClient
 
 
 class IconFinderService:
     def __init__(self):
-        self.vector_store = self.get_icons_vectorstore()
+        self.collection_name = "icons"
+        self.client = QdrantClient(path="qdrant")
+        print("Initializing icons collection...")
+        self.client.set_model("BAAI/bge-small-en-v1.5")
+        self._initialize_icons_collection()
+        print("Icons collection initialized")
 
-    def get_icons_vectorstore(self):
-        vector_store_path = "assets/icons_vectorstore.json"
-        embedding_model = FastembedEmbeddingModel.BGESmallENV15
+    def _initialize_icons_collection(self):
+        try:
+            self.client.get_collection(self.collection_name)
+        except Exception:
+            self._populate_icons_collection()
 
-        if os.path.exists(vector_store_path):
-            return FastembedVectorstore.load(embedding_model, vector_store_path)
-
-        vector_store = FastembedVectorstore(embedding_model)
+    def _populate_icons_collection(self):
         with open("assets/icons.json", "r") as f:
             icons = json.load(f)
+
         documents = []
+        metadata = []
+
         for each in icons["icons"]:
             if each["name"].split("-")[-1] == "bold":
-                documents.append(f"{each['name']}||{each['tags']}")
+                doc_text = f"{each['name']} {each['tags']}"
+                documents.append(doc_text)
+                metadata.append({"name": each["name"]})
 
-        vector_store.embed_documents(documents)
-        vector_store.save(vector_store_path)
-
-        return vector_store
+        if documents:
+            self.client.add(
+                collection_name=self.collection_name,
+                documents=documents,
+                metadata=metadata,
+            )
 
     async def search_icons(self, query: str, k: int = 1):
-        result = await asyncio.to_thread(self.vector_store.search, query, k)
-        return [
-            f"/static/icons/bold/{result[0].split('||')[0]}.png" for result in result
-        ]
+        result = await asyncio.to_thread(
+            self.client.query,
+            collection_name=self.collection_name,
+            query_text=query,
+            limit=k,
+        )
+        return [f"/static/icons/bold/{each.metadata['name']}.png" for each in result]

servers/fastapi/utils/process_slides.py

@@ -10,12 +10,10 @@ from utils.dict_utils import get_dict_at_path, get_dict_paths_with_key, set_dict
 
 
 async def process_slide_and_fetch_assets(
+    image_generation_service: ImageGenerationService,
+    icon_finder_service: IconFinderService,
     slide: SlideModel,
 ) -> List[ImageAsset]:
-    image_directory = get_images_directory()
-
-    image_generation_service = ImageGenerationService(image_directory)
-    icon_finder_service = IconFinderService()
 
     async_tasks = []
 
@@ -61,6 +59,8 @@ async def process_slide_and_fetch_assets(
 
 
 async def process_old_and_new_slides_and_fetch_assets(
+    image_generation_service: ImageGenerationService,
+    icon_finder_service: IconFinderService,
     old_slide_content: dict,
     new_slide_content: dict,
 ) -> List[ImageAsset]:
@@ -98,10 +98,6 @@ async def process_old_and_new_slides_and_fetch_assets(
         get_dict_at_path(new_slide_content, path) for path in new_icon_dict_paths
     ]
 
-    # Creates services
-    image_generation_service = ImageGenerationService(get_images_directory())
-    icon_finder_service = IconFinderService()
-
     # Creates async tasks for fetching new images
     async_image_fetch_tasks = []
     new_images_fetch_status = []