feat(fastapi): uses better json loader that parses dirty json

Dockerfile

@@ -29,7 +29,7 @@ RUN curl -fsSL https://ollama.com/install.sh | sh
 # Install dependencies for FastAPI
 RUN pip install aiohttp aiomysql aiosqlite asyncpg fastapi[standard] \
     pathvalidate pdfplumber chromadb sqlmodel \
-    anthropic google-genai openai fastmcp
+    anthropic google-genai openai fastmcp dirtyjson
 RUN pip install docling --extra-index-url https://download.pytorch.org/whl/cpu
 
 # Install dependencies for Next.js

Dockerfile.dev

@@ -29,7 +29,7 @@ RUN curl -fsSL http://ollama.com/install.sh | sh
 # Install dependencies for FastAPI
 RUN pip install aiohttp aiomysql aiosqlite asyncpg fastapi[standard] \
   pathvalidate pdfplumber chromadb sqlmodel \
-  anthropic google-genai openai fastmcp
+  anthropic google-genai openai fastmcp dirtyjson
 RUN pip install docling --extra-index-url https://download.pytorch.org/whl/cpu
 
 # Install dependencies for Next.js

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

@@ -2,6 +2,7 @@ import asyncio
 import json
 import math
 import uuid
+import dirtyjson
 from fastapi import APIRouter, Depends, HTTPException
 from fastapi.responses import StreamingResponse
 from sqlalchemy.ext.asyncio import AsyncSession
@@ -82,7 +83,7 @@ async def stream_outlines(
             presentation_outlines_text += chunk
 
         try:
-            presentation_outlines_json = json.loads(presentation_outlines_text)
+            presentation_outlines_json = dict(dirtyjson.loads(presentation_outlines_text))
         except Exception as e:
             raise HTTPException(
                 status_code=400,

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

@@ -4,6 +4,7 @@ import math
 import os
 import random
 from typing import Annotated, List, Literal, Optional
+import dirtyjson
 from fastapi import APIRouter, Body, Depends, HTTPException
 from fastapi.responses import StreamingResponse
 from sqlalchemy import delete
@@ -486,7 +487,7 @@ async def generate_presentation_api(
             presentation_outlines_text += chunk
 
         try:
-            presentation_outlines_json = json.loads(presentation_outlines_text)
+            presentation_outlines_json = dict(dirtyjson.loads(presentation_outlines_text))
         except Exception as e:
             print(e)
             raise HTTPException(

servers/fastapi/pyproject.toml

@@ -11,6 +11,7 @@ dependencies = [
     "anthropic>=0.60.0",
     "asyncpg>=0.30.0",
     "chromadb>=1.0.15",
+    "dirtyjson>=1.0.8",
     "docling>=2.43.0",
     "fastapi[standard]>=0.116.1",
     "fastmcp>=2.11.0",

servers/fastapi/services/documents_loader.py

@@ -96,11 +96,15 @@ class DocumentsLoader:
         return self.docling_service.parse_to_markdown(file_path)
 
     @classmethod
-    def get_page_images_from_pdf(cls, file_path: str, temp_dir: str):
+    def get_page_images_from_pdf(cls, file_path: str, temp_dir: str) -> List[str]:
         with pdfplumber.open(file_path) as pdf:
+            images = []
             for page in pdf.pages:
                 img = page.to_image(resolution=150)
-                img.save(os.path.join(temp_dir, f"page_{page.page_number}.png"))
+                image_path = os.path.join(temp_dir, f"page_{page.page_number}.png")
+                img.save(image_path)
+                images.append(image_path)
+            return images
 
     @classmethod
     async def get_page_images_from_pdf_async(cls, file_path: str, temp_dir: str):

servers/fastapi/services/llm_client.py

@@ -1,4 +1,5 @@
 import asyncio
+import dirtyjson
 import json
 from typing import AsyncGenerator, List, Optional
 from fastapi import HTTPException
@@ -554,7 +555,7 @@ class LLMClient:
                 )
         if content:
             if depth == 0:
-                return json.loads(content)
+                return dict(dirtyjson.loads(content))
             return content
         return None
 
@@ -655,7 +656,7 @@ class LLMClient:
             )
 
         if text_content:
-            return json.loads(text_content)
+            return dict(dirtyjson.loads(text_content))
         return None
 
     async def _generate_anthropic_structured(

servers/fastapi/utils/llm_calls/generate_presentation_outlines.py

@@ -39,6 +39,8 @@ def get_system_prompt(
         - Do not generate table of contents slide.
         - Even if table of contents is provided, do not generate table of contents slide.
         {"- Always make first slide a title slide." if include_title_slide else "- Do not include title slide in the presentation."}
+
+        **Search web to get latest information about the topic**
     """
 
 

servers/fastapi/utils/llm_client_error_handler.py

@@ -2,9 +2,11 @@ from fastapi import HTTPException
 from anthropic import APIError as AnthropicAPIError
 from openai import APIError as OpenAIAPIError
 from google.genai.errors import APIError as GoogleAPIError
+import traceback
 
 
 def handle_llm_client_exceptions(e: Exception) -> HTTPException:
+    traceback.print_exc()
     if isinstance(e, OpenAIAPIError):
         return HTTPException(status_code=500, detail=f"OpenAI API error: {e.message}")
     if isinstance(e, GoogleAPIError):

servers/fastapi/uv.lock

@@ -471,6 +471,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/50/3d/9373ad9c56321fdab5b41197068e1d8c25883b3fea29dd361f9b55116869/dill-0.4.0-py3-none-any.whl", hash = "sha256:44f54bf6412c2c8464c14e8243eb163690a9800dbe2c367330883b19c7561049", size = 119668, upload-time = "2025-04-16T00:41:47.671Z" },
 ]
 
+[[package]]
+name = "dirtyjson"
+version = "1.0.8"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/db/04/d24f6e645ad82ba0ef092fa17d9ef7a21953781663648a01c9371d9e8e98/dirtyjson-1.0.8.tar.gz", hash = "sha256:90ca4a18f3ff30ce849d100dcf4a003953c79d3a2348ef056f1d9c22231a25fd", size = 30782, upload-time = "2022-11-28T23:32:33.319Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/68/69/1bcf70f81de1b4a9f21b3a62ec0c83bdff991c88d6cc2267d02408457e88/dirtyjson-1.0.8-py3-none-any.whl", hash = "sha256:125e27248435a58acace26d5c2c4c11a1c0de0a9c5124c5a94ba78e517d74f53", size = 25197, upload-time = "2022-11-28T23:32:31.219Z" },
+]
+
 [[package]]
 name = "distro"
 version = "1.9.0"
@@ -1908,6 +1917,7 @@ dependencies = [
     { name = "anthropic" },
     { name = "asyncpg" },
     { name = "chromadb" },
+    { name = "dirtyjson" },
     { name = "docling" },
     { name = "fastapi", extra = ["standard"] },
     { name = "fastmcp" },
@@ -1930,6 +1940,7 @@ requires-dist = [
     { name = "anthropic", specifier = ">=0.60.0" },
     { name = "asyncpg", specifier = ">=0.30.0" },
     { name = "chromadb", specifier = ">=1.0.15" },
+    { name = "dirtyjson", specifier = ">=1.0.8" },
     { name = "docling", specifier = ">=2.43.0" },
     { name = "fastapi", extras = ["standard"], specifier = ">=0.116.1" },
     { name = "fastmcp", specifier = ">=2.11.0" },