Merge pull request #1436 from ManishMadan2882/main

React Widget: Search bar component

application/api/user/routes.py

@@ -339,6 +339,9 @@ def post(self):
                         ".json",
                         ".xlsx",
                         ".pptx",
+                        ".png",
+                        ".jpg",
+                        ".jpeg",
                     ],
                     job_name,
                     final_filename,
@@ -365,6 +368,9 @@ def post(self):
                         ".json",
                         ".xlsx",
                         ".pptx",
+                        ".png",
+                        ".jpg",
+                        ".jpeg",
                     ],
                     job_name,
                     final_filename,

application/core/settings.py

@@ -18,6 +18,7 @@ class Settings(BaseSettings):
     DEFAULT_MAX_HISTORY: int = 150
     MODEL_TOKEN_LIMITS: dict = {"gpt-3.5-turbo": 4096, "claude-2": 1e5}
     UPLOAD_FOLDER: str = "inputs"
+    PARSE_PDF_AS_IMAGE: bool = False
     VECTOR_STORE: str = "faiss" #  "faiss" or "elasticsearch" or "qdrant" or "milvus" or "lancedb"
     RETRIEVERS_ENABLED: list = ["classic_rag", "duckduck_search"] # also brave_search
 

application/parser/file/bulk.py

@@ -13,6 +13,7 @@
 from application.parser.file.tabular_parser import PandasCSVParser,ExcelParser
 from application.parser.file.json_parser import JSONParser
 from application.parser.file.pptx_parser import PPTXParser
+from application.parser.file.image_parser import ImageParser
 from application.parser.schema.base import Document
 
 DEFAULT_FILE_EXTRACTOR: Dict[str, BaseParser] = {
@@ -27,6 +28,9 @@
     ".mdx": MarkdownParser(),
     ".json":JSONParser(),
     ".pptx":PPTXParser(),
+    ".png": ImageParser(),
+    ".jpg": ImageParser(),
+    ".jpeg": ImageParser(),
 }
 
 

application/parser/file/docs_parser.py

@@ -7,7 +7,8 @@
 from typing import Dict
 
 from application.parser.file.base_parser import BaseParser
-
+from application.core.settings import settings
+import requests
 
 class PDFParser(BaseParser):
     """PDF parser."""
@@ -18,6 +19,15 @@ def _init_parser(self) -> Dict:
 
     def parse_file(self, file: Path, errors: str = "ignore") -> str:
         """Parse file."""
+        if settings.PARSE_PDF_AS_IMAGE:
+            doc2md_service = "https://llm.arc53.com/doc2md"
+            # alternatively you can use local vision capable LLM
+            with open(file, "rb") as file_loaded:
+                files = {'file': file_loaded}
+                response = requests.post(doc2md_service, files=files)   
+                data = response.json()["markdown"] 
+            return data
+
         try:
             import PyPDF2
         except ImportError:

application/parser/file/image_parser.py

@@ -0,0 +1,27 @@
+"""Image parser.
+
+Contains parser for .png, .jpg, .jpeg files.
+
+"""
+from pathlib import Path
+import requests
+from typing import Dict, Union
+
+from application.parser.file.base_parser import BaseParser
+
+
+class ImageParser(BaseParser):
+    """Image parser."""
+
+    def _init_parser(self) -> Dict:
+        """Init parser."""
+        return {}
+
+    def parse_file(self, file: Path, errors: str = "ignore") -> Union[str, list[str]]:
+        doc2md_service = "https://llm.arc53.com/doc2md"
+        # alternatively you can use local vision capable LLM
+        with open(file, "rb") as file_loaded:
+            files = {'file': file_loaded}
+            response = requests.post(doc2md_service, files=files)   
+            data = response.json()["markdown"] 
+        return data

extensions/react-widget/package.json

@@ -32,7 +32,8 @@
   "scripts": {
     "build": "parcel build src/main.tsx --public-url ./",
     "build:react": "parcel build src/index.ts",
-    "dev": "parcel src/index.html -p 3000",
+   "serve": "parcel serve -p 3000",
+    "dev": "parcel -p 3000",
     "test": "jest",
     "lint": "eslint",
     "check": "tsc --noEmit",

extensions/react-widget/src/App.tsx

@@ -1,11 +1,11 @@
 import React from "react"
 import {DocsGPTWidget} from "./components/DocsGPTWidget"
-const App = () => {
+import {SearchBar} from "./components/SearchBar"
+export const App = () => {
   return (
     <div>
+      <SearchBar/>
       <DocsGPTWidget/>
     </div>
   )
-}
-
-export default App
+}