Merge branch 'fix/webapp-upload-file' into deploy/dev

api/core/tools/provider/builtin/podcast_generator/tools/podcast_audio_generator.py

@@ -1,8 +1,8 @@
 import concurrent.futures
 import io
 import random
+import warnings
 from typing import Any, Literal, Optional, Union
-from warnings import catch_warnings
 
 import openai
 
@@ -10,7 +10,8 @@ from core.tools.entities.tool_entities import ToolInvokeMessage
 from core.tools.errors import ToolParameterValidationError, ToolProviderCredentialValidationError
 from core.tools.tool.builtin_tool import BuiltinTool
 
-with catch_warnings(action="ignore", category=RuntimeWarning):
+with warnings.catch_warnings():
+    warnings.simplefilter("ignore")
     from pydub import AudioSegment
 
 

api/core/workflow/nodes/document_extractor/node.py

@@ -6,12 +6,14 @@ import docx
 import pandas as pd
 import pypdfium2
 import yaml
+from unstructured.partition.api import partition_via_api
 from unstructured.partition.email import partition_email
 from unstructured.partition.epub import partition_epub
 from unstructured.partition.msg import partition_msg
 from unstructured.partition.ppt import partition_ppt
 from unstructured.partition.pptx import partition_pptx
 
+from configs import dify_config
 from core.file import File, FileTransferMethod, file_manager
 from core.helper import ssrf_proxy
 from core.variables import ArrayFileSegment
@@ -263,7 +265,14 @@ def _extract_text_from_ppt(file_content: bytes) -> str:
 def _extract_text_from_pptx(file_content: bytes) -> str:
     try:
         with io.BytesIO(file_content) as file:
-            elements = partition_pptx(file=file)
+            if dify_config.UNSTRUCTURED_API_URL and dify_config.UNSTRUCTURED_API_KEY:
+                elements = partition_via_api(
+                    file=file,
+                    api_url=dify_config.UNSTRUCTURED_API_URL,
+                    api_key=dify_config.UNSTRUCTURED_API_KEY,
+                )
+            else:
+                elements = partition_pptx(file=file)
         return "\n".join([getattr(element, "text", "") for element in elements])
     except Exception as e:
         raise TextExtractionError(f"Failed to extract text from PPTX: {str(e)}") from e

web/app/components/base/file-uploader/hooks.ts

@@ -216,7 +216,7 @@ export const useFile = (fileConfig: FileUpload) => {
     handleAddFile(uploadingFile)
     startProgressTimer(uploadingFile.id)
 
-    uploadRemoteFileInfo(url).then((res) => {
+    uploadRemoteFileInfo(url, !!params.token).then((res) => {
       const newFile = {
         ...uploadingFile,
         type: res.mime_type,

web/service/common.ts

@@ -324,8 +324,8 @@ export const verifyForgotPasswordToken: Fetcher<CommonResponse & { is_valid: boo
 export const changePasswordWithToken: Fetcher<CommonResponse, { url: string; body: { token: string; new_password: string; password_confirm: string } }> = ({ url, body }) =>
   post<CommonResponse>(url, { body })
 
-export const uploadRemoteFileInfo = (url: string) => {
-  return post<{ id: string; name: string; size: number; mime_type: string; url: string }>('/remote-files/upload', { body: { url } })
+export const uploadRemoteFileInfo = (url: string, isPublic?: boolean) => {
+  return post<{ id: string; name: string; size: number; mime_type: string; url: string }>('/remote-files/upload', { body: { url } }, { isPublicAPI: isPublic })
 }
 
 export const sendEMailLoginCode = (email: string, language = 'en-US') =>