Merge branch 'langgenius:main' into add-turbo-pack

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

@@ -397,19 +397,44 @@ def _extract_text_from_csv(file_content: bytes) -> str:
         if not rows:
             return ""
 
+        # Combine multi-line text in the header row
+        header_row = [cell.replace("\n", " ").replace("\r", "") for cell in rows[0]]
+
         # Create Markdown table
-        markdown_table = "| " + " | ".join(rows[0]) + " |\n"
+        markdown_table = "| " + " | ".join(header_row) + " |\n"
-        markdown_table += "| " + " | ".join(["---"] * len(rows[0])) + " |\n"
+        markdown_table += "| " + " | ".join(["-" * len(col) for col in rows[0]]) + " |\n"
+
+        # Process each data row and combine multi-line text in each cell
         for row in rows[1:]:
-            markdown_table += "| " + " | ".join(row) + " |\n"
+            processed_row = [cell.replace("\n", " ").replace("\r", "") for cell in row]
+            markdown_table += "| " + " | ".join(processed_row) + " |\n"
 
-        return markdown_table.strip()
+        return markdown_table
     except Exception as e:
         raise TextExtractionError(f"Failed to extract text from CSV: {str(e)}") from e
 
 
 def _extract_text_from_excel(file_content: bytes) -> str:
     """Extract text from an Excel file using pandas."""
+
+    def _construct_markdown_table(df: pd.DataFrame) -> str:
+        """Manually construct a Markdown table from a DataFrame."""
+        # Construct the header row
+        header_row = "| " + " | ".join(df.columns) + " |"
+
+        # Construct the separator row
+        separator_row = "| " + " | ".join(["-" * len(col) for col in df.columns]) + " |"
+
+        # Construct the data rows
+        data_rows = []
+        for _, row in df.iterrows():
+            data_row = "| " + " | ".join(map(str, row)) + " |"
+            data_rows.append(data_row)
+
+        # Combine all rows into a single string
+        markdown_table = "\n".join([header_row, separator_row] + data_rows)
+        return markdown_table
+
     try:
         excel_file = pd.ExcelFile(io.BytesIO(file_content))
         markdown_table = ""
@@ -417,8 +442,15 @@ def _extract_text_from_excel(file_content: bytes) -> str:
             try:
                 df = excel_file.parse(sheet_name=sheet_name)
                 df.dropna(how="all", inplace=True)
-                # Create Markdown table two times to separate tables with a newline
+
-                markdown_table += df.to_markdown(index=False, floatfmt="") + "\n\n"
+                # Combine multi-line text in each cell into a single line
+                df = df.applymap(lambda x: " ".join(str(x).splitlines()) if isinstance(x, str) else x)  # type: ignore
+
+                # Combine multi-line text in column names into a single line
+                df.columns = pd.Index([" ".join(col.splitlines()) for col in df.columns])
+
+                # Manually construct the Markdown table
+                markdown_table += _construct_markdown_table(df) + "\n\n"
             except Exception as e:
                 continue
         return markdown_table

api/tests/unit_tests/core/workflow/nodes/test_document_extractor_node.py

@@ -1,5 +1,7 @@
+import io
 from unittest.mock import Mock, patch
 
+import pandas as pd
 import pytest
 from docx.oxml.text.paragraph import CT_P
 
@@ -187,145 +189,134 @@ def test_node_type(document_extractor_node):
 
 @patch("pandas.ExcelFile")
 def test_extract_text_from_excel_single_sheet(mock_excel_file):
-    """Test extracting text from Excel file with single sheet."""
+    """Test extracting text from Excel file with single sheet and multiline content."""
-    # Mock DataFrame
+
-    mock_df = Mock()
+    # Test multi-line cell
-    mock_df.dropna = Mock()
+    data = {"Name\nwith\nnewline": ["John\nDoe", "Jane\nSmith"], "Age": [25, 30]}
-    mock_df.to_markdown.return_value = "| Name | Age |\n|------|-----|\n| John | 25  |"
+
+    df = pd.DataFrame(data)
 
     # Mock ExcelFile
     mock_excel_instance = Mock()
     mock_excel_instance.sheet_names = ["Sheet1"]
-    mock_excel_instance.parse.return_value = mock_df
+    mock_excel_instance.parse.return_value = df
     mock_excel_file.return_value = mock_excel_instance
 
     file_content = b"fake_excel_content"
     result = _extract_text_from_excel(file_content)
+    expected_manual = "| Name with newline | Age |\n| ----------------- | --- |\n\
+| John Doe | 25 |\n| Jane Smith | 30 |\n\n"
 
-    expected = "| Name | Age |\n|------|-----|\n| John | 25  |\n\n"
+    assert expected_manual == result
-    assert result == expected
+    mock_excel_instance.parse.assert_called_once_with(sheet_name="Sheet1")
-    mock_excel_file.assert_called_once()
-    mock_df.dropna.assert_called_once_with(how="all", inplace=True)
-    mock_df.to_markdown.assert_called_once_with(index=False, floatfmt="")
 
 
 @patch("pandas.ExcelFile")
 def test_extract_text_from_excel_multiple_sheets(mock_excel_file):
-    """Test extracting text from Excel file with multiple sheets."""
+    """Test extracting text from Excel file with multiple sheets and multiline content."""
-    # Mock DataFrames for different sheets
-    mock_df1 = Mock()
-    mock_df1.dropna = Mock()
-    mock_df1.to_markdown.return_value = "| Product | Price |\n|---------|-------|\n| Apple   | 1.50  |"
 
-    mock_df2 = Mock()
+    # Test multi-line cell
-    mock_df2.dropna = Mock()
+    data1 = {"Product\nName": ["Apple\nRed", "Banana\nYellow"], "Price": [1.50, 0.99]}
-    mock_df2.to_markdown.return_value = "| City | Population |\n|------|------------|\n| NYC  | 8000000    |"
+    df1 = pd.DataFrame(data1)
+
+    data2 = {"City\nName": ["New\nYork", "Los\nAngeles"], "Population": [8000000, 3900000]}
+    df2 = pd.DataFrame(data2)
 
     # Mock ExcelFile
     mock_excel_instance = Mock()
     mock_excel_instance.sheet_names = ["Products", "Cities"]
-    mock_excel_instance.parse.side_effect = [mock_df1, mock_df2]
+    mock_excel_instance.parse.side_effect = [df1, df2]
     mock_excel_file.return_value = mock_excel_instance
 
     file_content = b"fake_excel_content_multiple_sheets"
     result = _extract_text_from_excel(file_content)
 
-    expected = (
+    expected_manual1 = "| Product Name | Price |\n| ------------ | ----- |\n\
-        "| Product | Price |\n|---------|-------|\n| Apple   | 1.50  |\n\n"
+| Apple Red | 1.5 |\n| Banana Yellow | 0.99 |\n\n"
-        "| City | Population |\n|------|------------|\n| NYC  | 8000000    |\n\n"
+    expected_manual2 = "| City Name | Population |\n| --------- | ---------- |\n\
-    )
+| New York | 8000000 |\n| Los Angeles | 3900000 |\n\n"
-    assert result == expected
+
+    assert expected_manual1 in result
+    assert expected_manual2 in result
+
     assert mock_excel_instance.parse.call_count == 2
 
 
 @patch("pandas.ExcelFile")
 def test_extract_text_from_excel_empty_sheets(mock_excel_file):
     """Test extracting text from Excel file with empty sheets."""
-    # Mock empty DataFrame
+
-    mock_df = Mock()
+    # Empty excel
-    mock_df.dropna = Mock()
+    df = pd.DataFrame()
-    mock_df.to_markdown.return_value = ""
 
     # Mock ExcelFile
     mock_excel_instance = Mock()
     mock_excel_instance.sheet_names = ["EmptySheet"]
-    mock_excel_instance.parse.return_value = mock_df
+    mock_excel_instance.parse.return_value = df
     mock_excel_file.return_value = mock_excel_instance
 
     file_content = b"fake_excel_empty_content"
     result = _extract_text_from_excel(file_content)
 
-    expected = "\n\n"
+    expected = "|  |\n|  |\n\n"
     assert result == expected
 
+    mock_excel_instance.parse.assert_called_once_with(sheet_name="EmptySheet")
+
 
 @patch("pandas.ExcelFile")
 def test_extract_text_from_excel_sheet_parse_error(mock_excel_file):
     """Test handling of sheet parsing errors - should continue with other sheets."""
-    # Mock DataFrames - one successful, one that raises exception
+
-    mock_df_success = Mock()
+    # Test error
-    mock_df_success.dropna = Mock()
+    data = {"Data": ["Test"], "Value": [123]}
-    mock_df_success.to_markdown.return_value = "| Data | Value |\n|------|-------|\n| Test | 123   |"
+    df = pd.DataFrame(data)
 
     # Mock ExcelFile
     mock_excel_instance = Mock()
     mock_excel_instance.sheet_names = ["GoodSheet", "BadSheet"]
-    mock_excel_instance.parse.side_effect = [mock_df_success, Exception("Parse error")]
+    mock_excel_instance.parse.side_effect = [df, Exception("Parse error")]
     mock_excel_file.return_value = mock_excel_instance
 
     file_content = b"fake_excel_mixed_content"
     result = _extract_text_from_excel(file_content)
 
-    expected = "| Data | Value |\n|------|-------|\n| Test | 123   |\n\n"
+    expected_manual = "| Data | Value |\n| ---- | ----- |\n| Test | 123 |\n\n"
-    assert result == expected
-
-
-@patch("pandas.ExcelFile")
-def test_extract_text_from_excel_file_error(mock_excel_file):
-    """Test handling of Excel file reading errors."""
-    mock_excel_file.side_effect = Exception("Invalid Excel file")
-
-    file_content = b"invalid_excel_content"
 
-    with pytest.raises(Exception) as exc_info:
+    assert expected_manual == result
-        _extract_text_from_excel(file_content)
 
-    # Note: The function should raise TextExtractionError, but since it's not imported in the test,
+    assert mock_excel_instance.parse.call_count == 2
-    # we check for the general Exception pattern
-    assert "Failed to extract text from Excel file" in str(exc_info.value)
 
 
 @patch("pandas.ExcelFile")
 def test_extract_text_from_excel_io_bytesio_usage(mock_excel_file):
     """Test that BytesIO is properly used with the file content."""
-    import io
 
-    # Mock DataFrame
+    # Test bytesio
-    mock_df = Mock()
+    data = {"Test": [1], "Data": ["A"]}
-    mock_df.dropna = Mock()
+    df = pd.DataFrame(data)
-    mock_df.to_markdown.return_value = "| Test | Data |\n|------|------|\n| 1    | A    |"
 
     # Mock ExcelFile
     mock_excel_instance = Mock()
     mock_excel_instance.sheet_names = ["TestSheet"]
-    mock_excel_instance.parse.return_value = mock_df
+    mock_excel_instance.parse.return_value = df
     mock_excel_file.return_value = mock_excel_instance
 
     file_content = b"test_excel_bytes"
     result = _extract_text_from_excel(file_content)
 
-    # Verify that ExcelFile was called with a BytesIO object
     mock_excel_file.assert_called_once()
-    call_args = mock_excel_file.call_args[0][0]
+    call_arg = mock_excel_file.call_args[0][0]
-    assert isinstance(call_args, io.BytesIO)
+    assert isinstance(call_arg, io.BytesIO)
 
-    expected = "| Test | Data |\n|------|------|\n| 1    | A    |\n\n"
+    expected_manual = "| Test | Data |\n| ---- | ---- |\n| 1 | A |\n\n"
-    assert result == expected
+    assert expected_manual == result
 
 
 @patch("pandas.ExcelFile")
 def test_extract_text_from_excel_all_sheets_fail(mock_excel_file):
     """Test when all sheets fail to parse - should return empty string."""
+
     # Mock ExcelFile
     mock_excel_instance = Mock()
     mock_excel_instance.sheet_names = ["BadSheet1", "BadSheet2"]
@@ -335,29 +326,6 @@ def test_extract_text_from_excel_all_sheets_fail(mock_excel_file):
     file_content = b"fake_excel_all_bad_sheets"
     result = _extract_text_from_excel(file_content)
 
-    # Should return empty string when all sheets fail
     assert result == ""
 
-
+    assert mock_excel_instance.parse.call_count == 2
-@patch("pandas.ExcelFile")
-def test_extract_text_from_excel_markdown_formatting(mock_excel_file):
-    """Test that markdown formatting parameters are correctly applied."""
-    # Mock DataFrame
-    mock_df = Mock()
-    mock_df.dropna = Mock()
-    mock_df.to_markdown.return_value = "| Float | Int |\n|-------|-----|\n| 123456.78 | 42  |"
-
-    # Mock ExcelFile
-    mock_excel_instance = Mock()
-    mock_excel_instance.sheet_names = ["NumberSheet"]
-    mock_excel_instance.parse.return_value = mock_df
-    mock_excel_file.return_value = mock_excel_instance
-
-    file_content = b"fake_excel_numbers"
-    result = _extract_text_from_excel(file_content)
-
-    # Verify to_markdown was called with correct parameters
-    mock_df.to_markdown.assert_called_once_with(index=False, floatfmt="")
-
-    expected = "| Float | Int |\n|-------|-----|\n| 123456.78 | 42  |\n\n"
-    assert result == expected

sdks/python-client/dify_client/client.py

@@ -47,7 +47,7 @@ class DifyClient:
 
     def text_to_audio(self, text: str, user: str, streaming: bool = False):
         data = {"text": text, "user": user, "streaming": streaming}
-        return self._send_request("POST", "/text-to-audio", data=data)
+        return self._send_request("POST", "/text-to-audio", json=data)
 
     def get_meta(self, user):
         params = {"user": user}

web/app/(commonLayout)/apps/Apps.tsx

@@ -88,11 +88,11 @@ const Apps = () => {
   const anchorRef = useRef<HTMLDivElement>(null)
   const options = [
     { value: 'all', text: t('app.types.all'), icon: <RiApps2Line className='mr-1 h-[14px] w-[14px]' /> },
+    { value: 'workflow', text: t('app.types.workflow'), icon: <RiExchange2Line className='mr-1 h-[14px] w-[14px]' /> },
+    { value: 'advanced-chat', text: t('app.types.advanced'), icon: <RiMessage3Line className='mr-1 h-[14px] w-[14px]' /> },
     { value: 'chat', text: t('app.types.chatbot'), icon: <RiMessage3Line className='mr-1 h-[14px] w-[14px]' /> },
     { value: 'agent-chat', text: t('app.types.agent'), icon: <RiRobot3Line className='mr-1 h-[14px] w-[14px]' /> },
     { value: 'completion', text: t('app.types.completion'), icon: <RiFile4Line className='mr-1 h-[14px] w-[14px]' /> },
-    { value: 'advanced-chat', text: t('app.types.advanced'), icon: <RiMessage3Line className='mr-1 h-[14px] w-[14px]' /> },
-    { value: 'workflow', text: t('app.types.workflow'), icon: <RiExchange2Line className='mr-1 h-[14px] w-[14px]' /> },
   ]
 
   useEffect(() => {

web/app/components/app/create-app-modal/index.tsx

@@ -1,9 +1,9 @@
 'use client'
 
-import { useCallback, useEffect, useRef, useState } from 'react'
+import { useCallback, useRef, useState } from 'react'
 import { useTranslation } from 'react-i18next'
 
-import { useRouter, useSearchParams } from 'next/navigation'
+import { useRouter } from 'next/navigation'
 import { useContext, useContextSelector } from 'use-context-selector'
 import { RiArrowRightLine, RiArrowRightSLine, RiCommandLine, RiCornerDownLeftLine, RiExchange2Fill } from '@remixicon/react'
 import Link from 'next/link'
@@ -19,7 +19,6 @@ import AppsContext, { useAppContext } from '@/context/app-context'
 import { useProviderContext } from '@/context/provider-context'
 import { ToastContext } from '@/app/components/base/toast'
 import type { AppMode } from '@/types/app'
-import { AppModes } from '@/types/app'
 import { createApp } from '@/service/apps'
 import Input from '@/app/components/base/input'
 import Textarea from '@/app/components/base/textarea'
@@ -56,14 +55,6 @@ function CreateApp({ onClose, onSuccess, onCreateFromTemplate }: CreateAppProps)
 
   const isCreatingRef = useRef(false)
 
-  const searchParams = useSearchParams()
-
-  useEffect(() => {
-    const category = searchParams.get('category')
-    if (category && AppModes.includes(category as AppMode))
-      setAppMode(category as AppMode)
-  }, [searchParams])
-
   const onCreate = useCallback(async () => {
     if (!appMode) {
       notify({ type: 'error', message: t('app.newApp.appTypeRequired') })

web/app/components/app/type-selector/index.tsx

@@ -15,7 +15,7 @@ export type AppSelectorProps = {
   onChange: (value: AppSelectorProps['value']) => void
 }
 
-const allTypes: AppMode[] = ['chat', 'agent-chat', 'completion', 'advanced-chat', 'workflow']
+const allTypes: AppMode[] = ['workflow', 'advanced-chat', 'chat', 'agent-chat', 'completion']
 
 const AppTypeSelector = ({ value, onChange }: AppSelectorProps) => {
   const [open, setOpen] = useState(false)

web/app/components/base/chat/chat/hooks.ts

@@ -366,7 +366,7 @@ export const useChat = (
             if (!newResponseItem)
               return
 
-            const isUseAgentThought = newResponseItem.agent_thoughts?.length > 0
+            const isUseAgentThought = newResponseItem.agent_thoughts?.length > 0 && newResponseItem.agent_thoughts[newResponseItem.agent_thoughts?.length - 1].thought === newResponseItem.answer
             updateChatTreeNode(responseItem.id, {
               content: isUseAgentThought ? '' : newResponseItem.answer,
               log: [

web/app/components/base/radio/component/radio/index.tsx

@@ -38,7 +38,7 @@ export default function Radio({
   const divClassName = `
     flex items-center py-1 relative
     px-7 cursor-pointer text-text-secondary rounded
-    bg-components-option-card-option-bg hover:bg-components-option-card-option-bg-hover hover:shadow-xs
+    hover:bg-components-option-card-option-bg-hover hover:shadow-xs
   `
 
   return (