Squashed 'packages/tools/' changes from 78317b9c..0b3f00e6

0b3f00e6 chore: update project version to 0.73.0 and revise uv.lock dependencies (#455)
ad19b074 feat: replace embedchain with native crewai adapter (#451)

git-subtree-dir: packages/tools
git-subtree-split: 0b3f00e67c0dae24d188c292dc99759fd1c841f7

tests/tools/test_search_tools.py

@@ -1,11 +1,11 @@
 import os
 import tempfile
 from pathlib import Path
-from unittest.mock import ANY, MagicMock
+from unittest.mock import MagicMock
 
 import pytest
-from embedchain.models.data_type import DataType
 
+from crewai_tools.rag.data_types import DataType
 from crewai_tools.tools import (
     CodeDocsSearchTool,
     CSVSearchTool,
@@ -49,7 +49,7 @@ def test_pdf_search_tool(mock_adapter):
     result = tool._run(query="test content")
     assert "this is a test" in result.lower()
     mock_adapter.add.assert_called_once_with("test.pdf", data_type=DataType.PDF_FILE)
-    mock_adapter.query.assert_called_once_with("test content")
+    mock_adapter.query.assert_called_once_with("test content", similarity_threshold=0.6, limit=5)
 
     mock_adapter.query.reset_mock()
     mock_adapter.add.reset_mock()
@@ -58,7 +58,7 @@ def test_pdf_search_tool(mock_adapter):
     result = tool._run(pdf="test.pdf", query="test content")
     assert "this is a test" in result.lower()
     mock_adapter.add.assert_called_once_with("test.pdf", data_type=DataType.PDF_FILE)
-    mock_adapter.query.assert_called_once_with("test content")
+    mock_adapter.query.assert_called_once_with("test content", similarity_threshold=0.6, limit=5)
 
 
 def test_txt_search_tool():
@@ -82,7 +82,7 @@ def test_docx_search_tool(mock_adapter):
     result = tool._run(search_query="test content")
     assert "this is a test" in result.lower()
     mock_adapter.add.assert_called_once_with("test.docx", data_type=DataType.DOCX)
-    mock_adapter.query.assert_called_once_with("test content")
+    mock_adapter.query.assert_called_once_with("test content", similarity_threshold=0.6, limit=5)
 
     mock_adapter.query.reset_mock()
     mock_adapter.add.reset_mock()
@@ -91,7 +91,7 @@ def test_docx_search_tool(mock_adapter):
     result = tool._run(docx="test.docx", search_query="test content")
     assert "this is a test" in result.lower()
     mock_adapter.add.assert_called_once_with("test.docx", data_type=DataType.DOCX)
-    mock_adapter.query.assert_called_once_with("test content")
+    mock_adapter.query.assert_called_once_with("test content", similarity_threshold=0.6, limit=5)
 
 
 def test_json_search_tool():
@@ -114,7 +114,7 @@ def test_xml_search_tool(mock_adapter):
     result = tool._run(search_query="test XML", xml="test.xml")
     assert "this is a test" in result.lower()
     mock_adapter.add.assert_called_once_with("test.xml")
-    mock_adapter.query.assert_called_once_with("test XML")
+    mock_adapter.query.assert_called_once_with("test XML", similarity_threshold=0.6, limit=5)
 
 
 def test_csv_search_tool():
@@ -153,8 +153,8 @@ def test_website_search_tool(mock_adapter):
     tool = WebsiteSearchTool(website=website, adapter=mock_adapter)
     result = tool._run(search_query=search_query)
 
-    mock_adapter.query.assert_called_once_with("what is crewai?")
-    mock_adapter.add.assert_called_once_with(website, data_type=DataType.WEB_PAGE)
+    mock_adapter.query.assert_called_once_with("what is crewai?", similarity_threshold=0.6, limit=5)
+    mock_adapter.add.assert_called_once_with(website, data_type=DataType.WEBSITE)
 
     assert "this is a test" in result.lower()
 
@@ -164,8 +164,8 @@ def test_website_search_tool(mock_adapter):
     tool = WebsiteSearchTool(adapter=mock_adapter)
     result = tool._run(website=website, search_query=search_query)
 
-    mock_adapter.query.assert_called_once_with("what is crewai?")
-    mock_adapter.add.assert_called_once_with(website, data_type=DataType.WEB_PAGE)
+    mock_adapter.query.assert_called_once_with("what is crewai?", similarity_threshold=0.6, limit=5)
+    mock_adapter.add.assert_called_once_with(website, data_type=DataType.WEBSITE)
 
     assert "this is a test" in result.lower()
 
@@ -185,7 +185,7 @@ def test_youtube_video_search_tool(mock_adapter):
     mock_adapter.add.assert_called_once_with(
         youtube_video_url, data_type=DataType.YOUTUBE_VIDEO
     )
-    mock_adapter.query.assert_called_once_with(search_query)
+    mock_adapter.query.assert_called_once_with(search_query, similarity_threshold=0.6, limit=5)
 
     mock_adapter.query.reset_mock()
     mock_adapter.add.reset_mock()
@@ -197,7 +197,7 @@ def test_youtube_video_search_tool(mock_adapter):
     mock_adapter.add.assert_called_once_with(
         youtube_video_url, data_type=DataType.YOUTUBE_VIDEO
     )
-    mock_adapter.query.assert_called_once_with(search_query)
+    mock_adapter.query.assert_called_once_with(search_query, similarity_threshold=0.6, limit=5)
 
 
 def test_youtube_channel_search_tool(mock_adapter):
@@ -213,7 +213,7 @@ def test_youtube_channel_search_tool(mock_adapter):
     mock_adapter.add.assert_called_once_with(
         youtube_channel_handle, data_type=DataType.YOUTUBE_CHANNEL
     )
-    mock_adapter.query.assert_called_once_with(search_query)
+    mock_adapter.query.assert_called_once_with(search_query, similarity_threshold=0.6, limit=5)
 
     mock_adapter.query.reset_mock()
     mock_adapter.add.reset_mock()
@@ -227,7 +227,7 @@ def test_youtube_channel_search_tool(mock_adapter):
     mock_adapter.add.assert_called_once_with(
         youtube_channel_handle, data_type=DataType.YOUTUBE_CHANNEL
     )
-    mock_adapter.query.assert_called_once_with(search_query)
+    mock_adapter.query.assert_called_once_with(search_query, similarity_threshold=0.6, limit=5)
 
 
 def test_code_docs_search_tool(mock_adapter):
@@ -239,7 +239,7 @@ def test_code_docs_search_tool(mock_adapter):
     result = tool._run(search_query=search_query)
     assert "test documentation" in result
     mock_adapter.add.assert_called_once_with(docs_url, data_type=DataType.DOCS_SITE)
-    mock_adapter.query.assert_called_once_with(search_query)
+    mock_adapter.query.assert_called_once_with(search_query, similarity_threshold=0.6, limit=5)
 
     mock_adapter.query.reset_mock()
     mock_adapter.add.reset_mock()
@@ -248,7 +248,7 @@ def test_code_docs_search_tool(mock_adapter):
     result = tool._run(docs_url=docs_url, search_query=search_query)
     assert "test documentation" in result
     mock_adapter.add.assert_called_once_with(docs_url, data_type=DataType.DOCS_SITE)
-    mock_adapter.query.assert_called_once_with(search_query)
+    mock_adapter.query.assert_called_once_with(search_query, similarity_threshold=0.6, limit=5)
 
 
 def test_github_search_tool(mock_adapter):
@@ -264,9 +264,11 @@ def test_github_search_tool(mock_adapter):
     result = tool._run(search_query="tell me about crewai repo")
     assert "repo description" in result
     mock_adapter.add.assert_called_once_with(
-        "repo:crewai/crewai type:code", data_type="github", loader=ANY
+        "https://github.com/crewai/crewai",
+        data_type=DataType.GITHUB,
+        metadata={"content_types": ["code"], "gh_token": "test_token"}
     )
-    mock_adapter.query.assert_called_once_with("tell me about crewai repo")
+    mock_adapter.query.assert_called_once_with("tell me about crewai repo", similarity_threshold=0.6, limit=5)
 
     # ensure content types provided by run call is used
     mock_adapter.query.reset_mock()
@@ -280,9 +282,11 @@ def test_github_search_tool(mock_adapter):
     )
     assert "repo description" in result
     mock_adapter.add.assert_called_once_with(
-        "repo:crewai/crewai type:code,issue", data_type="github", loader=ANY
+        "https://github.com/crewai/crewai",
+        data_type=DataType.GITHUB,
+        metadata={"content_types": ["code", "issue"], "gh_token": "test_token"}
     )
-    mock_adapter.query.assert_called_once_with("tell me about crewai repo")
+    mock_adapter.query.assert_called_once_with("tell me about crewai repo", similarity_threshold=0.6, limit=5)
 
     # ensure default content types are used if not provided
     mock_adapter.query.reset_mock()
@@ -295,9 +299,11 @@ def test_github_search_tool(mock_adapter):
     )
     assert "repo description" in result
     mock_adapter.add.assert_called_once_with(
-        "repo:crewai/crewai type:code,repo,pr,issue", data_type="github", loader=ANY
+        "https://github.com/crewai/crewai",
+        data_type=DataType.GITHUB,
+        metadata={"content_types": ["code", "repo", "pr", "issue"], "gh_token": "test_token"}
     )
-    mock_adapter.query.assert_called_once_with("tell me about crewai repo")
+    mock_adapter.query.assert_called_once_with("tell me about crewai repo", similarity_threshold=0.6, limit=5)
 
     # ensure nothing is added if no repo is provided
     mock_adapter.query.reset_mock()
@@ -306,4 +312,4 @@ def test_github_search_tool(mock_adapter):
     tool = GithubSearchTool(gh_token="test_token", adapter=mock_adapter)
     result = tool._run(search_query="tell me about crewai repo")
     mock_adapter.add.assert_not_called()
-    mock_adapter.query.assert_called_once_with("tell me about crewai repo")
+    mock_adapter.query.assert_called_once_with("tell me about crewai repo", similarity_threshold=0.6, limit=5)