Spaces:
Runtime error
Runtime error
updatetool
Browse files
tools.py
CHANGED
|
@@ -17,6 +17,61 @@ from huggingface_hub import list_models
|
|
| 17 |
from pathlib import Path
|
| 18 |
import json
|
| 19 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
|
| 21 |
def extract_image_text(img_path: str) -> str:
|
| 22 |
"""
|
|
|
|
| 17 |
from pathlib import Path
|
| 18 |
import json
|
| 19 |
|
| 20 |
+
def wiki_search(query: str) -> str:
|
| 21 |
+
"""Search Wikipedia for a query and return maximum 2 results.
|
| 22 |
+
|
| 23 |
+
Args:
|
| 24 |
+
query: The search query."""
|
| 25 |
+
search_docs = WikipediaLoader(query=query, load_max_docs=2).load()
|
| 26 |
+
formatted_search_docs = "\n\n---\n\n".join(
|
| 27 |
+
[
|
| 28 |
+
f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
|
| 29 |
+
for doc in search_docs
|
| 30 |
+
])
|
| 31 |
+
return {"wiki_results": formatted_search_docs}
|
| 32 |
+
|
| 33 |
+
wiki_search_tool = Tool(
|
| 34 |
+
name="wiki_search",
|
| 35 |
+
func=wiki_search,
|
| 36 |
+
description="Searches for and retrieves information on Wikipedia using an inputted query."
|
| 37 |
+
)
|
| 38 |
+
|
| 39 |
+
def web_search(query: str) -> str:
|
| 40 |
+
"""Search Tavily for a query and return maximum 3 results.
|
| 41 |
+
|
| 42 |
+
Args:
|
| 43 |
+
query: The search query."""
|
| 44 |
+
search_docs = TavilySearchResults(max_results=3).invoke(query=query)
|
| 45 |
+
formatted_search_docs = "\n\n---\n\n".join(
|
| 46 |
+
[
|
| 47 |
+
f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
|
| 48 |
+
for doc in search_docs
|
| 49 |
+
])
|
| 50 |
+
return {"web_results": formatted_search_docs}
|
| 51 |
+
web_search_tool = Tool(
|
| 52 |
+
name="web_search",
|
| 53 |
+
func=web_search,
|
| 54 |
+
description="Searches for and retrieves information online using an inputted query using Tavily."
|
| 55 |
+
)
|
| 56 |
+
|
| 57 |
+
def arvix_search(query: str) -> str:
|
| 58 |
+
"""Search Arxiv for a query and return maximum 3 result.
|
| 59 |
+
|
| 60 |
+
Args:
|
| 61 |
+
query: The search query."""
|
| 62 |
+
search_docs = ArxivLoader(query=query, load_max_docs=3).load()
|
| 63 |
+
formatted_search_docs = "\n\n---\n\n".join(
|
| 64 |
+
[
|
| 65 |
+
f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content[:1000]}\n</Document>'
|
| 66 |
+
for doc in search_docs
|
| 67 |
+
])
|
| 68 |
+
return {"arvix_results": formatted_search_docs}
|
| 69 |
+
|
| 70 |
+
arvix_search_tool = Tool(
|
| 71 |
+
name="arvix_search",
|
| 72 |
+
func=arvix_search,
|
| 73 |
+
description="Searches for and retrieves information online using an inputted query on the website arvix, that is a database storing many factual scientific papers."
|
| 74 |
+
)
|
| 75 |
|
| 76 |
def extract_image_text(img_path: str) -> str:
|
| 77 |
"""
|