Spaces:
Paused
Paused
Upload search.py
Browse files
search.py
ADDED
|
@@ -0,0 +1,165 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import requests
|
| 2 |
+
from pydantic import BaseModel, Field
|
| 3 |
+
from typing import Callable, Any, Optional
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
class Valves(BaseModel):
|
| 7 |
+
SEARXNG_URL: str = Field(
|
| 8 |
+
default="https://search.ricky.cloudns.be/search",
|
| 9 |
+
description="SearXNG search URL",
|
| 10 |
+
)
|
| 11 |
+
BAIDU_URL: str = Field(
|
| 12 |
+
default="https://www.baidu.com/s", description="Baidu search URL"
|
| 13 |
+
)
|
| 14 |
+
# 添加 Bing 搜索 URL
|
| 15 |
+
BING_URL: str = Field(
|
| 16 |
+
default="https://www.bing.com/search", description="Bing search URL"
|
| 17 |
+
)
|
| 18 |
+
TIMEOUT: int = Field(default=30, description="Request timeout in seconds")
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
class Tools:
|
| 22 |
+
def __init__(self):
|
| 23 |
+
self.valves = Valves()
|
| 24 |
+
self.reader_api = "https://r.jina.ai/"
|
| 25 |
+
self.citation = True
|
| 26 |
+
|
| 27 |
+
async def read_url(
|
| 28 |
+
self, url: str, __event_emitter__: Optional[Callable[[dict], Any]] = None
|
| 29 |
+
) -> str:
|
| 30 |
+
"""
|
| 31 |
+
Read and extract the main content from a given URL.
|
| 32 |
+
:param url: The URL to read from.
|
| 33 |
+
:return: The main content of the page in clean, LLM-friendly text.
|
| 34 |
+
"""
|
| 35 |
+
try:
|
| 36 |
+
if __event_emitter__:
|
| 37 |
+
await __event_emitter__(
|
| 38 |
+
{
|
| 39 |
+
"type": "status",
|
| 40 |
+
"data": {
|
| 41 |
+
"description": f"Reading content from {url}",
|
| 42 |
+
"status": "in_progress",
|
| 43 |
+
},
|
| 44 |
+
}
|
| 45 |
+
)
|
| 46 |
+
|
| 47 |
+
data = {"url": url}
|
| 48 |
+
response = requests.post(
|
| 49 |
+
self.reader_api, data=data, timeout=self.valves.TIMEOUT
|
| 50 |
+
)
|
| 51 |
+
response.raise_for_status()
|
| 52 |
+
content = response.text
|
| 53 |
+
|
| 54 |
+
if __event_emitter__:
|
| 55 |
+
await __event_emitter__(
|
| 56 |
+
{
|
| 57 |
+
"type": "status",
|
| 58 |
+
"data": {
|
| 59 |
+
"description": "Content retrieved",
|
| 60 |
+
"status": "complete",
|
| 61 |
+
},
|
| 62 |
+
}
|
| 63 |
+
)
|
| 64 |
+
|
| 65 |
+
result_presentation = """
|
| 66 |
+
|
| 67 |
+
Result Presentation
|
| 68 |
+
STRICTLY ADHERE TO THIS FORMAT:
|
| 69 |
+
---
|
| 70 |
+
## [Site Name - Title](URL)
|
| 71 |
+
核心信息(Key Insights):
|
| 72 |
+
- Point 1
|
| 73 |
+
- Point 2
|
| 74 |
+
...
|
| 75 |
+
总结(Summary):
|
| 76 |
+
|
| 77 |
+
(Leave a blank line between each result summary)
|
| 78 |
+
---"""
|
| 79 |
+
|
| 80 |
+
return f"{content}\n---\n{result_presentation}"
|
| 81 |
+
except Exception as e:
|
| 82 |
+
if __event_emitter__:
|
| 83 |
+
await __event_emitter__(
|
| 84 |
+
{
|
| 85 |
+
"type": "status",
|
| 86 |
+
"data": {
|
| 87 |
+
"description": f"Error reading URL: {str(e)}",
|
| 88 |
+
"status": "error",
|
| 89 |
+
},
|
| 90 |
+
}
|
| 91 |
+
)
|
| 92 |
+
return f"Error reading URL: {str(e)}"
|
| 93 |
+
|
| 94 |
+
async def search(
|
| 95 |
+
self,
|
| 96 |
+
search_term: str,
|
| 97 |
+
engine: str = "google",
|
| 98 |
+
__event_emitter__: Optional[Callable[[dict], Any]] = None,
|
| 99 |
+
) -> str:
|
| 100 |
+
"""
|
| 101 |
+
Issues a query to a search engine and displays the results. (Note: Use this when the query requires gathering information from external sources, such as current events, unfamiliar terms, or when explicit browsing is requested.)
|
| 102 |
+
:param search_term: The search term or phrase to look up. (Note: think the keywords most likely to help users retrieve information as the "search_term",don't include search engine)
|
| 103 |
+
:param engine: The search engine to use ('google', 'baidu', or 'bing'). Note:bing(Default, suitable for most queries), baidu(Best for content about China), google(Relative equilibrium choices)
|
| 104 |
+
:return: The search results in text format.
|
| 105 |
+
"""
|
| 106 |
+
try:
|
| 107 |
+
if __event_emitter__:
|
| 108 |
+
await __event_emitter__(
|
| 109 |
+
{
|
| 110 |
+
"type": "status",
|
| 111 |
+
"data": {
|
| 112 |
+
"description": f"Searching with {engine}: {search_term}",
|
| 113 |
+
"status": "in_progress",
|
| 114 |
+
},
|
| 115 |
+
}
|
| 116 |
+
)
|
| 117 |
+
|
| 118 |
+
if engine == "baidu":
|
| 119 |
+
url = f"{self.reader_api}{self.valves.BAIDU_URL}?wd={search_term}"
|
| 120 |
+
headers = {"X-Target-Selector": "#content_left"}
|
| 121 |
+
elif engine == "bing":
|
| 122 |
+
# 添加 Bing 搜索的处理
|
| 123 |
+
url = f"{self.reader_api}{self.valves.BING_URL}?q={search_term}"
|
| 124 |
+
headers = {"X-Target-Selector": "#b_results"}
|
| 125 |
+
else:
|
| 126 |
+
url = f"{self.reader_api}{self.valves.SEARXNG_URL}?q={search_term}"
|
| 127 |
+
headers = {"X-Target-Selector": "#urls"}
|
| 128 |
+
|
| 129 |
+
response = requests.get(url, headers=headers, timeout=self.valves.TIMEOUT)
|
| 130 |
+
response.raise_for_status()
|
| 131 |
+
content = response.text
|
| 132 |
+
|
| 133 |
+
if __event_emitter__:
|
| 134 |
+
await __event_emitter__(
|
| 135 |
+
{
|
| 136 |
+
"type": "status",
|
| 137 |
+
"data": {
|
| 138 |
+
"description": f"Search completed with {engine}",
|
| 139 |
+
"status": "complete",
|
| 140 |
+
},
|
| 141 |
+
}
|
| 142 |
+
)
|
| 143 |
+
|
| 144 |
+
search_result_processing = """
|
| 145 |
+
Search Result Processing:
|
| 146 |
+
- Identify 1-3 highly relevant results from the search.
|
| 147 |
+
- Focus on the most informative and reliable sources.
|
| 148 |
+
- Integrate the key points from these sources to formulate a comprehensive answer to the query.
|
| 149 |
+
- If applicable, mention additional relevant information inferred from visible snippets in the search results.
|
| 150 |
+
"""
|
| 151 |
+
|
| 152 |
+
return f"{content}\n---\n{search_result_processing}"
|
| 153 |
+
|
| 154 |
+
except Exception as e:
|
| 155 |
+
if __event_emitter__:
|
| 156 |
+
await __event_emitter__(
|
| 157 |
+
{
|
| 158 |
+
"type": "status",
|
| 159 |
+
"data": {
|
| 160 |
+
"description": f"Error during search: {str(e)}",
|
| 161 |
+
"status": "error",
|
| 162 |
+
},
|
| 163 |
+
}
|
| 164 |
+
)
|
| 165 |
+
return f"Error during search: {str(e)}"
|