feat: add new proxy rotation function

scrapegraphai/nodes/fetch_node.py

@@ -7,6 +7,7 @@
 from langchain_core.documents import Document
 from .base_node import BaseNode
 from ..utils.remover import remover
+from ..utils.proxy_generator import proxy_generator
 
 
 class FetchNode(BaseNode):
@@ -37,13 +38,16 @@ class FetchNode(BaseNode):
                         to succeed.
     """
 
-    def __init__(self, input: str, output: List[str], node_name: str = "Fetch"):
+    def __init__(self, input: str, output: List[str], num_prox: int = True,
+                 node_name: str = "Fetch"):
         """
         Initializes the FetchHTMLNode with a node name and node type.
         Arguments:
             node_name (str): name of the node
+            prox_rotation (bool): if you wamt to rotate proxies
         """
         super().__init__(node_name, "node", input, output, 1)
+        self.num_prox = num_prox
 
     def execute(self, state):
         """
@@ -78,7 +82,11 @@ def execute(self, state):
 
         # if it is a URL
         else:
-            loader = AsyncHtmlLoader(source)
+            if self.num_prox > 1:
+                loader = AsyncHtmlLoader(
+                    source, proxies=proxy_generator(self.num_prox))
+            else:
+                loader = AsyncHtmlLoader(source)
             document = loader.load()
             compressed_document = [
                 Document(page_content=remover(str(document)))]

scrapegraphai/utils/__init__.py

@@ -5,3 +5,4 @@
 from .convert_to_csv import convert_to_csv
 from .convert_to_json import convert_to_json
 from .prettify_exec_info import prettify_exec_info
+from .proxy_generator import proxy_generator

scrapegraphai/utils/proxy_rotation.py

@@ -0,0 +1,31 @@
+"""
+Module for rotating proxies
+"""
+from fp.fp import FreeProxy
+
+
+def proxy_generator(num_ips: int):
+    """
+    Rotates through a specified number of proxy IPs using the FreeProxy library.
+
+    Args:
+        num_ips (int): The number of proxy IPs to rotate through.
+
+    Returns:
+        dict: A dictionary containing the rotated proxy IPs, indexed by their position in rotation.
+
+    Example:
+        >>> proxy_generator(5)
+        {
+            0: '192.168.1.1:8080',
+            1: '103.10.63.135:8080',
+            2: '176.9.75.42:8080',
+            3: '37.57.216.2:8080',
+            4: '113.20.31.250:8080'
+        }
+    """
+    res = []
+
+    for i in range(0, num_ips):
+        res.append(FreeProxy().get())
+    return res