455 lines
12 KiB
Plaintext
455 lines
12 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "a92df66b-68c9-4288-b881-45d1fd948c18",
|
|
"metadata": {},
|
|
"source": [
|
|
"### Week 1 Contribution: Selenium-enhanced Website Summarizer\n",
|
|
"This notebook attempts to summarize content from any website using a BeautifulSoup-first strategy with a Selenium fallback for JavaScript-heavy pages. Llama 3.2 is used to generate a markdown-formatted summary.\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "407ea4b4-7c1b-4f94-a48d-f3ee3273bc61",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"import os\n",
|
|
"import requests\n",
|
|
"from bs4 import BeautifulSoup\n",
|
|
"from IPython.display import Markdown,display\n",
|
|
"from openai import OpenAI"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "040e97a8-9a5f-4903-9d0e-fa19bb719b4f",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"MODEL=\"llama3.2\"\n",
|
|
"openai=OpenAI(base_url=\"http://localhost:11434/v1\",api_key=\"ollama\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "cac3c9ae-31ce-45b1-bbc1-70577a198e84",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"message=\"Hi, write a snarky poem for me.\" \n",
|
|
"response=openai.chat.completions.create(\n",
|
|
" model=MODEL,\n",
|
|
" messages=[{\n",
|
|
" \"role\":\"user\",\n",
|
|
" \"content\":message\n",
|
|
" }]\n",
|
|
")\n",
|
|
"print(response.choices[0].message.content)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "a27514f6-d7a5-4292-b98b-dc166416a2fc",
|
|
"metadata": {},
|
|
"source": [
|
|
"### Beautiful Soup Version"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "678901b6-5da1-4df7-8b73-a1c69dc758b0",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"headers = {\n",
|
|
" \"User-Agent\": \"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/117.0.0.0 Safari/537.36\"\n",
|
|
"} # to make sure we're not blocked as bots from websites\n",
|
|
"\n",
|
|
"class bsWebsite:\n",
|
|
" \"\"\"\n",
|
|
" Attributes:\n",
|
|
" url (str): The URL of the page\n",
|
|
" title (str): The title of the page\n",
|
|
" text (str): The readable text from the page\n",
|
|
" \"\"\"\n",
|
|
"\n",
|
|
" def __init__(self,url):\n",
|
|
" self.url=url\n",
|
|
" response=requests.get(url,headers=headers) # gets the content of the page in response variable\n",
|
|
"\n",
|
|
" soup=BeautifulSoup(response.content,'html.parser') # content of response is accessed using html parser for structure\n",
|
|
" self.title=soup.title.string if soup.title else \"No title\"\n",
|
|
"\n",
|
|
" for irrelevant in soup.body([\"script\", \"style\", \"img\", \"input\"]):\n",
|
|
" irrelevant.decompose()\n",
|
|
"\n",
|
|
" self.text=soup.body.get_text(separator='\\n',strip=True)\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "9a1a5ddd-7907-46fd-a1b7-ceeb876262f7",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"ed = bsWebsite(\"https://edwarddonner.com\")\n",
|
|
"\n",
|
|
"print(ed.url)\n",
|
|
"print(ed.text)\n",
|
|
"print(ed.title)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "b7e965e4-7d20-4980-8cb2-871b8ca63c45",
|
|
"metadata": {},
|
|
"source": [
|
|
"#### Now, let's create a detailed summary for how selenium works using what we just made"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "b71a05c6-669b-4632-aeb9-b51daa4429a1",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"sel=bsWebsite(\"https://www.geeksforgeeks.org/software-engineering/selenium-webdriver-tutorial/\")\n",
|
|
"print(sel.url)\n",
|
|
"print(sel.title)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "4c463c67-2a9c-4fcd-99aa-cab0e2cdf936",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"def user_prompt_for(web):\n",
|
|
" user_prompt=f\"\"\"You are looking at a website called {web.title}. \n",
|
|
" Provide a detailed summary of the given content and the concepts in markdown:\\n[{web.text}]\"\"\"\n",
|
|
"\n",
|
|
" return user_prompt"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "b2118ac4-3355-4f90-b799-ba375ceeafc1",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"system_prompt=\"\"\"You are an assistant that analyses the contents of a website based on request of user, \n",
|
|
"while ignoring text that is navigation related. Respond in markdown.\"\"\""
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "716b3772-3c73-4010-b089-8bc374cab9de",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"print(user_prompt_for(ed))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "b23b39b4-78a3-4694-8c89-f2ce56b628f2",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"user_prompt=user_prompt_for(sel)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "ce29c83c-7b47-43a8-8f92-c2a1aa36f8f5",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"messages=[\n",
|
|
" { \"role\":\"system\", \"content\":system_prompt},\n",
|
|
" { \"role\":\"user\", \"content\":user_prompt}\n",
|
|
"]"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "1f120702-029e-4c1a-8ffb-2c4944110aa8",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"response=openai.chat.completions.create(model=MODEL,messages=messages)\n",
|
|
"\n",
|
|
"print(response.choices[0].message.content)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "e9326415-6d35-4750-b9b1-1ae83a86d6f7",
|
|
"metadata": {},
|
|
"source": [
|
|
"### Selenium Version"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "ba86d4cc-cf4c-4f75-aa57-4126b15463b7",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# making sure we're in the virtual environment\n",
|
|
"import sys\n",
|
|
"print(sys.executable)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "2ba86dfa-1e91-4535-9c93-3838c46aee52",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# !pip install selenium"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "01771002-b10f-4681-8710-0f1515866c92",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# !pip install webdriver-manager"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "c19b582d-a355-4c20-8028-42a802e7dca5",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"from selenium import webdriver\n",
|
|
"from selenium.webdriver.edge.service import Service\n",
|
|
"# for edge only:\n",
|
|
"from webdriver_manager.microsoft import EdgeChromiumDriverManager"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "978ab0b9-b42b-4136-8383-79b3f84e084b",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# works for edge only. Do not close the window that pops up as t will be used to open sites given.\n",
|
|
"driver=webdriver.Edge(service=Service(EdgeChromiumDriverManager().install()))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "7dfdeb48-562e-44d3-9044-157d616835fd",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# creating a similar class as bsWebsie but using selenium\n",
|
|
"class SelWebsite:\n",
|
|
"\n",
|
|
" def __init__(self,url,driver):\n",
|
|
" self.driver=driver\n",
|
|
" self.driver.get(url)\n",
|
|
" \n",
|
|
" self.url=self.driver.current_url\n",
|
|
" self.title=self.driver.title\n",
|
|
" self.text=self.driver.find_element(By.TAG_NAME,\"body\").text"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "6174105d-c123-4032-afa8-75588c0f1133",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# testing it on OpenAI website\n",
|
|
"gpt=SelWebsite(\"https://openai.com\",driver)\n",
|
|
"print(gpt.url)\n",
|
|
"print(gpt.driver)\n",
|
|
"print(gpt.title)\n",
|
|
"print(gpt.text)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "bde84abf-09dd-4a56-b6a7-4e5a34c1098e",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": []
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "b7208f3f-6245-48a4-a5ae-d0b59550ee28",
|
|
"metadata": {},
|
|
"source": [
|
|
"##### Troubleshooting in case of errors:\n",
|
|
"1. Make sure the window popped up wasn't closed.\n",
|
|
"2. If the below cell results in any text except an error - driver ID is valid. In this case, quit and restart the driver again.\n",
|
|
"3. If driver ID is invalid, activate driver again using below cells."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "30afa4d1-1ce6-4bad-820e-b72cf3eef959",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# use the following code to check for valid session ID for driver if error occurs:\n",
|
|
"print(driver.session_id)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "154ace93-47b2-40ea-9d49-c6c598a67144",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# if above is valid but still results in trouble, run both; otherwise run only the second part:\n",
|
|
"# driver.quit()\n",
|
|
"# driver = webdriver.Edge(service=Service(EdgeChromiumDriverManager().install()))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "07e74ec5-fda6-462f-b929-7d173b0bdb31",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"print(user_prompt_for(gpt))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "b5d0fd2e-949a-4358-b963-1395157618d2",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"messages2=[\n",
|
|
" {\"role\":\"system\",\"content\":system_prompt},\n",
|
|
" {\"role\":\"user\",\"content\":user_prompt_for(gpt)}\n",
|
|
"]"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "db457f5c-e1be-4087-932d-25ba4880b3ac",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"response=openai.chat.completions.create(model=MODEL,messages=messages2)\n",
|
|
"\n",
|
|
"print(response.choices[0].message.content)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "d448018f-f363-4af9-8ae3-88cc4408da91",
|
|
"metadata": {},
|
|
"source": [
|
|
"### Now let's build a summarize function which can be called directly to summarize any site."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "690ca16b-4b9c-4ddc-b21e-1e69b1d3135a",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"def summarize(site_url):\n",
|
|
" \"\"\"\n",
|
|
" Summarizes the visible content of a website.\n",
|
|
" - Tries BeautifulSoup parsing first (bsWebsite)\n",
|
|
" - Falls back to Selenium parsing (SelWebsite) if BS4 fails\n",
|
|
" - Uses llama3.2 to generate a summary in Markdown\n",
|
|
" \"\"\"\n",
|
|
" try:\n",
|
|
" site=bsWebsite(site_url)\n",
|
|
" except Exception as e:\n",
|
|
" print(f\"BS4 failed: {e}\\nTrying Selenium...\\n\")\n",
|
|
" site=SelWebsite(site_url,driver)\n",
|
|
"\n",
|
|
" messages3=[\n",
|
|
" {\"role\":\"system\",\"content\":system_prompt},\n",
|
|
" {\"role\":\"user\",\"content\":user_prompt_for(site)}\n",
|
|
" ]\n",
|
|
"\n",
|
|
" print(f\"\\nSummarizing: {site.title}\\nURL: {site.url}\\n\")\n",
|
|
"\n",
|
|
" response=openai.chat.completions.create(model=MODEL,messages=messages3)\n",
|
|
"\n",
|
|
" print(response.choices[0].message.content)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "2744296c-ebbd-4696-8517-d14234af9a65",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"summarize(\"https://www.udemy.com\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "9d0d2379-c8b3-4900-8671-179303c00929",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": []
|
|
}
|
|
],
|
|
"metadata": {
|
|
"kernelspec": {
|
|
"display_name": "Python 3 (ipykernel)",
|
|
"language": "python",
|
|
"name": "python3"
|
|
},
|
|
"language_info": {
|
|
"codemirror_mode": {
|
|
"name": "ipython",
|
|
"version": 3
|
|
},
|
|
"file_extension": ".py",
|
|
"mimetype": "text/x-python",
|
|
"name": "python",
|
|
"nbconvert_exporter": "python",
|
|
"pygments_lexer": "ipython3",
|
|
"version": "3.11.13"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|