fix: clean description (#88)

README.md

@@ -67,6 +67,7 @@ Optional
 ├── job_type (enum): fulltime, parttime, internship, contract
 ├── proxy (str): in format 'http://user:pass@host:port' or [https, socks]
 ├── is_remote (bool)
+├── full_description (bool): fetches full description for Indeed / LinkedIn (much slower)
 ├── results_wanted (int): number of job results to retrieve for each site specified in 'site_type'
 ├── easy_apply (bool): filters for jobs that are hosted on LinkedIn
 ├── country_indeed (enum): filters the country on Indeed (see below for correct spelling)

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "python-jobspy"
-version = "1.1.34"
+version = "1.1.36"
 description = "Job scraper for LinkedIn, Indeed, Glassdoor & ZipRecruiter"
 authors = ["Zachary Hampton <zachary@bunsly.com>", "Cullen Watson <cullen@bunsly.com>"]
 homepage = "https://github.com/Bunsly/JobSpy"

src/jobspy/__init__.py

@@ -40,6 +40,7 @@ def scrape_jobs(
     country_indeed: str = "usa",
     hyperlinks: bool = False,
     proxy: Optional[str] = None,
+    full_description: Optional[bool] = False,
     offset: Optional[int] = 0,
 ) -> pd.DataFrame:
     """
@@ -74,6 +75,7 @@ def scrape_jobs(
         is_remote=is_remote,
         job_type=job_type,
         easy_apply=easy_apply,
+        full_description=full_description,
         results_wanted=results_wanted,
         offset=offset,
     )

src/jobspy/jobs/__init__.py

@@ -1,7 +1,7 @@
-from typing import Union, Optional
+from typing import Optional
 from datetime import date
 from enum import Enum
-from pydantic import BaseModel, validator
+from pydantic import BaseModel
 
 
 class JobType(Enum):

src/jobspy/scrapers/__init__.py

@@ -19,6 +19,7 @@ class ScraperInput(BaseModel):
     is_remote: bool = False
     job_type: Optional[JobType] = None
     easy_apply: bool = None  # linkedin
+    full_description: bool = False
     offset: int = 0
 
     results_wanted: int = 15

src/jobspy/scrapers/glassdoor/__init__.py

@@ -5,12 +5,16 @@ jobspy.scrapers.glassdoor
 This module contains routines to scrape Glassdoor.
 """
 import json
-from typing import Optional, Any
+import requests
+from bs4 import BeautifulSoup
+from typing import Optional
 from datetime import datetime, timedelta
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from ..utils import count_urgent_words, extract_emails_from_text
 
 from .. import Scraper, ScraperInput, Site
 from ..exceptions import GlassdoorException
-from ..utils import create_session
+from ..utils import create_session, modify_and_get_description
 from ...jobs import (
     JobPost,
     Compensation,
@@ -66,50 +70,70 @@ class GlassdoorScraper(Scraper):
         jobs_data = res_json["data"]["jobListings"]["jobListings"]
 
         jobs = []
-        for i, job in enumerate(jobs_data):
-            job_url = res_json["data"]["jobListings"]["jobListingSeoLinks"][
-                "linkItems"
-            ][i]["url"]
-            if job_url in self.seen_urls:
-                continue
-            self.seen_urls.add(job_url)
-            job = job["jobview"]
-            title = job["job"]["jobTitleText"]
-            company_name = job["header"]["employerNameFromSearch"]
-            location_name = job["header"].get("locationName", "")
-            location_type = job["header"].get("locationType", "")
-            age_in_days = job["header"].get("ageInDays")
-            is_remote, location = False, None
-            date_posted = (datetime.now() - timedelta(days=age_in_days)).date() if age_in_days else None
-
-            if location_type == "S":
-                is_remote = True
-            else:
-                location = self.parse_location(location_name)
-
-            compensation = self.parse_compensation(job["header"])
-
-            job = JobPost(
-                title=title,
-                company_name=company_name,
-                date_posted=date_posted,
-                job_url=job_url,
-                location=location,
-                compensation=compensation,
-                is_remote=is_remote
-            )
-            jobs.append(job)
+        with ThreadPoolExecutor(max_workers=self.jobs_per_page) as executor:
+            future_to_job_data = {executor.submit(self.process_job, job): job for job in jobs_data}
+            for future in as_completed(future_to_job_data):
+                job_data = future_to_job_data[future]
+                try:
+                    job_post = future.result()
+                    if job_post:
+                        jobs.append(job_post)
+                except Exception as exc:
+                    raise GlassdoorException(f'Glassdoor generated an exception: {exc}')
 
         return jobs, self.get_cursor_for_page(
             res_json["data"]["jobListings"]["paginationCursors"], page_num + 1
         )
 
+    def process_job(self, job_data):
+        """Processes a single job and fetches its description."""
+        job_id = job_data["jobview"]["job"]["listingId"]
+        job_url = f'{self.url}/job-listing/?jl={job_id}'
+        if job_url in self.seen_urls:
+            return None
+        self.seen_urls.add(job_url)
+        job = job_data["jobview"]
+        title = job["job"]["jobTitleText"]
+        company_name = job["header"]["employerNameFromSearch"]
+        location_name = job["header"].get("locationName", "")
+        location_type = job["header"].get("locationType", "")
+        age_in_days = job["header"].get("ageInDays")
+        is_remote, location = False, None
+        date_posted = (datetime.now() - timedelta(days=age_in_days)).date() if age_in_days else None
+
+        if location_type == "S":
+            is_remote = True
+        else:
+            location = self.parse_location(location_name)
+
+        compensation = self.parse_compensation(job["header"])
+
+        try:
+            description = self.fetch_job_description(job_id)
+        except Exception as e :
+            description = None
+
+        job_post = JobPost(
+            title=title,
+            company_name=company_name,
+            date_posted=date_posted,
+            job_url=job_url,
+            location=location,
+            compensation=compensation,
+            is_remote=is_remote,
+            description=description,
+            emails=extract_emails_from_text(description) if description else None,
+            num_urgent_words=count_urgent_words(description) if description else None,
+        )
+        return job_post
+
     def scrape(self, scraper_input: ScraperInput) -> JobResponse:
         """
         Scrapes Glassdoor for jobs with scraper_input criteria.
         :param scraper_input: Information about job search criteria.
         :return: JobResponse containing a list of jobs.
         """
+        scraper_input.results_wanted = min(900, scraper_input.results_wanted)
         self.country = scraper_input.country
         self.url = self.country.get_url()
 
@@ -143,6 +167,41 @@ class GlassdoorScraper(Scraper):
 
         return JobResponse(jobs=all_jobs)
 
+    def fetch_job_description(self, job_id):
+        """Fetches the job description for a single job ID."""
+        url = f"{self.url}/graph"
+        body = [
+            {
+                "operationName": "JobDetailQuery",
+                "variables": {
+                    "jl": job_id,
+                    "queryString": "q",
+                    "pageTypeEnum": "SERP"
+                },
+                "query": """
+                query JobDetailQuery($jl: Long!, $queryString: String, $pageTypeEnum: PageTypeEnum) {
+                    jobview: jobView(
+                        listingId: $jl
+                        contextHolder: {queryString: $queryString, pageTypeEnum: $pageTypeEnum}
+                    ) {
+                        job {
+                            description
+                            __typename
+                        }
+                        __typename
+                    }
+                }
+                """
+            }
+        ]
+        response = requests.post(url, json=body, headers=GlassdoorScraper.headers())
+        if response.status_code != 200:
+            return None
+        data = response.json()[0]
+        desc = data['data']['jobview']['job']['description']
+        soup = BeautifulSoup(desc, 'html.parser')
+        return modify_and_get_description(soup)
+
     @staticmethod
     def parse_compensation(data: dict) -> Optional[Compensation]:
         pay_period = data.get("payPeriod")
@@ -231,12 +290,11 @@ class GlassdoorScraper(Scraper):
         for job_type in JobType:
             if job_type_str in job_type.value:
                 return [job_type]
-        return None
 
     @staticmethod
-    def parse_location(location_name: str) -> Location:
+    def parse_location(location_name: str) -> Location | None:
         if not location_name or location_name == "Remote":
-            return None
+            return
         city, _, state = location_name.partition(", ")
         return Location(city=city, state=state)
 
@@ -245,7 +303,6 @@ class GlassdoorScraper(Scraper):
         for cursor_data in pagination_cursors:
             if cursor_data["pageNumber"] == page_num:
                 return cursor_data["cursor"]
-        return None
 
     @staticmethod
     def headers() -> dict:

src/jobspy/scrapers/indeed/__init__.py

@@ -21,6 +21,7 @@ from ..utils import (
     extract_emails_from_text,
     create_session,
     get_enum_from_job_type,
+    modify_and_get_description
 )
 from ...jobs import (
     JobPost,
@@ -78,7 +79,7 @@ class IndeedScraper(Scraper):
         if sc_values:
             params["sc"] = "0kf:" + "".join(sc_values) + ";"
         try:
-            session = create_session(self.proxy, is_tls=True)
+            session = create_session(self.proxy)
             response = session.get(
                 f"{self.url}/jobs",
                 headers=self.get_headers(),
@@ -140,7 +141,8 @@ class IndeedScraper(Scraper):
             date_posted = datetime.fromtimestamp(timestamp_seconds)
             date_posted = date_posted.strftime("%Y-%m-%d")
 
-            description = self.get_description(job_url)
+            description = self.get_description(job_url) if scraper_input.full_description else None
+
             with io.StringIO(job["snippet"]) as f:
                 soup_io = BeautifulSoup(f, "html.parser")
                 li_elements = soup_io.find_all("li")
@@ -246,9 +248,7 @@ class IndeedScraper(Scraper):
             return None
 
         soup = BeautifulSoup(job_description, "html.parser")
-        text_content = " ".join(soup.get_text(separator=" ").split()).strip()
-
-        return text_content
+        return modify_and_get_description(soup)
 
     @staticmethod
     def get_job_type(job: dict) -> list[JobType] | None:

src/jobspy/scrapers/linkedin/__init__.py

@@ -4,23 +4,36 @@ jobspy.scrapers.linkedin
 
 This module contains routines to scrape LinkedIn.
 """
+import time
 import random
 from typing import Optional
 from datetime import datetime
 
 import requests
-import time
 from requests.exceptions import ProxyError
-from bs4 import BeautifulSoup
-from bs4.element import Tag
 from threading import Lock
+from bs4.element import Tag
+from bs4 import BeautifulSoup
 from urllib.parse import urlparse, urlunparse
 
 from .. import Scraper, ScraperInput, Site
 from ..exceptions import LinkedInException
 from ..utils import create_session
-from ...jobs import JobPost, Location, JobResponse, JobType, Country, Compensation
-from ..utils import count_urgent_words, extract_emails_from_text, get_enum_from_job_type, currency_parser
+from ...jobs import (
+    JobPost,
+    Location,
+    JobResponse,
+    JobType,
+    Country,
+    Compensation
+)
+from ..utils import (
+    count_urgent_words,
+    extract_emails_from_text,
+    get_enum_from_job_type,
+    currency_parser,
+    modify_and_get_description
+)
 
 
 class LinkedInScraper(Scraper):
@@ -111,7 +124,7 @@ class LinkedInScraper(Scraper):
 
                 # Call process_job directly without threading
                 try:
-                    job_post = self.process_job(job_card, job_url)
+                    job_post = self.process_job(job_card, job_url, scraper_input.full_description)
                     if job_post:
                         job_list.append(job_post)
                 except Exception as e:
@@ -123,7 +136,7 @@ class LinkedInScraper(Scraper):
         job_list = job_list[: scraper_input.results_wanted]
         return JobResponse(jobs=job_list)
 
-    def process_job(self, job_card: Tag, job_url: str) -> Optional[JobPost]:
+    def process_job(self, job_card: Tag, job_url: str, full_descr: bool) -> Optional[JobPost]:
         salary_tag = job_card.find('span', class_='job-search-card__salary-info')
 
         compensation = None
@@ -160,7 +173,7 @@ class LinkedInScraper(Scraper):
             if metadata_card
             else None
         )
-        date_posted = None
+        date_posted = description = job_type = None
         if datetime_tag and "datetime" in datetime_tag.attrs:
             datetime_str = datetime_tag["datetime"]
             try:
@@ -169,9 +182,8 @@ class LinkedInScraper(Scraper):
                 date_posted = None
         benefits_tag = job_card.find("span", class_="result-benefits__text")
         benefits = " ".join(benefits_tag.get_text().split()) if benefits_tag else None
-
-        # removed to speed up scraping
-        # description, job_type = self.get_job_description(job_url)
+        if full_descr:
+            description, job_type = self.get_job_description(job_url)
 
         return JobPost(
             title=title,
@@ -182,10 +194,10 @@ class LinkedInScraper(Scraper):
             job_url=job_url,
             compensation=compensation,
             benefits=benefits,
-            # job_type=job_type,
-            # description=description,
-            # emails=extract_emails_from_text(description) if description else None,
-            # num_urgent_words=count_urgent_words(description) if description else None,
+            job_type=job_type,
+            description=description,
+            emails=extract_emails_from_text(description) if description else None,
+            num_urgent_words=count_urgent_words(description) if description else None,
         )
 
     def get_job_description(
@@ -214,7 +226,7 @@ class LinkedInScraper(Scraper):
 
         description = None
         if div_content:
-            description = " ".join(div_content.get_text().split()).strip()
+            description = modify_and_get_description(div_content)
 
         def get_job_type(
             soup_job_type: BeautifulSoup,

src/jobspy/scrapers/utils.py

@@ -8,6 +8,15 @@ from requests.adapters import HTTPAdapter, Retry
 from ..jobs import JobType
 
 
+def modify_and_get_description(soup):
+    for li in soup.find_all('li'):
+        li.string = "- " + li.get_text()
+
+    description = soup.get_text(separator='\n').strip()
+    description = re.sub(r'\n+', '\n', description)
+    return description
+
+
 def count_urgent_words(description: str) -> int:
     """
     Count the number of urgent words or phrases in a job description.

src/jobspy/scrapers/ziprecruiter/__init__.py

@@ -10,14 +10,13 @@ import re
 from datetime import datetime, date
 from typing import Optional, Tuple, Any
 
-import requests
 from bs4 import BeautifulSoup
 from concurrent.futures import ThreadPoolExecutor
 
 from .. import Scraper, ScraperInput, Site
 from ..exceptions import ZipRecruiterException
-from ..utils import count_urgent_words, extract_emails_from_text, create_session
 from ...jobs import JobPost, Compensation, Location, JobResponse, JobType, Country
+from ..utils import count_urgent_words, extract_emails_from_text, create_session, modify_and_get_description
 
 
 class ZipRecruiterScraper(Scraper):
@@ -107,9 +106,9 @@ class ZipRecruiterScraper(Scraper):
         title = job.get("name")
         job_url = job.get("job_url")
 
-        description = BeautifulSoup(
-            job.get("job_description", "").strip(), "html.parser"
-        ).get_text()
+        job_description_html = job.get("job_description", "").strip()
+        description_soup = BeautifulSoup(job_description_html, "html.parser")
+        description = modify_and_get_description(description_soup)
 
         company = job["hiring_company"].get("name") if "hiring_company" in job else None
         country_value = "usa" if job.get("job_country") == "US" else "canada"